UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Can countertransference manifestations be identified during counselling sessions? De Vita, Elsie Lorna 2002

You don't seem to have a PDF reader installed, try download the pdf

Item Metadata

Download

Media
[if-you-see-this-DO-NOT-CLICK]
[if-you-see-this-DO-NOT-CLICK]
ubc_2002-750078.pdf [ 8.72MB ]
Metadata
JSON: 1.0054557.json
JSON-LD: 1.0054557+ld.json
RDF/XML (Pretty): 1.0054557.xml
RDF/JSON: 1.0054557+rdf.json
Turtle: 1.0054557+rdf-turtle.txt
N-Triples: 1.0054557+rdf-ntriples.txt
Original Record: 1.0054557 +original-record.json
Full Text
1.0054557.txt
Citation
1.0054557.ris

Full Text

Can Countertransference Manifestations be Identified During Counselling Sessions? By ELSIE LORNA DE VITA B.A., University of British Columbia, 1989 M A , University of British Columbia, 1995  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTORATE OF PHILOSOPHY In THE FACULTY OF GRADUATE STUDIES (Department of Educational and Counselling Psychology, and Special Education)  We accept this dissertation as conforming to the required standard  UNIVERSITY OF BRITISH COLUMBIA July, 2002 © ELSIE LORNA DE VITA, 2002  UBC  Rare Books and Special Collections - Thesis Authorisation Form  In p r e s e n t i n g t h i s t h e s i s i n p a r t i a l f u l f i l m e n t o f the requirements f o r an advanced degree a t the U n i v e r s i t y o f B r i t i s h Columbia, I a g r e e t h a t t h e L i b r a r y s h a l l make i t f r e e l y a v a i l a b l e f o r r e f e r e n c e and s t u d y . I f u r t h e r a g r e e t h a t p e r m i s s i o n f o r e x t e n s i v e c o p y i n g o f t h i s t h e s i s f o r s c h o l a r l y p u r p o s e s may b e g r a n t e d b y t h e h e a d o f my department o r by h i s o r her r e p r e s e n t a t i v e s . I t i s understood t h a t copying o r p u b l i c a t i o n of t h i s t h e s i s for f i n a n c i a l gain s h a l l not be a l l o w e d w i t h o u t my w r i t t e n p e r m i s s i o n .  Department o f The U n i v e r s i t y o f B r i t i s h V a n c o u v e r , Canada  Columbia  http .//www.library.ubc.ca/spcoll/thesauth.html  9/23/02  Abstract This study investigated two main research questions: first, could countertransference manifestations, operationalized as counsellor over-involvement and under-involvement, be reliably identified by independent judges observing videotapes o f actual therapy sessions; and second, was there evidence to support the contention that counsellor over-involvement and under-involvement were valid indicators o f countertransference behaviour. A multiple case study research design was employed to research this phenomenon. In order to respond to the second research question, this study had to determine whether the first research question could be answered with confidence. Thus, a methodology was employed for the first research question that maximized the reliability o f measuring counsellor over-involvement and under-involvement. A generalizability (G) study was conducted to assess the dependability o f the behavioural measure o f countertransference. The G study helped to design the decision (D) study (e.g., how many counsellor-client dyads, sessions, and judges would be needed to obtain a reliable measure o f counsellor over-involvement and under-involvement). The D study included two counsellor-client dyads across eight therapy sessions. Three judges used videotapes and transcriptions o f the sessions to rate each counsellor response for over-involvement and under-involvement using a 7-point Likert scale (i.e., 3 = under-involved; -2 = somewhat under-involved; -1 = possibly under-involved; 0 = empathically involved; +1 = possibly over-involved; +2 = somewhat over-involved; +3 = over-involved). This study confirmed that counsellor over-involvement and underinvolvement could be reliably observed by independent judges. The average intra-class correlation across eight therapy sessions was .76 for Counsellor One and .79 for Counsellor Two. A moving averages graphing procedure was used to identify episodes where each counsellor's over or under-involved response departed from their individual baseline, using sessions with reliability coefficients greater than .75. These episodes were  used as the focus for research question two, investigating indicators o f countertransference. A Qualitative (Q) study was conducted to respond to the second research question. Data was collected from multiple information sources (e.g., episodes o f over and under-involvement from session transcripts, counsellor session notes, supervision notes, and counsellor and supervisors ratings). These data were then analyzed qualitatively by triangulating the data and identifying themes. The results suggested that counsellor over-involvement and under-involvement are interpretable as valid indicators o f countertransference behaviour.  iv Table o f Contents  Abstract....:  ii  Table o f Contents  iv  List o f Tables  viii  List o f Figures  i  List o f Graphs  x  List o f Appendices  xi  Acknowledgement  xii  Chapter I: Introduction  x  1  Statement o f Problem  2  Purpose o f the Study  4  Design  4  Implications  5  Chapter II: Literature Review  7  The History o f the Construct  7  Classical Perspective  7  Totalistic Perspective  8  Theoretical Models  10  Clinical Research on Countertransference  14  Empirical Research on Countertransference  15  C T as Perceptual Distortion  15  C T as Mental Activity  17  C T as Withdrawal o f Personal Involvement or Avoidance  18  C T as Over-Involvement and Under-Involvement  25  C T as Deviations from Baseline  26  Summary  28  V  Restatement o f Purpose Chapter III: Methodology Step 1: Design  29 31 31  Multiple Case Research  31  Generalizability Theory  33  Generalizability Study  34  Research Design Summary  40  Decision Study: Research Question One  41  Qualitative Study: Research Question T w o  41  Step 2: Recruitment o f Participants Counsellor Recruitment Counsellor demographics Client Recruitment Client demographics  41 42 43 .44 45  Research Supervisor Recruitment and Demographics  45  Judges  45 Judges' demographics  Step 3: Treatment Implementation and Data Collection  46 46  Treatment  46  Session Notes  48  Supervision Sessions  48  Counsellors' & Supervisors' Ratings Identifying Potential Triggers  48  Descriptive Measures  49  Demographic questionnaire  49  B r i e f Symptom Inventory  50  Countertransference Factors Inventory - Revised  50  Step 4: D Study: Research Question One  51  vi Data Preparation  51  Rating Countertransference Behavioural Manifestations  51  Judges' training Data Analysis Step 5: Q Study: Research Question T w o Data Preparation  52 53 53 54  Client stimulus  54  Counsellor verbalizations  54  Session notes  54  Supervision notes  55  Counsellors' & supervisors' ratings identifying potential triggers Data Analysis Summary Chapter I V : Results Counsellor and Client Measures  55 56 56 58 58  Countertransference Factors Inventory- Revised  58  B r i e f Symptom Inventory  58  The D Study: Research Question One  59  Descriptive statistics  59  Reliability  61  Identification o f countertransference episodes  64  The Q Study: Research Question T w o  65  Convergence o f data sources: Establishing support for episodes as Indicators o f C T  65  Direct and Indirect Confirmation : Counsellor One  77  Direct and Indirect Confirmation : Counsellor T w o  80  vii Thematic Analysis: Identifying Common Triggers  109  Findings o f the Q Study  109  Chapter V : Discussion  112  Overview o f Study  112  K e y Findings: D Study  114  K e y Findings: Q Study  115  Study L i n k s to the Literature  116  Methodological Contributions o f the Study  119  Limits o f the Study  121  External validity  121  Internal validity  122  Future Research  124  Implications for Practice and Training  126  Summary  126  Footnotes  128  References  129  viii List o f Tables  Table 1:  Decision Study for Therapist's C T Behaviour Observations  40  Table 2:  Percentage o f Ratings  60  Table 3:  Means and Standard Deviations  61  Table 4:  Inter-rater Reliability - Three Judges  62  Table 5:  Inter-rater Reliability - Pairs o f Judges  64  Table 6:  C T Episodes for Counsellor One  87  Table 7:  C T Episodes for Counsellor T w o  95  Table 8:  Potential Triggers Identified in Client Stimulus  110  ix List of Figures  Figure 1:  Wilson and L i n d y ' s Conceptualization o f C T  11  Figure 2:  Flowchart Depicting Methodology  32  Figure 3:  Flowchart Depicting Sources o f Data  47  List o f Graphs Graph 1: M o v i n g Averages for Combined Ratings: Counsellor One - Session 3 ... .66 Graph 2: M o v i n g Averages for Combined Ratings: Counsellor One - Session 4 ... .67 Graph 3: M o v i n g Averages for Combined Ratings: Counsellor One - Session 5 ... .68 Graph 4: M o v i n g Averages for Combined Ratings: Counsellor One - Session 6 ... .69 Graph 5: M o v i n g Averages for Combined Ratings: Counsellor One - Session 7 ... .70 Graph 6: M o v i n g Averages for Combined Ratings: Counsellor T w o - Session 3 ... .71 Graph 7: M o v i n g Averages for Combined Ratings: Counsellor T w o - Session 4 ... .72 Graph 8: M o v i n g Averages for Combined Ratings: Counsellor T w o - Session 5 ... .73 Graph 9: M o v i n g Averages for Combined Ratings: Counsellor T w o - Session 6 ... .74 Graph 10: M o v i n g Averages for Combined Ratings: Counsellor T w o - Session 1...75  List o f Appendices  Appendix A : Generalizability Study  133  Appendix B :  Training Manual and Rating Sheet  145  Appendix C :  Counsellors and Clients' Information Sheets and Consent Forms  152  Previous Supervisors' Information Sheet  159  Research Supervisor's Information Sheet  160  Judges Information Sheet  161  Appendix D : Demographic Questionnaire  162  Appendix E :  C T Triggers Rating F o r m for Counsellors and Supervisors  165  Appendix F :  Countertransference Factors Inventory-Revised Face Sheet ( C F I - R ) . . . 171  Appendix G :  Study Announcements  176  Appendix H : B r i e f Symptom Inventory Face Sheet (BSI)  180  Appendix I:  Session Notes  184  Appendix J:  Research Supervisor's Questions  186  Appendix K : Frequency Distributions  188  Appendix L :  Graphs for Three Judges  195  Appendix N :  Ethical Approval for G Study and D Study  206  Appendix O: Debriefing  209  Xll  Acknowledgement  I would like to thank my research supervisors, Beth Haverkamp, Richard Young, and Nand Kishor for remaining interested in this project, particularly when I was losing steam and doubting myself. Thank you to the counsellors and clients who participated in this study. Without their willingness to allow their intimate moments to be observed, this type o f process research could not be conducted. I am indebted to them for their courage and generousity. I am grateful for the love and support from my friends, old and new, who have stuck by me during my darkest hours. They provided a space for my tears, black humour, and laughter. I am honoured to have them in my life - Deena Chochinov, Geoff Clarke, Andrew Feldmar, Meredith Feldmar, Soma Feldmar, Daniela Gola, Marcie Harrison, Ramona Hass, Zoran Kondzulovic, Delyse Ledgard, Scott Lowrey, Eric Posen, Jamie Sork, T o m Sork, Patricia Wilensky, and Marshall Wilensky. I would like to give special thanks to Heather Gretton, Nadine K e l l n , L e e Kotsalis, and Olivia Scalzo for their friendship and the countless hours they devoted to this project. I owe all o f you big time. Thank you to my family for their perpetual patience as I finished this project. Being able to "come and go" as I needed was a relief. Finally, this dissertation is dedicated to the memory o f my parents, Antonio and Caterina D e Vita. Although they did not understand what a "psychologist" did for a living, they barely blinked an eye when I announced, "I want to study psychology!" I am grateful that they encouraged me to follow my heart. A n d now, I dance Dance when you're broken open. Dance when you've torn the bandage off. Dance in the middle o f the fighting. Dance in your blood. Dance, when you're perfectly free. Rumi  1 Chapter I Introduction Sigmund Freud first wrote about countertransference in 1910; he described countertransference as the therapist's conflictual emotional reactions to a client's material that stemmed from unresolved issues in the therapist's unconscious (e.g., cited in Robbins & Jolkovski, 1987; Singer & Luborsky, 1977; Y u l i s & Kiesler, 1968). According to Freud and supporters o f the classical psychoanalytic position, these reactions - the result o f anxiety and defense - were deemed antithetical to the therapeutic process. The therapist was encouraged to understand his or her reactions to the client, either through supervision or analysis, so that their effects on the process could be minimized. Over the years others have conceptualized countertransference more broadly, adopting what is often referred to as the "totalistic" perspective (e.g. Kernberg, 1965). From this view, all feelings and thoughts a therapist holds in response to a client, both conscious and unconscious, are called countertransference (e.g., Heimann, 1950). These feelings are thought to increase the therapist's understanding o f the client and facilitate the empathic process. In other words, in this view countertransference is seen as an important therapeutic tool, essential to client growth (e.g., Winnicott, 1949). Some have incorporated both the classical and totalistic perspectives. For example, countertransference has also been described as a concept that includes elements that influence the therapeutic relationship both positively and negatively (e.g., Blanck & Blanck, 1979; Watkins, 1985). In other words, the therapist's blind spots or unresolved issues are seen to have the potential to derail the therapeutic process i f acted upon; however, countertransferential reactions can also serve as a tool for the therapist to understand the client's material (Blanck & Blanck, 1979). This view o f countertransference allows for both conscious and unconscious aspects o f countertransference and does not overlook the potentially detrimental aspects o f countertransference behaviour. Thus, consistent with others' perspective (e.g., Peabody & Gelso, 1982; M c C l u r e & Hodge, 1987; Robbins & Jolkovski, 1987), this combined view clearly differentiates between the effects o f countertransference "feelings and thoughts" and countertransference "behaviour" on the therapeutic process. Another important distinction in the literature, perhaps first articulated by Winnicott  2 (1949), was between objective and subjective countertransference (e.g., Kiesler, 2000; Wilson & Lindy, 1994). Wilson and Lindy (1994) used the terms subjective and objective countertransference to differentiate between reactions on the part of the therapist that "originate from the therapist's personal conflicts, idiosyncrasies, or unresolved issues from life course development" (p. 16) with those that are more universal, or objective, in that most therapists would respond to the client's material in a similar manner. Gelso and Carter (1985, 1994) have argued that where the classical conceptualization of countertransference was too restrictive, the totalistic perspective was too broad. Gelso and Carter (1985) differentiate between reality-based and irrational aspects o f the counsellor's response to the client: the former refers to the therapist's response to the client based on the client's material, whereas the latter refers to the therapist's response to the client based on the therapist's own unresolved issues. Gelso and Carter (1994) defined countertransference as "the therapist's transference to the client's material, both to the transference and nontransference communications presented by the client (p.297). Statement of the Problem Although definitions and conceptualizations o f countertransference have evolved over the years, and continue to be debated, current knowledge o f countertransference is based primarily on clinical writings and analogue research (Hayes, McCracken, McClanahan, H i l l , Harp, & Carozzoni,1998). Unfortunately, attempts to measure countertransference have been mostly superficial to date (e.g., Hayes, McCraken, McClanahan, H i l l , Harp, & Carazzoni; 1998; Singer & Luborsky, 1977).With few exceptions (e.g., Cutler, 1958; M c C l u r e & Hodge, 1987; Rosenberger & Hayes, 1998), there have been no attempts to study countertransference, particularly countertransference behaviour, in a naturalistic setting. O f this research in naturalistic settings, no attempts have been made to address the issue o f construct validity when assessing countertransference. Gelso, Fassinger, Gomez, and Latts (1995) noted that empirical research has fallen behind clinical theory on countertransference: A key impediment to such research revolves around how to operationalize a highly abstract and global construct, such as countertransference. In attempting to simplify this construct for the sake o f measurement, researchers have tended to pick up on only one or another aspect of the global construct. What has been needed is an  3 operational ization o f countertransference that reflects its complex and multidimensional nature, (p. 356) Empirical research has attempted to measure countertransference in a variety o f ways: by measuring the therapist's misperception o f the client (e.g., Cutler, 1958; Fiedler, 1951; M c C l u r e & Hodge, 1987; Snyder & Snyder, 1961); by measuring the influence o f anxiety on the therapeutic relationship (e.g., Bandura, 1956); by observing the therapist's approachavoidance reactions in regards to various client presentations, such as hostility (Bandura, Lipsher, & Miller, 1960; Latts & Gelso, 1995; Rosenberger & Hayes, 1998); by the therapist's withdrawal o f personal involvement i n therapy by excluding h i m or herself from interpretations to the client (e.g., Peabody & Gelso, 1982; Robbins & Jolkovski, 1987; Y u l i s & Kiesler, 1968); by the therapist's mental activity (e.g., Dube & Normandin, 1999); and by the therapist's deviations from his or her typical style (e.g., Holmqvist, 2001). M o r e recently, Hayes, McCraken, McClanahan, H i l l , Harp, and Carazzoni (1998) qualitatively analyzed data from interviews with therapists immediately following their counselling sessions to gain their perspectives on countertransference. They presented a preliminary theory o f countertransference that included various dimensions o f countertransference called origins, triggers, and manifestations. Countertransference origins were defined as areas o f unresolved intrapsychic conflict for the therapist which may serve as "blind spots" for the therapist that can impact the therapeutic relationship i f triggered. Countertransference triggers included therapy events that evoked the therapist's unresolved issues. Frequently researched triggers include the clients' presenting problems, and their presentation style. Countertransference manifestations included therapists' behaviours, thoughts, or feelings that stemmed from the triggering o f their unresolved issues. Consistent with others' conceptualizations o f countertransference behaviour (e.g., Friedman & Gelso, 2000; W i l s o n & Lindy, 1994); Hayes et al.'s (1998) qualitative analysis o f counsellors' interviews identified both the therapists' over-identification or over-involvement with the client, as well as their avoidance or under-involvement with the client as aspects o f countertransference. Hayes et al.'s (1998) preliminary theory seems helpful in conceptualizing the domains o f countertransference (origins, triggers, and manifestations) i n a manner that can be empirically validated. The authors suggested that empirical research should work backward  4  to first try to identify countertransference manifestations, then proceed to identify triggers and origins within the therapist. Until countertransference manifestations can be reliably and validly identified during therapy sessions, research projects investigating the relationship between countertransference and other variables, such as the therapeutic alliance or treatment outcome, w i l l be built on a foundation o f sand. Purpose The purpose of this exploratory research is to attempt to identify behavioural manifestations o f therapist countertransference reactions during actual counselling sessions in a reliable and valid manner. Specifically, this study w i l l address the following research questions: 1.  Can countertransference behavioural manifestations, defined as therapist overinvolvement and under-involvement, be reliably identified during counselling sessions with clients?  2.  Is there evidence to support the contention that counsellor over-involvement and under-involvement are valid indicators o f countertransference behaviour. This study focused on behavioural manifestations o f countertransference for two  reasons: first, behaviour was external to the therapist, hence facilitating observation, and second, it has been identified as the aspect o f countertransference that is more problematic to the therapeutic relationship. Operationalizing countertransference behavioural manifestations as over-involvement and under-involvement (Friedman & Gelso, 2000) was thought to have both clinical and empirical value. Previous research focused solely on therapists' avoidance behaviour (e.g., Rosenberger & Hayes, 1998) and overlooked negative aspects o f therapists' seemingly facilitative behaviours, such as over-supporting or colluding with clients. Design This study employed a mixed research design to investigate countertransference manifestations. First, a multiple case study approach was selected because it was well suited for analyzing interactions intensely within naturalistic settings such as therapy sessions (e.g., Jones, 1993; Y i n , 1989). In addition, generalizability theory (GT: Cronbach, Gleser, Nanda, & Rajartnam, cited in Shavelson & Webb, 1991), a statistical theory about the dependability of behavioural measurements, was also applied to help design the study (e.g., to determine the number o f dyads, sessions, and judges necessary to arrive at dependable ratings o f over-  5 involvement and under-involvement). Given that a central premise o f this study weighed on establishing a dependable measure o f countertransference behaviour, it seemed prudent to address this issue. Generalizability theory w i l l be described further in the methodology section o f this document. Because feelings and thoughts are internal to the therapist, it is challenging to access those aspects o f countertransference during the therapy session. The very act o f asking therapists to attend to their thoughts and feelings towards the client may alter or bias their responses. In order to access the therapist's thoughts and feelings and potential countertransference triggers to support the behavioural observations, this study examined potential convergence o f various sources o f data (e.g., therapists' session notes, research supervisors' supervision notes, post-study ratings by counsellors and supervisors). Clinical supervision is thought to increase understanding of therapists' countertransference reactions (e.g., Singer & Luborsky, 1977); however, little existing research, i f any, includes this facet in the research design. Singer and Luborsky (1977) also suggested that countertransference reactions were generally noted first by third parties (e.g., supervisors) rather than the therapist. Consequently, supervision sessions were also used to further explore the therapists' feelings and reactions to the client and the counselling sessions. The therapists' selfevaluations were requested upon completion o f the therapy sessions in order to avoid biasing their behaviour during sessions. Implications Although the term countertransference has strong links with psychodynamic traditions in therapy, the concept has utility for therapists regardless of theoretical orientation. F o r example, research on the counselling relationship has acknowledged the importance o f both transference and countertransference (Gelso & Carter, 1985). Unfortunately, to date, empirical research on the construct has been limited. I f potential countertransference manifestations, defined as therapist over-involvement and underinvolvement, can be identified by independent judges in this research study, an additional step towards construct validation w i l l have been achieved. A s a result, the relationship between countertransference origins, triggers, and manifestations can be explored further, as well as the relationship between countertransference manifestations and the therapeutic alliance and treatment outcome.  6  If successful, the method employed to identify countertransference manifestations may have important implications for how counselling supervision is conducted. Supervisors would be able to review a counsellor trainee's work for moments of over-involvement and under-involvement. These moments could then be explored with the trainee for possible "blind spots" or personal vulnerabilities. This exploration process could provide depth and insight to supervision and facilitate therapist development.  7  Chapter II Literature Review This section will begin with a historical overview o f the evolution of the term countertransference, including more recent theoretical conceptualizations o f the construct (e.g., Hayes, McCracken, McClanahan, H i l l , Harp, & Carozzoni, 1998; Wilson & Lindy, 1994). Next, the clinical and empirical research on countertransference will be reviewed to summarize what has been established in the field and what remains to be explored in terms o f operationalizing and measuring such an elusive construct. The History o f the Construct: Definitions and Conceptualizations o f Countertransference The definition and therapeutic usefulness o f the term countertransference has evoked much debate during the 20th century. This section will outline the evolution o f the construct, primarily as it has been defined in the psychodynamic literature. There are several comprehensive reviews tracing the history of the concept o f countertransference (e.g., Jacobs, 1999; Singer & Luborsky, 1977; Slakter,1987). This review will highlight the most salient contributions, starting with Sigmund Freud who first coined the term in 1910, and continuing with more recent theoretical models o f countertransference. Classical perspective. According to Freud (1910) and many o f his followers, countertransference is the analyst's unconscious response to the patient's transference, and i f unrecognized by the analyst and uncontrolled, this response can have an adverse effect on the therapeutic process by interfering with the analyst's ability to understand his or her patient. Thus, all attempts must be made by the analyst to minimize the effects o f countertransference. This definition describes the classical psychoanalytical view o f countertransference. Anne Reich (1951, 1960) wrote two notable papers supporting the classical position. In her first paper she stated that countertransference, "comprises the effects of the analyst's own unconscious needs and conflicts on his understanding or technique" (1951, p. 26). She distinguished between acute manifestations o f countertransference, which occurred suddenly under specific circumstances with specific clients, and permanent or chronic manifestations o f countertransference, which reflected more habitual neurotic difficulties o f the therapist that could arise in a variety o f situations across different clients. Reich's (1960) second paper on countertransference attempted to refute the notion  8  that countertransference was a useful force within the therapeutic process. She argued that this idea was a result of a failure to differentiate between the therapist's countertransference, which is generally unconscious, and his or her total response, which includes conscious responses. Reich (1960) stressed that "conscious responses should be regarded as countertransference only if they reach an inordinate intensity or are strongly tainted by inappropriate sexual or aggressive feelings, thus revealing themselves to be determined by unconscious infantile strivings" (p. 390). Ferenczi (cited in Jacobs, 1999), a follower of Freud, disagreed with his mentor's perspective on countertransference. He pointed out that the analyst's reactions to his or her clients were an essential ingredient to empathic understanding. Although highly criticized, Ferenczi was interested in the "role of metacommunications in analysis and of the interplay between the minds of patient and analyst," a view "that was quite remarkable for its time." (Jacobs, 1999, p. 578). Perhaps, Ferenczi's queries paved the way for what later came to be known as the "totalistic perspective" on countertransference. Totalistic perspective. Roughly thirty years after Ferenczi's questioning of Freud's views, others challenged the classical position that countertransference reactions were primarily unconscious and detrimental to the therapeutic process (e.g., Heimann, 1950; Kernberg, 1965; Little, 1951; Winnicott, 1949). This view, labelled the "totalistic" perspective, defined countertransference more broadly to include all feelings and thoughts a therapist holds in response to a client. In this view, the therapist's reactions are thought to increase his or her understanding of the client and facilitate the empathic process. According to Heimann (1950), in order to achieve an empathic understanding of the client's experience and internal world, the therapist engages in a trial identification with the client in order to arouse feelings within him or herself that shed light on the client's material. Thus, in order to respond empathically to clients, the therapist must be able to first feel what the client is feeling (e.g., harness his or her countertransferential feelings). In his classic paper, "Hate in the Countertransference," Winnicott (1949) differentiated between objective countertransference reactions within the therapist (which are based on the client's actual personality) from those countertransference reactions that are more subjective in nature (based on the therapist's personal experiences and development). Winnicott argued  9  that severely disturbed clients could evoke intense countertransference reactions of hate within the therapist. These reactions were to be understood by the therapist so that he or she did not act out during the therapy hour to the detriment of the client. Winnicott (1949) was perhaps thefirstto distinguish between objective and subjective countertransference. Other writers have since adopted this distinction (e.g., Kiesler, 2001; Wilson & Lindy, 1994). In his review of the literature, Kiesler (2001) adapting Winnicott's definition, described objective countertransference as "the constricted feelings, attitudes, and reactions of a therapist that are evoked primarily by the client's maladaptive behaviour and that are generalizable to other therapists and to other significant persons in the client's life" (p. 1057). Kiesler described subjective countertransference as "the defensive and irrational reactions and feelings a therapist experiences with a particular client that represent residual effects of the therapist's own unresolved conflicts and anxieties" (p. 1057). Similarly, Wilson and Lindy's (1994) theoretical model, described below, differentiated between objective and subjective countertransference. Bouchard, Normandin, and Sequin (1995) distinguished between three types of countertransference: rational-objective countertransference, reactive countertransference, and reflective countertransference. Whereas thefirsttwo types correspond to the definitions of objective and subjective countertransference noted above, the latter is consistent with definitions of empathy. The authors developed the Countertransference Rating System (CRS) to measure therapists' mental activity along these dimensions. (This research is described later in this document under the section titled empirical research). Over the years countertransference has come to be viewed by many to include elements that influence the therapeutic relationship both positively and negatively (e.g., Blanck & Blanck, 1979; Watkins, 1985). If acted upon, the therapist's blind spots or unresolved issues may interfere with the therapeutic process. However, if analyzed and understood, the therapist may also use his or her countertransferential feelings as a tool to understand the client's material (Blanck & Blanck, 1979). This view of countertransference allows for both conscious and unconscious aspects of countertransference and does not overlook the potentially detrimental aspects of countertransference behaviour. Where the supporters of the totalistic perspective have argued that the classical  10 conceptualization o f countertransference was too restrictive, others have argued that the totalistic perspective was too broad (e.g., Gelso and Carter 1985, 1994; Gelso & Hayes, 1998, 2002a, 2002b). Gelso and Carter (1985) differentiated between reality-based and irrational aspects o f the counsellor's response to the client: the former refers to the therapist's response to the client based on the client's material, whereas the latter refers to the therapist's response to the client based on the therapist's own unresolved issues. Gelso and Carter (1994) defined countertransference as "the therapist's transference to the client's material, both to the transference and non-transference communications presented by the client (p.297). This definition is more closely aligned with the description o f subjective countertransference, locating the origins o f the therapist's conflictual reactions in his or her unresolved intrapsychic issues. M o s t writers in the area o f countertransference further differentiate between countertransference feelings and behaviours (e.g., McClure & Hodge, 1987; Peabody & Gelso, 1982; Robbins & Jolkovski, 1987). Countertransference feelings can take various forms such as feeling affectionate, nurturant, sexually aroused, pity, frustration, annoyance, and hostility (Snyder & Snyder, 1961). Countertransference behaviour can include acting overly solicitous, seductive, withdrawn, or punitive towards clients. It is generally agreed that acting out countertransference behaviour leads to poor counselling outcomes (e.g., Gelso & Carter, 1985). Similarly, most clinicians and researchers agree that therapists must strive to bring their reactions to clients into awareness so that they can be understood and managed (e.g., Hayes, Riker, & Ingram 1997; Singer & Luborsky, 1977) or even employed therapeutically (e.g., Blanck & Blanck, 1979, Winnicott, 1949). Theoretical models. Theoretical models have also been developed to shed light on this complex construct. Wilson and Lindy (1994) presented a model to study therapists' empathic difficulties and countertransference reactions with clients who experienced post-traumatic stress. Although their model was originally constructed to depict the experience o f therapists who work in the field o f trauma, the authors acknowledge its usefulness with other client populations. The four quadrants o f the model are depicted in Figure 1. Extending the work o f others (e.g., Lindy, 1988; Maroda, 1991; Slatker, 1987; and Wilson, 1989), Wilson and Lindy (1994) differentiated between two categories o f  11 countertransference reactions, Type I (avoidance) and Type II (over-identification). Type I reactions include forms o f denial, minimization, distortions, avoidance, detachment, and withdrawal, whereas Type II reactions involve forms o f over-identification, over-idealization, enmeshment, and excessive advocacy for the client. Both types o f countertransference reactions result in a movement away from an empathic stance by the therapist. The authors note that therapists may experience one style more than another (e.g., over-identification versus avoidance.) Wilson and Lindy's (1994) model also differentiated between objective and subjective countertransference, along with four types o f empathic strain that relate to the different types o f countertransference reactions. The authors referred to subjective and objective countertransference to differentiate between reactions on the part o f the therapist that "originate from the therapist's personal conflicts, idiosyncrasies, or unresolved issues from life  Figure 1. Modes o f Empathic Strain in Countertransference Reactions (CTRs) (Wilson & Lindy, 1994) OBJECTIVE COUNTERTRANSFERENCE Empathic Disequilibrium  Empathic Withdrawal  uncertainty  blank screen facade  vulnerability  intellectualization  unmodulated affect  misperception o f dynamics  Type II Countertransference  Type I Countertransference  Over-identification  Avoidance  Empathic Enmeshment  Empathic Repression  loss o f boundaries  withdrawal  over-involvement  denial  reciprocal dependency  distancing  SUBJECTIVE COUNTERTRANSFERENCE  12 course development" (p. 16), with those that are more universal, or objective, in that most therapists would respond to the client's material in a similar manner. This distinction between objective and subjective countertransference is similar to that made by Winnicott (1949). M o r e recently, Hayes, McCracken, McClanahan, H i l l , Harp, and Carozzoni (1998) developed a preliminary theory based on a qualitative analysis o f data from interviews with eight psychologists (four men and four women), immediately following their counselling sessions to gain their perspectives on countertransference. Therapy sessions ranged between 12-20 session, for a total o f 127 sessions. Post-session interviews with the psychologists were conducted to gather information about their impressions and reactions regarding the session. The researchers analyzed the 127 post-session interviews to identify sections pertaining to countertransference. Sections of the interviews were considered to depict countertransference reactions only if the therapists self-identified their reactions as deriving from unresolved intrapsychic conflict. From this data, Hayes et al. (1998) presented a preliminary theory o f countertransference which included various dimensions o f countertransference called origins, triggers, and manifestations. Countertransference origins were defined as areas of unresolved intrapsychic conflict for the therapist that may serve as "blind spots" for the therapist and which can impact the therapeutic relationship i f triggered. Examples o f origins include: family issues, therapist's needs and values, therapy issues such as termination, and cultural issues. Countertransference triggers were defined as therapy events that evoke the therapist's unresolved issues. Examples of triggers include: the content of the client's material (e.g., death, family o f origin), changes in the therapy structure, therapist's perception of the client (e.g., as dependent or hostile), and the client expressing negative emotion. Countertransference manifestations were described as therapist's behaviours, thoughts, or feelings that are a consequence of unresolved issues being triggered. Examples o f manifestations include: "approach" responses by the therapist (e.g., nurturing, identification, positive feelings towards the client), "avoidance" responses by the therapist (e.g., distancing self from the client, boredom or fatigue, disappointment with the client), and negative feelings by the therapist. Hayes, McCracken, McClanahan, H i l l , Harp, and Carozzoni's (1998) preliminary theory seems helpful in conceptualizing the domains o f countertransference (origins, triggers,  13  and manifestations) in a manner that can be empirically validated. Their conceptualization o f countertransference manifestations includes both approach and avoidance responses by the therapist, which is consistent with others' formulations (e.g., Friedman and Gelso, 2000; Wilson and Lindy, 1994). Unlike previous definitions o f therapist "approach" responses as facilitative o f the therapeutic process, in this context "approach" responses refer to those responses by the therapist that are seemingly helpful, such as over-supporting, but can serve to de-rail the process. Based on others difficulty predicting countertransference behaviour from potential countertransference origins (i.e., Rosenberger & Hayes, 1998), the authors suggested that empirical research should work backwards to first try to identify countertransference manifestations, then proceed to identify triggers and origins within the therapist. To summarize, this section briefly outlined the evolution o f the construct o f countertransference from Freud's (1910) early writings, to more recent conceptualizations. Depending on which definition is applied, the impact o f countertransference reactions on the therapeutic process can be viewed as negative, positive, or both. The differing definitions o f countertransference have led to varying opinions about how it relates to the empathic process. A s noted previously, Heimann (1950), one o f the earliest writers on the "positive" view o f countertransference, described using the therapist's emotional responses, or countertransferential feelings, as the basis to formulate empathic responses. Reich (1960) believed that the therapist's empathic failure in the trial identification with the client is the result o f his or her countertransference. In other words, breaches in the empathic process or a movement away from an empathic stance by the therapist, are due to the therapist's countertransference reactions. Consistent with the "totalistic" view o f countertransference, this researcher believes that countertransference reactions can have both a positive and negative impact on the therapeutic process. In addition, there can be both conscious and unconscious aspects o f countertransference which relate to either the client's material (e.g., objective countertransference) or the therapist's unresolved intrapsychic conflicts (e.g., subjective countertransference). Countertransference "feelings" can help deepen the therapist's understanding o f the client's experience; however, countertransference "behaviour" can  14  interfere with the therapeutic process and lead to empathic failures. This notion o f empathic breaches as a possible indicator that the therapist's countertransference reactions are being aroused has implications for observing countertransference reactions during counselling sessions. Clinical Research on Countertransference The vast clinical literature on countertransference spans over 80 years and includes mostly anecdotal reports in psychoanalytic journals. According to these reports, countertransference feelings towards clients can take various forms such as feeling affectionate, nurturant, sexually aroused, pity, frustration, annoyance, and hostility (e.g., Snyder & Snyder, 1961). A s a result, therapists may find themselves having dreams about their clients, or behaving in an overly solicitous, supportive, or punitive manner towards them. It was generally thought that countertransference behaviour, not feelings, negatively impacted the therapy process (e.g., Gelso & Carter, 1985). In their review o f this clinical literature, Singer and Luborsky (1977) noted that the findings from the extensive number o f anecdotal reports were varied and complex. They summarized the following points: first, countertransference could impede effective treatment because it interfered with the therapist's ability to form a proper identification with the client. Identification was thought to be a necessary part o f the process o f understanding. Second, a sign that countertransference was in operation was i f the therapist had intense sexual or aggressive feelings towards the client. Third, countertransference could be one o f two types: it may arise in response to specific situations and specific clients (acute); or it may occur across clients and conflicts, reflecting a habitual need o f the therapist (chronic). Fourth, the more general definition o f countertransference (totalistic perspective) may be an effective therapeutic tool to assist therapists in empathizing with clients. Five, all authors emphasized the importance o f having countertransference reactions under conscious awareness and control to minimize their detrimental effects on the therapeutic process. Six, the therapist's emotional maturity and self-understanding, usually gained through personal psychotherapy or psychoanalysis, helped to minimize the enactment o f countertransference needs. Seven, countertransference feelings and behaviour could be managed through self-analysis or by consulting with a supervisor or colleague. Finally, countertransference may be observed  15  through peripheral cues such as body movement or changes in the tone o f the therapist's voice. Current reviews o f the clinical literature (e.g., Hinshelwood, 1999; Jacobs, 1999; and Kernberg, 1999) articulated similar themes to those summarized by Singer and Luborsky (1977) twenty five years ago. It seems that theoretical discussions and anecdotal reports continue to be the favoured method o f presenting countertransference reactions in psychoanalytic journals. Recently, Kiesler (2001) questioned whether we can "anchor our divergent constructs o f C T in agreed upon actual therapist behaviours?" (p. 1058). H e proposed that counsellors' "deviations from baselines" could be a possible method to detect both subjective and objective countertransference during therapy sessions. In other words, subjective countertransference may be in operation when the therapist's reactions to a particular client deviate noticeably from his or her usual response (baseline) with the same client or with other clients; objective countertransference may be in operation when the therapist's reactions to the client, or deviations from baseline, are similar to those o f the therapist's colleagues and significant others in the client's life. Prior assessment o f the therapist, his or her colleagues, clients, and the clients' significant others would be necessary to accumulate baseline data. This suggestion holds promise for stimulating research on how to operationalize and measure countertransference behaviour. Empirical Research on Countertransference Several attempts have been made to study countertransference empirically using a variety o f methods. The following section will review these studies according to their conceptualization o f countertransference as, 1) perceptual distortion; 2) mental activity; 3) withdrawal o f personal involvement or avoidance reactions; 4) over-involvement and/or under-involvement; and 5) deviations from baseline. Countertransference as perceptual distortion. In an exploratory study, Fiedler (1951) attempted to measure therapist distortions using a Q-sort technique. Distortions in the therapists' sorts were deemed to indicate the presence o f countertransference (i.e., the amount to which the therapist over or underestimated the client's similarity to himself or the client's similarity to the therapist's ideal). Unfortunately the results were inconclusive, primarily due to small sample size and methodological problems. However, Fiedler's attempt to quantify  16 countertransference influenced future research on the topic (McClure & Hodge, 1987). A classic study by Cutler (1958) operationalized countertransference according to Brunner's (cited in Cutler, 1958) theory o f perception which notes that, "strong need-satisfying hypotheses will tend to be confirmed on the basis o f minimal appropriate information from the environment" (p. 350). The author questioned whether countertransference may be a special case o f perception being influenced by need. T w o therapists and five clients participated: therapist one saw three clients across three consecutive sessions and therapist two saw two clients across four consecutive sessions. A self-report measure was used to collect data on the therapists' conflictual areas. Cutler (1958) concluded that the therapists were less accurate in reporting material about a client whose needs directly related to needs identified in the therapists' own personalities. A s part o f an in-depth study on the psychotherapy relationship, Snyder and Snyder (1961) collected extensive data on one therapist and several o f his clients. The authors developed an affect scale to measure both the therapist's and the clients' emotional reactions during the session. After each therapy session, the therapist and the clients would complete the measure. In addition, the therapist would fill out the affect scale according to how he believed the clients responded. The difference between each client's actual score and the therapist's perception o f his or her ratings was viewed as the therapist's perceptual distortion and was considered an index o f countertransference. The authors found that more negative countertransference effects occurred with clients that had less successful therapy outcomes. Also, there was a trend for countertransference to increase as sessions continued. M c C l u r e and Hodge (1987) explored the relationship between the therapist's attitude o f liking or disliking his or her client and countertransference. Countertransference was operationalized as "the difference between the therapist's perception o f the client's personality and the client's own perception" as measured by the Taylor-Johnson Temperament Analysis (TIJA; Nash, cited in M c C l u r e & Hodge, 1987). Ten doctoral level therapists affiliated with a university counselling centre, two established therapists in private practice, and their 28 clients living in Southern California participated in the study. The authors noted that the T I J A was developed to assess the personality configuration o f either an individual client, a couple, or a family, as well as measure one person's view o f another. The therapists completed the T I J A  17  for themselves and for each o f their clients. The clients also completed the measure for themselves. A n attitude scale embedded within the T I J A was used to assess therapists' attitudes towards their clients (positive, neutral, or negative). The authors found that, for cases in which the therapists held a positive attitude toward the client, they distorted the client's personality in such a way as to make it more similar to their own personality profile. F o r cases in which the therapists indicated a negative attitude towards the client, they distorted the client's personality to be more dissimilar than their own. For cases in which the therapists held a neutral or intermediate attitude towards the client, the therapists did not distort the client's personality to be more or less similar than thenown profiles. The authors argued that the distortions in the therapists' ratings were evidence o f the presence o f countertransference. Countertransference as therapist's mental activity. A s described earlier, Bouchard, Normandin, and Seguin (1995) distinguished between three types o f countertransference: rational-objective countertransference, reactive countertransference, and reflective countertransference. The first type referred to mental activity by the therapist that was detached, observing rather than participating; the second type referred to mental activity by the therapist that was defensive, consistent with the classical definition o f countertransference; and, the third type referred to mental activity by the therapist that was consistent with an engaged, empathic stance. The authors developed the Countertransference Rating System ( C R S ) to measure therapists' mental activity along these dimensions. Recently, Dube and Normandin (1999) applied the C R S to investigate 27 trainee therapists' spontaneous reactions to five clinical vignettes depicting actual client-therapist interactions. The researchers explored whether trainees' personal psychotherapy impacted their listening process (mental activity). After reading each vignette, the trainees were instructed to record their reactions to what was occurring in the vignette. Judges scored the trainees' responses to the vignettes using the C R S . Agreement between the judges was moderately good for the three types o f mental activity (rational-objective k = .52; reactive k = .60; and reflective k = .63). The researchers found that the reflective category, with its various subcategories, was the most frequently observed type o f mental activity. Trainees who had had their own personal therapy were less likely to block out or act on their reactions  18 compared with those who had not had personal therapy. In addition, those who had experienced their own therapy tended to elaborate the client's internal world more extensively than did the "no therapy" group. Personal therapy was found to have no effect on rationalobjective or reactive countertransference. The authors noted that they continue to refine the C R S for research purposes and hope to develop distinct scores for each mental activity to allow profile analyses. Countertransference as withdrawal o f personal involvement or avoidance. Yulis and Kiesler (1968) characterized countertransference as therapist withdrawal o f involvement. They developed a procedure to assess countertransference behaviour in response to three stimulus tapes o f an actress portraying a client who presented as sexual, hostile, and neutral (in terms of the sexual and hostile portrayals). Each 15 minute tape consisted o f 10 segments, allowing for therapist interventions at each stopping point. After each segment, participants were instructed to choose between two written responses: both responses were deemed therapeutically appropriate by expert judges but differed in terms o f degree o f personal involvement. Hence, counsellor involvement or withdrawal was determined by his or her selection o f intervention responses. Consistently selecting responses that excluded personal involvement was viewed as countertransferential behaviour. The authors supported their prediction that participants with low anxiety scores would be more personally involved (i.e., show less countertransference) with their clients than participants with high anxiety scores. However, they did not find that participants showed less countertransference with the neutral tape scenario versus the sexual and hostile scenario. Using Yulis and Kiesler's (1968) procedure to investigate countertransference, Peabody and Gelso (1982) studied the complex relationship between counsellor empathy, awareness, and amount o f countertransference behaviour. Twenty-two male doctoral students in counselling psychology participated in the study. Overall, the authors found that empathy was negatively related to countertransference behaviour in the sexual analog scenario, but not in the other two scenarios. In addition, counsellor empathic ability was positively related to counsellor reports o f openness to countertransference feelings. Countertransference reactions by the therapist have often been operationalized as "avoidance reactions." In general, avoidance reactions refer to those responses by the  19 therapist that are designed to inhibit further exploration o f a topic (e.g.. disapproval, changing topics, silence), whereas approach reactions refer to those responses by the therapist that are designed to facilitate further exploration o f a topic (e.g., reflection, approval, instigation) (e.g., Bandura, Lipsher, & Miller, 1960; Latts & Gelso, 1995). Bandura, Lipsher, and Miller, (1960), investigated therapists' countertransference, operationalized as avoidance reactions, to clients' expressions o f hostility. They investigated three hypotheses: 1) that therapists who experienced high anxiety around client hostility would exhibit avoidance behaviour, (rather than encouragement or approach behaviour) compared to therapists who experienced low hostility anxiety; 2) that patients would be encouraged to express further hostility i f the therapist responded with approach versus avoidance reactions; and 3) that i f the therapists responded with an avoidance versus an approach response to the client's expression o f hostility, the client would be more likely to change the object o f their hostility. The authors analyzed tape recordings from 110 therapy sessions obtained from 17 clients and 12 therapists. The counsellors' responses were coded for approach and avoidance reactions by two judges. Inter-judge reliability for the counsellor's approach and avoidance responses was assessed by the degree o f agreement between the two judges in coding the response units. T w o hundred and sixty one o f the response units scored were in perfect agreement; 100 units showed minor discrepancies, mainly due to some overlap in the categories that were coded; and 37 units were rated in the opposite direction by the judges (e.g., one judge rated the unit as an avoidance response and the other rated as an approach response). Clinical psychology staff rated the therapists' personality characteristics. The interactions between therapists and their clients were coded for the following elements: 1) the number o f times the therapists responded with approach or avoidance responses to the clients' expression o f hostility; 2) the frequency with which clients continued to express hostility immediately following therapists' approach or avoidance responses; and 3) the objects towards whom the clients expressed their hostility. Although the authors did not confirm their first two hypotheses, they did find support for their third hypothesis. In addition, they found that: 1) therapists who had a l o w need for  20  approval and who expressed their own hostility directly were more likely to encourage their clients' expression o f hostility compared with therapists who had a high need for approval and who expressed their hostility indirectly; and, 2) therapists were more likely to avoid clients' hostility when it was directed towards them versus other objects. This research also provided a useful operationalization o f countertransference as "therapist avoidance reactions" that has been adopted and expanded by others (e.g., Friedman & Gelso, 2000; Gelso, Fassinger, Gomez, & Latts, 1995; Hayes & Gelso, 1993; Rosenberger & Hayes, 1998; Yulis & Kiesler, 1968). Hayes and Gelso (1993) studied 34 male counsellors' reactions to gay and H I V positive clients using an analogue research design. The independent variables included client sexual orientation and H I V status; the dependent variable was counsellor discomfort, assessed using affective, cognitive, and behavioural measures. The affective component was defined as counsellor state anxiety using a self-report measure; the cognitive component was defined as counsellor inaccuracy in recalling client material (number o f words related to sexuality or death); and the behavioral component was defined as counsellor approach-avoidance responses based on their verbal responses to the videotaped clients. Counsellors' responses were coded as approach, avoidance, or neither by three judges. Interrater reliability for the proportion o f ratings on which pairs o f judges agreed were .67, .69, and .73. A cumulative ratio o f the number o f avoidance responses to the number o f approach and avoidance responses was calculated. In addition, counsellor's degree o f homophobia and death anxiety was assessed using self-report measures. T w o young male actors played the four client scenarios: HIV-positive/gay; HIV-negative/gay; HIV-positive/ heterosexual; and H I V negative/heterosexual. The counsellors were randomly assigned to the eight scenarios (2 actors X 4 conditions). Each video-tape had five pre-designated places where the tape stopped and the counsellor could record his or her response into a microphone. The researchers found that there was no difference between counsellors' discomfort with gay versus heterosexual clients (no main effect for sexual orientation). The interaction between H I V status and sexual orientation was also not significant. A s hypothesized, counsellors' higher ratings on the homophobia measure predicted their discomfort with gay male clients, and counsellors  21  reported greater discomfort with HIV-positive versus HIV-negative male clients. The researchers did not support their hypothesis that higher scores on the death anxiety measure would predict counsellors' discomfort with HIV-positive clients better than their discomfort with HIV-negative clients. Gelso, Fassinger, Gomez, and Latts (1995) were interested in the role of homophobia, counselor gender, and countertransference management on countertransference reactions to lesbian clients. Once again, an analog research design was employed with 68 masters and doctoral students in counselling programs observing a video tape simulation of a client and completing various self-report inventories to measure the constructs of interest. A similar procedure employed by other researchers (e.g., Hayes & Gelso, 1993; Latts & Gelso, 1995) was also used in this study. Counsellors were randomly assigned to one of the two sexual orientation scripts. Each tape contained eight segments; after each segment participants could verbally respond into a microphone as they would in an actual therapy session. Countertransference behaviour was defined as the ratio of avoidance responses to the sum of approach and avoidance behaviours in the counsellor's verbal responses. As predicted, the researchers found that the higher the level of homophobia, the greater the counsellors' avoidance response to the client's material. A large scale study on psychotherapy (The Menninger Foundation's Psychotherapy Project, cited in Singer & Luborsky, 1977), attempted a retrospective study of countertransference that had occurred within the context of actual therapy sessions. After treatment had ended, the researchers tried to reconstruct the influence of countertransference by reviewing various sources of information, such as post-treatment interviews with the client, the therapist, and supervisor, and the complete process notes. The researchers found it difficult to determine the influence of countertransference. Applying the tentative theory of countertransference origins, triggers, and manifestations outlined by Hayes et al. (1998), Rosenberger and Hayes (1998) attempted to investigate the relationships among the origins of countertransference, client verbalizations thought to trigger countertransference behaviour, manifestations of countertransference behaviour, and the effects of countertransference behaviours on the client's and therapist's perceptions of session depth and smoothness. They employed an intensive case study design,  22 following one counsellor-client dyad across 12 counselling sessions. The client and the counsellor were both white males. The client was 20 years old and the therapist was 32 years old. Rosenberger and Hayes (1998) attempted to identify the therapist's countertransference origins (e.g., unresolved intrapsychic conflicts) by having the therapist and three people who knew the therapist well complete the Adjective Check List ( A C L ; Gough & Heilbrun, cited in Rosenberger & Hayes, 1998). Countertransference origins, or "blind spots," for the therapist were identified by the scales on which therapist and cohort T scores differed by 10 points. The researchers also tried to identify countertransference origins by conducting a pre-treatment interview with the therapist to ask him "to identify and discuss themes which might tap into resolved and unresolved intrapsychic conflicts." (Rosenberger & Hayes, 1998, p. 11). A total o f 10 themes emerged from the A C L (three) and the interview (seven). Countertransference triggers were identified by having three trained raters view the therapy sessions and code the content o f each client talking turn based on the 10 countertransference origins. A n additional category called "other" was also included to code the talking turns. The inter-rater reliability for the three raters was .67, with individual pairs o f raters ranging form .61 to .76. (The researchers did not specify whether these numbers indicate reliability coefficients or percent agreement for raters). Rosenberger and Hayes (1998) identified manifestations o f countertransference behaviour in the following manner: first, countertransference behaviours were assessed for every speaking turn by rating approach and avoidance responses. This technique was similar to the one developed by Bandura et al. (1960) and used in previous analogue research on countertransference. Bandura et al. (1960) defined avoidance behaviours as "those verbal responses designed to inhibit, discourage, or divert the patient's hostile expressions" (p. 3). Raters were presented with a grid, with the left column containing both approach reactions (approval, exploration, instigation, reflection, and labelling) and avoidance reactions (disapproval, topical transitions, silence, ignoring, and mislabelling), as well as a separate category for "unclassifiable responses" which represented all speaking turns which were not classified as approach or avoidance. Rater reliability for the three 10 minute training videos  23  was .97 for the three judges. Inter-rater agreement for the case study was not reported. The authors also identified countertransference behaviours by session using the Inventory o f Countertransference Behaviour (ICB; Friedman & Gelso, 2000). The I C B provides a global score o f the therapist's countertransference behaviour, operationalized as therapist over-involvement and under-involvement, for the therapy session. A separate team o f judges independently rated the 12 counselling sessions on the I C B . Rosenberger and Hayes (1998) tested three hypotheses. Their first hypothesis was that client speaking turns containing conflictual material would tend to be followed by therapist speaking turns containing avoidance behaviour. They employed a sequential analysis technique to calculate the conditional probability o f the therapist's avoidance response immediately following the client's speaking turn containing conflictual material. The results did not support the authors' hypothesis either across or within the counselling sessions. Their second hypothesis was that there would be a positive relationship between the frequency with which the client talked about issues that were conflict-relevant for the therapist and the frequency o f countertransference behaviour. This hypothesis was assessed in two ways. One, the total number o f client speaking turns containing conflictual material was correlated with the total number o f therapist speaking turns containing avoidant responses; and two, Pearson correlation coefficients were calculated between scores on the Inventory o f Countertransference Behaviour and number o f conflictual client speaking turns for each session. B o t h attempts to test this hypothesis were not supported. There was a trend towards a positive relationship between the Negative Countertransference factor on the I C B and the number o f conflictual client speaking turns. Rosenberger and Hayes' (1998) third hypothesis tested whether sessions with higher frequencies o f therapist countertransference behaviour would be rated less favourably by the client and therapist dyad than would sessions with lower frequencies o f countertransference behaviour. Again, this hypothesis was tested in two ways. One, Pearson correlation coefficients were calculated between the number o f avoidant therapist speaking turns and the dyad's ratings on the Session Evaluation Questionnaire ( S E Q ; Stiles & Snow, cited in Rosenberger & Hayes, 1998); and two, Pearson correlation coefficients between the I C B and client and therapist ratings on the S E Q were calculated. The first test o f this hypothesis was  not supported. Surprisingly, it was found that the more the therapist engaged in avoidance behaviour, the deeper he perceived the session to be. A l l other relationships regarding this hypothesis did not reach statistical significance, but were in the expected direction. The second test o f the authors' hypothesis received some support. They found an inverse relationship between the I C B ' s Negative Countertransference factor, and the therapist's perceptions o f the depth and smoothness o f the sessions. N o other significant relationships were found (e.g. Positive Countertransference factor and Total Countertransference). Recently, Rosenberger and Hayes (2002) replicated the research methodology described above with another counsellor-client dyad across 13 therapy sessions. The counsellor was a 34 year old, White female psychologist. The client was a 21 year old, White female. The authors tested the same three hypotheses: first, that client speaking turns containing conflictual material would tend to be followed by counsellor speaking turns containing avoidance behaviour; second, that there would be a positive relationship between the frequency with which the client talked about issues that were conflict-relevant for the counsellor and the frequency o f countertransference behaviour; and third, that sessions with higher frequencies o f counsellor countertransference behaviour would be rated less favourably by the client and counsellor dyad than would sessions with lower frequencies o f countertransference behaviour. The authors did not support their first and second hypotheses. In fact, the test for the second hypothesis was in the opposite direction than expected; the counsellor's avoidance behaviour was inversely related to the frequency with which the client addressed conflictrelevant material. The third hypothesis received partial support. The counsellor's ratings o f session depth was inversely related to the frequency with which the client addressed conflictrelevant material that the counsellor was both unaware (r = -.31) and aware (r = -.48). There was also a positive relationship between session smoothness and frequency with which the client addressed conflict-relevant material that the counsellor was unaware (r = .51). The client's ratings o f these dimensions (e.g., session depth and smoothness) were unrelated to the counsellor's avoidance behaviour. Rosenberger and Hayes (2002) also explored the impact o f the client addressing material identified as conflictual for the counsellor on the counsellor's ratings o f the working  alliance and the counsellor's social influence. The impact o f the counsellor's avoidance behaviour on the client's ratings o f the working alliance and the counsellor's social influence were also explored. A positive relationship was found between the counsellor's ratings o f the working alliance and the frequency with which the client addressed conflict-relevant material that the counsellor was aware (r = .47). N o other significant relationships were found for the counsellor or the client regarding the working alliance. In terms o f the counsellor's social influence, there was a negative correlation between the frequency with which the client addressed conflict-relevant material that the counsellor was aware and the counsellor's ratings o f expertness (r = -.50) and attractiveness (r = -.36). There was also a negative relationship between conflictual material that the counsellor was unaware and her ratings o f trustworthiness (r = -.52). N o other significant relationships were found for the counsellor regarding this variable. Due to the lack o f variability in the client's ratings, no correlations could be calculated to explore the impact o f the counsellor's avoidance behaviour on the client's ratings o f the counsellor's social influence. The research by Rosenberger and Hayes (1998; 2002) was innovative and served to highlight the challenges o f studying countertransference in a naturalistic setting. It may have been ambitious to attempt to predict countertransference behaviour from predetermined countertransference origins, given the difficulty at ascertaining valid measures o f "therapists' unresolved intrapsychic conflicts." Also, conducting a pre-treatment interview with the therapist to identify countertransference origins could have unknowingly biased the results. T o avoid some o f these methodological challenges, Hayes et al. (1998) suggested "working backwards" to try to identify countertransference manifestations first, then triggers, and then origins. Countertransference as over-involvement and under-involvement. The Inventory o f Countertransference Behaviour ( I C B ; Friedman & Gelso, 2000) was recently developed to measure two dimensions o f countertransference: therapist over-involvement and underinvolvement. This conceptualization o f countertransference is consistent with Wilson and Lindy's (1994) model o f countertransference described earlier (e.g., describes countertransference as either the therapist's over-identification with or avoidance o f the client's material). The I C B originally contained 32 items on which supervisors or judges could  26 rate the extent to which a therapist-trainee's behaviour in a session reflected specific manifestations o f countertransference. Respondents rated the trainee's reaction to a particular client in a given session on a five-point Likert type scale, where 1 = to a little or no extent and 5 = to a great extent. The instrument was designed with a stem ("The counsellor") followed by 32 stem endings. The items were hypothesized to represent behaviours that reflected both over-involvement and under-involvement by the counsellor. Exploratory factor analysis was conducted as part o f the development process. This analysis suggested that the two subscales that emerged from the analysis were not indicative o f the hypothesized dimensions, but rather described: 1) inappropriate therapist behaviours that were disapproving o f clients (titled Negative Countertransference - 1 1 items), and 2) inappropriate therapist behaviours that were overly familiar or supportive (titled Positive Countertransference - 1 0 items). A total score was calculated for the I C B , with higher totals indicating increased levels o f countertransference behaviour. Thus, the I C B provided a global score o f countertransference behaviour for the therapy session. Individual scale scores can also be calculated. Counsellor over-involvement, or positive countertransference, included items such as: seemed to agree too often with the client, over-supported the client, befriended the client, frequently changed the topic, talked too much, acted in a submissive way. Counsellor underinvolvement, or negative countertransference, included items such as: treated client in a punitive manner, was critical o f the client, spent time complaining, provided too much structure, inappropriately questioned the client's motives. The authors acknowledged that operationalizing countertransference behaviour as therapist over-involvement and under-involvement (which they renamed positive and negative countertransference after factor analysis) did not capture the complete complexity o f the phenomena. However, it did attempt to include overt behaviours that were more likely to be observed when assessing actual counselling sessions. T o date, Rosenberger and Hayes' (1998) case study described above was the only research that has employed the I C B as a global measure o f countertransference. However, they relied on the "avoidance" definition o f countertransference when analyzing the therapist's behaviour more closely within sessions. Countertransference as deviations from baseline. A recent study investigated patterns  o f consistency and deviation in therapists' feelings across clients (Holmqvist, 2001). Nine  27  therapists and 28 clients participated in the study. The number o f clients a therapists worked with ranged from one to six. Therapy sessions ranged from nine to 88. The therapists reported their feelings after each therapy session using a feeling checklist. The author attempted to identify four deviating reactions in the therapists: 1) therapist-characteristic countertransference (therapists' habitual feelings that differed from other therapists' habitual feelings); 2) patient-characteristic countertransference (therapists' feelings over a whole therapy with one patient that differed compared to their feelings over a whole therapy with other patients; 3) session-characteristic countertransference with respect to the therapist (therapists' feelings in individual sessions that differed from their usual response style across patients; and 4) session-characteristic countertransference with respect to the client (therapists' feelings in individual sessions that differed from their usual response style with that specific patient). The author used analysis o f variance to investigate differences in reactions between therapists generally, and between feelings toward individual patients. Discriminant analysis was used to discriminate the therapists by the feelings they reported, and to discriminate the therapies for individual clients by the feelings that were reported from them. The results suggested that therapists were consistent in their feeling style across time. In other words, which therapist had produced which feeling checklist could be predicted accurately most o f the time. The therapists' reactions were less consistent towards individual patients. The therapists' deviations from their usual responses were thought to be possible examples o f countertransference reactions. Holmqvist (2001) stated: deviating tendencies or deviating reactions in individual sessions could be regarded as indicators o f important processes in the therapy. The statistical method does not give any information about the reason for the deviating reactions. What the method does achieve is to point out those reactions that deviate from the therapist's normal reaction pattern. The discrirninant analysis does this without relying on the therapist to report that the reaction is unusual. In this way, it was possible to bypass one o f the problems in studying countertransference. The therapist's consciousness o f countertransference reactions as  28  clinically described may vary, and methods o f mapping countertransference that presuppose that the therapist reports the reaction as unusual may consequently be less apt. With the method presented here, this question is irrelevant, (p. 114) Because the data violated various assumptions for the statistical methods used in this study (e.g., small sample size, significant violations o f equality o f variances), the results should be interpreted with caution and require replication. What was noteworthy about this research was that the manner in which countertransference was conceptualized introduced interesting possibilities for future research. Summary To summarize, this section attempted to review some o f the clinical, theoretical, and empirical research on countertransference. Although the clinical literature makes an important contribution to understanding the phenomenon, a major weakness o f the literature on countertransference is the abundance o f anecdotal reports and the paucity o f empirical research. The empirical research on countertransference that does exist contains various limitations. Overall, the lack o f research in naturalistic settings and the over-reliance on analogue research designs are a major weakness o f the empirical research on countertransference. Based on the studies reviewed, there has been an over-reliance on selfreport inventories with questionable validity in operationalizing countertransference. In order to have confidence in research results, one requires confidence in how the concepts have been operationalized (construct validity), as well as confidence in the methods used to study the phenomena. Gelso and Hayes (2002a) succinctly captured this struggle: The likely culprits for the slow pace o f research were twofold. First, countertransference originated from and was firmly embedded in psychoanalysis, a discipline containing a decidedly anti-empirical bent and an opposition to the simplification that appears to be an inherent part o f scientific research. Second, and perhaps more telling, the construct itself is awesomely complex, focusing as it does on unconscious processes, defense mechanisms, and indeed often one person's unconscious reactions to another person's unconscious reactions. A d d to these the definitional ambiguity that seems to be a part o f virtually all high-level constructs, and  29  the road was paved for little research, (p. 3) The authors note that the construct has generated much interest in recent years in terms o f its impact on the therapy relationship. Countertransference is no longer a topic that is only o f interest to psychoanalysts. A s stated earlier, this researcher subscribes to a "combined definition" o f countertransference. From this perspective, countertransference is viewed as both positive and negative. Countertransference reactions, i f understood, can facilitate empathic understanding; however, i f unchecked, these reactions can interfere with the therapeutic process. In other words, there is a distinction between countertransference feelings and behaviour. Because the latter instances tend to be more problematic, this study focused on behavioural manifestations o f countertransference. Thus, regardless o f whether the therapist's countertransference reactions stemmed from his or her unresolved issues from past significant relationships (e.g., subjective countertransference) or from the client's issues and/or behaviour (e.g., objective countertransference), this study focused on observable behavioural manifestations o f countertransference during therapy sessions. Restatement o f Purpose To date there have been few attempts to observe countertransference behaviour during actual counselling sessions (e.g., Bandura et. al, 1960; Rosenberger & Hayes, 1998). Recent theoretical conceptualizations o f countertransference have provided helpful frameworks to investigate the construct further (e.g., Hayes et al., 1998; Wilson & Lindy, 1994). In their preliminary theory o f countertransference origins, triggers, and manifestations, Hayes et al. (1998) recommended that researchers "work backwards" and first try to identify countertransference manifestations (thoughts, feelings, and behaviours), then try to identify potential triggers and origins for the therapist. Because the purpose o f this research was to identify possible countertransference manifestations during actual counselling sessions, selecting criteria that were potentially observable was desirable, particularly when some have stated that countertransference manifestations are more likely to be observed by third parties (e.g., Singer & Luborsky, 1977). Thus, operationalizing countertransference behavioural manifestations as therapist overinvolvement and under-involvement not only seemed heuristically sound, but was thought to  30  have both clinical and empirical value (Friedman & Gelso, 2000). Previous research focused solely on therapists' avoidance behaviour (e.g., Rosenberger & Hayes, 1998) and overlooked negative aspects o f therapists' seemingly facultative behaviours, such as over-supporting or colluding with clients. This broader conceptualization provided some flexibility in the range o f potential therapist behaviours. The purpose o f this exploratory research was to address some o f the limits o f previous research by designing a study that attempted to identify behavioural manifestations o f therapist countertransference reactions during actual counselling sessions. Specifically, this study addressed the following research questions: 1.  Can countertransference behavioural manifestations, defined as therapist overinvolvement and under-involvement, be reliably identified during counselling sessions with clients?  2.  Is there evidence to support the contention that counsellor over-involvement and under-involvement are valid indicators o f countertransference behaviour? A s will be described in the method chapter, in this study, judges' agreement  constituted a reliable rating o f over-involvement and under-involvement. In addition, Kiesler's (2001) suggestion to identify countertransference behaviour as therapist behaviour that "deviates from baseline" provided a guideline on selecting significant episodes o f overinvolvement and under-involvement based on each counsellors' typical response style. Evidence that over-involvement and under-involvement were valid indicators o f countertransference behaviour was investigated by triangulating multiple data sources to identify potential triggers in the dialogue preceding the behaviour (e.g., content o f client dialogue, counsellor session notes, supervision notes, counsellor and supervisors' ratings). Hayes, McCracken, McClanahan, H i l l , Harp, and Carozzoni (1998) hypothesized that countertransference triggers, defined as therapy events that elicit a reaction in the therapist, preceded countertransference behaviour. The more evidence that the preceding client dialogue contained triggers for the therapist, the more confidence one could have that over and underinvolvement was an indicator o f countertransference.  Chapter III Method This research sought to address two questions: one, can countertransference behavioural manifestations, conceptualized as counsellor over-involvement and underinvolvement, be reliably identified by independent judges in a naturalistic setting; and, two, is there evidence to support the contention that counsellor over-involvement and underinvolvement were valid indicators o f countertransference behaviour? These research questions can be phrased more informally as, "can judges reliably identify counsellors' departures from empathic connection, defined as counsellor over-involvement and underinvolvement" and "what are the triggers associated with counsellors' empathic failures?" Because o f the complex nature o f this study, the initial section o f this chapter outlines the methods used to answer the research questions under the following five main headings: step 1: design and use o f Generalizability Theory; step 2: recruitment o f participants; step 3: treatment implementation and data collection; step 4: Decision (D) study: Question one; and, step 5: Qualitative (Q) study: Question two. A diagram illustrating the multiple and progressive steps o f the method, and their relationship to each other, appears in Figure 2. Step 1: Design In order to study potential countertransference manifestations systematically, this study employed a mixed research design. This section briefly describes the rationale for applying Generalizability Theory ( G T ) and a Generalizability, or G study, with a multiple case study design. The decision and qualitative studies, henceforth referred to as the D and Q studies, are also described briefly to orient the reader. Further details regarding measurement o f the research questions are provided below (e.g., steps 4 and 5). Multiple Case Study Research A multiple case study approach was used to intensely analyze the intimate interaction that occurs between clients and their therapists. This approach was selected because it allowed a level o f analysis that could yield insights concerning the therapeutic process that may have been overlooked by other research designs (e.g., Jones, 1993; Y i n 1989). Single and multiple-case research designs have various strengths. Heppner et al. (1992) identified the following advantages: first, they are a means o f collecting information  32  Figure 2: Overview of Methodology  Figure 2  How do we measure countertransference behavioural manifestations in a naturalistic setting? Identification of Research Questions (RQ)  RQ1: Can CT behavioural manifestations, operationalized as therapist over and under-invotvement, be reliably identified during counselling sessions?  RQ2: Is there evidence to support the contention that counsellor over and under-  involvement are valid indicators of CT behaviour?  r  What is required for reliable measurement?  What evidence is required to support validity?  \  Generalizability (G) Study  Reliable measurement (D Study)  Generalizability Theory: conduct a G study to assess the dependability of the behavioural measure.  And  Use G Study specifications (number of judges, counsellor/client dyads and sessions) to design and conduct the Decision (D) study.  Qualitative (Q) Study Use multiple sources of data in the design of the study to establish convergence of evidence.  Decision (D) Study D study answers RQ1: measurement is reliable (intraclass correlations, descriptive statistics)  Identification of over/underinvolvement episodes during counselling sessions (using moving averages graphs) Qualitative Analysis of multiple data sources answers RQ2  Combine results of RQ1 and RQ2  33  and ideas, and generating hypotheses about the therapeutic process; second, they are a means o f testing therapeutic techniques and o f testing new methodologies; and third, they are a means o f studying individuals, rare phenomena, and o f providing exemplars and counterinstances. In addition, single-case designs are versatile and can be employed creatively with a variety o f phenomena (Galassi & Gersh, 1993). M o r e recently, there has been more attention given to intensive single-case designs which incorporate repeated measures and adopt a systematic approach (e.g., Hilliard, 1993). This study employed video tapes o f actual counselling sessions as the basis for the ratings o f counsellors' countertransference behaviour. Because research question two could not be addressed unless research question one could be answered with confidence, this study adopted methodology for the first question that would maximize the reliability o f measuring countertransference behaviour, defined as over-involvement and under-involvement. This involved application o f Generalizability Theory ( G T ; Cronbach, Gleser, Nanda, & Rajartnam, cited in Shavelson, Webb, & Rowley, 1989) described below, and conducting a preliminary study (the G study) to assess the dependability o f the behavioural measure o f countertransference and design the eventual investigation, termed the decision (D) study. In other words, the G study determined the necessary and sufficient number o f counsellor-client dyads, counselling sessions, and judges required to achieve a desirable level o f reliability in the D study. Once the primary investigation o f research question one was completed, the study could then proceed to research question two. Research question two required a different form o f case study. In addition to the video taped counselling sessions for the counsellor/client dyads, multiple information sources were used to collect data. These data were then analyzed qualitatively in the Q Study. Generalizability Theory To address the dependability o f the behavioural measure o f countertransference, generalizability theory (GT; Cronbach, Gleser, Nanda, & Rajartnam, cited in Shavelson, Webb, & Rowley, 1989), a statistical theory about the dependability o f behavioural measurements, was used to evaluate the reliability o f the measurement. Although multiple case study research frowns upon applying sampling logic (i.e., a smaller number o f  34  participants or events are thought to be a representative sample collected from the entire pool of participants or events) to case studies ( Y i n , 1994), it seems reasonable to query the dependability o f the behavioural measure and to question whether events (i.e., therapist overinvolvement and under-involvement) across time for a single therapist generalize to future events for that therapist. Because case study research has generally been criticized for lacking external validity, employing this mixed research design helped to protect against threats to external validity. Generalizability theory ( G T ) has proved useful in gauging the dependability o f behavioural measurements by assessing the multiple sources o f measurement error and attempting to reduce their effects. Other researchers interested in assessing the dependability o f their measures have applied generalizability theory to a variety o f problems, in a variety o f settings (e.g., Erlich & Shavelson, 1976). In G T , dependability describes the accuracy o f generalizing from a person's observed score or rating on a measure (e.g., behavioural observation o f countertransference manifestation) to his or her average score or rating across observations (Shavelson & Webb, 1991). Thus, a single rating, across one occasion, with one judge is not a fully dependable measure o f the behaviour. A central component o f G T is that it differentiates between generalizability studies ( G studies) and decision studies ( D studies): " G studies estimate the magnitude o f as many potential sources o f measurement error as possible. D studies use information from a G study to design a measurement that minimizes error for a particular purpose." (Shavelson, Webb, & Rowley, 1989, p.923). Consequently, applying the principles o f G T increases the reliability o f the generalizations researchers can make from their findings obtained in a D study, as is the intent here. The generalizability study. T w o female, doctoral level counsellors from the Department o f Counselling Psychology at the University o f British Columbia and two o f their clients participated in the G study. The counsellors were in their first year o f doctoral study and were enrolled in a full day clinic practicum at a training facility in N e w Westminster. The clinic offers free counselling to members o f the community by Master's and Doctoral level student counsellors. Clients at the clinic sign consent forms permitting their sessions to be videotaped and observed by the clinic team for supervision purposes. The two counsellor-client dyads who allowed their video-taped counselling sessions to be  35 released for this G study completed additional informed consent documents and were provided with a written description o f the purpose o f the study (see Appendix A ) . Counsellor-Client dyads. Counsellor A was a White female in her middle to late thirties. She described her theoretical orientation as client-centred. H e r client (client A ) was a White female in her late thirties to early forties. Client A presented with relationship and parenting issues. She attended individual and joint sessions with her boyfriend and teenage daughter. Only sessions in which client A attended alone were included in the G study. The counselling sessions were conducted from September 1997 to March 1998 at the N e w Westminster Counselling Centre. The first and third sessions were analyzed in the G study. Counsellor B was a White female in her early thirties. She described her theoretical orientation as client-centred. H e r client (client B ) was a White male in his early fifties. Client B presented with depression, career issues, and relationship difficulties. H e attended individual sessions. The counselling sessions were conducted from September 1998 to March 1999 at the N e w Westminster Counselling Centre. The third and fourth sessions were analyzed in the G study. The first session o f counsellor-client dyad B was used for training purposes. Judges. T w o judges participated in the G study. Both judges were White females in their early thirties. Judge A the author, was a doctoral student in the department o f Counselling Psychology at the University o f British Columbia. She was in her final year o f doctoral work. Her theoretical orientation was influenced by existential and psychodynamic theories. Judge B recently completed her final requirements for her Ph.D. in the department o f Clinical Psychology at the University o f British Columbia. Her theoretical orientation was primarily cognitive-behavioural. A third judge, Judge C , participated i n the G study training as a "back up." Judge C graduated with her Master's degree in Counselling Psychology from the University o f British Columbia in 1995. Her theoretical orientation was influenced by client-centred and cognitive theories. She has worked as a substance abuse counsellor since completing her degree. Data preparation. Four counselling sessions (two from dyad A and two from dyad B ) were transcribed verbatim for the G study. A n additional session from dyad B was also transcribed for training purposes. In the transcripts, each statement made by the counsellor and client was numbered according to talking turns. For example, the clients' first statements  36  were numbered l a and the counsellors' first statements, or responses, were numbered l b . This numbering system was used throughout the transcript (e.g., 2a, 2b, 3a, 3b, etcetera). Rating countertransference behavioural manifestations. Countertransference behavioural manifestations, operationalized as therapist over-involvement and underinvolvement, were assessed by three trained judges observing video-tapes o f counselling sessions, paired with verbatim transcripts, and rating each o f the responses made by the counsellors using a 7-point Likert scale (i.e., +3 = over-involved, +2 = somewhat overinvolved, +1 = possibly over-involved, 0 = empathic connection, -1 = possibly underinvolved, -2 = somewhat under-involved, and -3 = under-involved). The judges' agreement for over or under-involvement constituted a rating o f countertransference behaviour. Bandura et al. (1960) originally developed a similar method o f coding every counsellor speaking turn for countertransference behaviour and operationalized countertransference as avoidance behaviour. Their method and definition has been adopted by others (e.g., Hayes & Gelso, 1993; Rosenberger & Hayes, 1998). The G study applied the same method o f rating each counsellor response, but expanded the operationalization o f countertransference to include both over-involved and under-involved responses. B y focusing solely on avoidance behaviour, previous research has overlooked negative aspects o f therapist's seemingly positive behaviour, such as over-supporting and befriending the client. In order to capture this broader conceptualization o f countertransference, descriptors for the dimensions over-involvement and under-involvement were taken from the Inventory o f Countertransference Behaviour ( I C B ; Friedman & Gelso, 2000). Counsellor overinvolvement included items such as: seemed to agree too often with the client, oversupported the client, colluded with the client, frequently changed the topic, talked too much, acted in a submissive way. Counsellor under-involvement included items such as: treated client in a punitive manner, was critical o f the client, spent time complaining, provided too much structure, inappropriately questioned the client's motives. Although Friedman and Gelso (2000) renamed the scales o f the I C B , "Negative Countertransference" and "Positive Countertransference," for this research, the previous labels over-involvement and underinvolvement were retained. The original labels were deemed easier to grasp conceptually for the purpose o f training judges, and were consistent with others' conceptualization o f  37  countertransference (e.g., Wilson & Lindy, 1994). A third dimension, empathic connection, was added to provide definitions o f appropriate counsellor behaviour to help judges differentiate the phenomenon under study. Counsellor empathic involvement included items such as: used facilitating skills such as reflection, summaries, clarification, labelling, and empathy to demonstrate an understanding o f the client's experience and willingness to explore issues further, was warm and caring, appeared genuine in his/her interactions with the client. Although the I C B is typically used to provide a global score o f countertransference behaviour for the session, the current research was interested in capturing moments o f countertransference behaviour within therapy sessions. Thus, the I C B descriptors were employed to assist judges in their ratings o f each therapist response during the session. In other words, rather than providing a global rating o f the therapists' behaviour for each session, the judges rated each therapist response for its level o f countertransference behaviour using the I C B descriptors and empathy descriptors. A manual describing these dimensions was developed for the purpose o f training the judges in the G study (see Appendix B ) . While watching the videotapes o f the counselling sessions and reading the corresponding transcripts, the trained judges used the seven point Likert scale to rate each counsellor response according to his or her level o f empathic involvement or over/underinvolvement. They recorded their ratings using the rating sheet in Appendix B . Training procedures. The author, Judge A reviewed the training tape (dyad B ' s first session) along with the corresponding transcript. Ratings for each counsellor response were completed using the manual in Appendix B . The counsellor's responses were rated according to her level o f over-involvement, under-involvement, and empathic involvement using a seven point Likert scale. The scale ranged from positive three at one end (over-involvement), zero in the middle (empathic involvement), and negative three at the other end (underinvolvement). Once Judge A completed the ratings using the training tape, the training procedures were repeated with Judge B using the training tape, transcript, and manual. Judge B was directed to read the instructions in the manual and to review the descriptors for counsellor over-involvement, under-involvement, and empathic involvement. A n example o f each counsellor behaviour, over-involvement, under-involvement, and empathic involvement, was shown from the training tape based on Judge A ' s initial ratings o f  the training tape. Judge B , blind to the ratings, was directed to rate the three examples. Judge A and B were in agreement regarding the ratings for the three examples o f counsellor behaviour. Judge A and Judge B then proceeded to review the training tape, stopping at five minute intervals so Judge B could complete her ratings for that segment. Ratings between Judge A and B were compared after each five minute interval before proceeding to the next segment. Discrepancies in ratings o f two points or more (e.g., Judge A rated the counsellor's response 0 and Judge B rated it +2) were discussed to help refine and clarify ratings. A n attempt was made to come to consensus for those ratings (approximately 10 % o f total number o f ratings). U p o n completing the ratings o f the training tape, Judges A and B proceeded to rate separately the two sessions from dyad A and the two sessions from dyad B . Intra-class correlations were computed for the judges' ratings. The judges were trained to an agreement level o f .82. A third judge, Judge C , who later served as a rater for the D study, participated in a portion o f the G study training. Although her data was not used for the G study, her training is described here. The training session with Judge C was less rigourous compared with the training o f Judge B . Judge C was also informed to read the instructions in the manual and to review the descriptors for counsellor over-involvement, under-involvement, and empathic involvement. A n example o f each counsellor behaviour, over-involvement, underinvolvement, and empathic involvement, was shown from the training tape. Judge C was encouraged to ask questions regarding the three dimensions she would be rating. A t this point Judge C was requested to review the training tape and transcript and to record her ratings at five minute intervals. Unlike Judge B ' s training which involved comparing ratings at the end o f every five minute segment, Judge A and C only compared and discussed ratings up to the first 15 minutes o f the session. Judge C then proceeded to rate the remainder o f the training tape on her own. She also completed the ratings for the two sessions from dyad A and the two sessions from dyad B . Because this G study provided an opportunity to clarify both methodological and procedural issues, feedback from the judges concerning the procedures were elicited throughout the training. Suggestions from the judges included: 1) it was helpful to read through the transcripts first before watching the video taped sessions, 2) it was easier to  follow i f the video was stopped after each two pages o f transcript, and 3) it was easier to write ratings on the transcript right next to the talking turn and transfer them to the rating sheet afterwards. G Study Results. The G study helped to design the current, or decision study (D study) by providing a statistical means o f estimating the dependability o f the behavioural measurement o f countertransference manifestations for various combinations o f judges, sessions, etcetera. Using information from the G study made it possible to reduce the amount o f measurement error in the D study. This task was accomplished by computing variance components for each parameter o f interest in the study (i.e., judge, sessions, therapists, interactions, error). G theory helps the researcher assess the major sources o f variation so that unwanted variation can be reduced in collecting future data ( D study) (Shavelson & Webb, 1991). The object o f measurement in the G study was therapist behaviour, denoted as therapist (t). The facets were sessions (s) and judges (j). This G study is described as a partially nested design because it had both crossed and nested effects: sessions were nested within therapists because there were two sessions per therapist and the sessions differed for both therapists. The judge facet was crossed with both sessions and therapists because each judge rated all the sessions for both therapists. With nested designs fewer variance components can be estimated separately (Shavelson & Webb, 1991). F o r example, because sessions are nested within therapists, it is impossible to separate the session main effect from the interaction between therapists and sessions. A table containing the estimated variance components from the analysis o f the two counsellors (e.g., percentage o f total variance attributed by dyad, judge, and session) is contained in Appendix A . Table 1 below provides the estimated error variances based on the G study used in the present case and alternative D-study designs. It includes the variance estimates and generalizability coefficients (dp) for several combinations o f facets (i.e., number o f judges and sessions). The generalizability coefficient is comparable to the reliability coefficient (i.e., an intraclass correlation coefficient) in classical theory, which represents true score variance divided by expected observed-score variance (Shavelson & Webb, 1991). The formulas used to calculate the G coefficient are listed in Appendix A .  40  Table 1 Decision Study for Therapist Countertransference Behaviour Observations [(s:t)xj Design]  G Study  Alternative D Studies  Source of  n'j =  1  2  2  2  3  3  3  Variation  n s=  1  2  5  10  6  8  10  .194  .194  .194  .194  .194  .00002  .00002  .00001  .00001  .00001  .0035  .0009  Therapist (t)  .194  .194  .00003 .00002  Judge (j) Session:Therapist(s:t) Therapist* Judge(tj)  .028  .014  .0056  .0028  .0047  0  0  0  0  0  0  0  Error(s:tj,e)  1.186  .2965  .1186  .0593  .066  .049  .0395  oRel  1.214  .3105  .1242  .0621  .0707  .0525  .0404  o Abs  1.214  .3105  .1242  .0621  .0707  .0525  .0404  p2  .14  .38  .61  .76  .73  .79  .83  .14  .38  .61  .76  .73  .79  .83  2  Using one rater and one session to measure therapists' countertransference behaviour would yield l o w generalizability and phi coefficients  (.14).  Because the variability due to the  residual term was substantial, it was necessary to increase the number o f sessions and judges to reduce this source o f unwanted variance. Given the time required to train judges, the decision to increase sessions more substantially than judges was made for practical purposes. In order to yield a generalizability coefficient o f .79, the decision study included three judges, all rating the two counsellor-client dyads across 8 therapy sessions. A generalizability coefficient o f .79 was within the range considered acceptable when assessing the dependability o f a behavioural measure and was consistent with what other researchers have accepted when designing their D study (e.g., Shavelson, Webb, & Rowley, 1989). Research Design Summary To summarize, a mixed research design was used in this study. Multiple case research and Generalizability theory (e.g., G study) were integrated to design a study that intensely analyzed the interactions between clients and their counsellors and maximized the reliability and validity o f measuring countertransference. The D study investigated research question  41  one and the Q study explored research question two. The D study: Research question one. A s noted above, the results o f the G study helped to design the decision (D) study in a manner that increased the dependability o f the behavioural measure o f countertransference. The D study implemented the specifications from the G study to address the first research question. This question explored whether countertransference behavioural manifestations could be reliably identified during therapy sessions. Manifestations were operationalized as therapist departures from empathic connection or, more specifically, as over and under-involvement. This variable was assessed by three trained judges, who observed two counsellor-client dyads across sixteen counselling sessions (eight sessions each). The counsellors' behaviour was rated based on the same 7point Likert scale employed in the G study. The judges' agreement constituted a rating o f countertransference behaviour. The O study: Research question two. The Q study addressed the second research question. This question explored whether there was evidence that counsellor over and underinvolvement were valid indicators o f countertransference behaviour. T o apply measurement theory logic to this research question, construct validity was investigated by identifying links between variables hypothesized to be related (Anastasi, 1988). Countertransference triggers, defined as therapy events that elicited a reaction in the therapist, are hypothesized as preceding countertransference behaviour (Hayes et al., 1998). Evidence supporting the validity o f conceptualizing countertransference behaviour as over-involvement and underinvolvement was accumulated by establishing a temporal relationship between countertransference triggers and behaviour. Potential triggers for the therapist were identified by reviewing various sources o f information: 1) the content o f the client's dialogue and emotional tone during the therapy sessions; 2) the counsellors' session notes; 3) the research supervision notes from two supervision sessions; and 4) ratings by the counsellors, the research supervisor, and the counsellors' previous supervisor regarding the counsellors' reactivity to various client issues. Further descriptions o f the D and Q studies appear in subsequent sections o f step 4 and 5. Step 2: Recruitment o f Participants This study was promoted as a counsellor development project. Participants in this study included two counsellors, two clients, one clinical research supervisor, and three  42  judges. Descriptions of the recruitment procedures and participant demographics are provided in this section. Counsellor Recruitment Counsellors were invited to participate in this study to gain paid, supervised experience working with clients who have serious health concerns. They were recruited from three sources: 1) the pool of current and graduated Master's level counselling students from the Department of Counselling Psychology at the University of British Columbia; 2) the pool of graduated studentsfroma two year counsellor training program (CURA Institute for Integrated Learning) in Vancouver, British Columbia; and 3) the pool of established counsellors in the local community. Information packages regarding the study were posted within the Department of Counselling Psychology at the University of British Columbia asking counsellors to contact the primary researcher directly for an information package (see Appendix C). In addition, information packages were given to the clinical director of CURA Institute for Integrated Learning and to several psychologists conducting supervision with counsellors in the community. Again, counsellors interested in participating were asked to contact the primary researcher directly. Four counsellorsfromthe community and two counsellorsfromCURA contacted the researcher. There were several requirements for inclusion in the study. The counsellors had to have either completed a Master's degree in counselling psychology or be enrolled in a Master's program. If they were currently enrolled in a program, potential participants had to have completed all their clinical requirements. In addition, the counsellors had to have undergone supervised counselling experience as part of their program training and to have a recent clinical supervisor who would be willing to complete an evaluation of the counsellor. To participate, counsellors had to be available during the summer months when the study was being conducted. Of the six counsellors who responded, three met the above criteria.  1  A meeting was arranged with therapists who satisfied the inclusion criteria to complete the following tasks:first,a general statement regarding the nature the study without disclosing the construct of interest was provided (see Appendix C); second, the time commitment and the therapist's responsibilities in the study were explained; and third, if acceptable to the therapist and the researcher, the therapist signed an informed consent form  and completed the demographic questionnaire (see Appendix D ) . The therapists were asked to contact a previous supervisor to inform him or her o f their involvement in the study and to ask him or her to complete a brief evaluation o f the therapist. They were given a package containing a general description o f the study and the evaluation forms to be mailed to their previous supervisor (see Appendices C , E , and F ) . The packages also contained a selfaddressed envelope for the supervisors to return the forms to the researcher. The therapists were also given the name and phone number o f the research supervisor for the study and were asked to make initial contact with her before starting the therapy sessions. During this meeting, therapists were given a tour o f the counselling rooms in the Faculty o f Education at the University o f British Columbia and shown how to operate the audio-visual equipment. They were also given the following materials: 1) eight blank video and audio tapes, 2) eight session note forms (to be completed by the therapist after each therapy session), and 3) two copies o f the B r i e f Symptom Inventory (BSI; Derogatis, 1993) (to be administered to the client after the fourth session and again after the final session). Counsellors' demographics. Counsellor One was a 49 year old, White female. She was married with no children. She identified her ethnic background as Canadian European. She had a Master's degree in counselling psychology from the University o f British Columbia, with four years o f post-degree counselling experience. She described her theoretical framework as eclectic, and as influenced by Feminist, Existential, Narrative, Family Systems, and Psychodynamic theories. She had the most experience with the following client issues and populations: violence and abuse in relationships; women's issues (e.g., eating disorders, depression), cross-cultural, career, mid-life issues, and life transitions (e.g., divorce, illness, death). Counsellor One had previously worked as a writer before training as a counsellor. Counsellor T w o was a 55 year old, White male. H e was divorced with no children. He identified his ethnic background as British. H e was in the process o f completing his Master's degree in mental health counselling from City University. H e had also completed a two year intensive counsellor training program through C U R A Institute for Integrated Learning. Including pre-university and university training, Counsellor T w o reported eight years o f pre-degree counselling experience. H e described his theoretical orientation as eclectic, and as influenced by Client-centred, B r i e f Solution- Focused, and Narrative  44  approaches. H e had the most experience with adults and youth and had completed an internship at a mental health clinic working with clients exhibiting a variety o f psychiatric disorders. Counsellor Two had previously worked as a civil engineer before training as a counsellor. Client Recruitment Clients with serious health concerns were recruited through postings and information packages placed in local agencies and clinics serving this client population (see Appendices C and G). In addition, advertisements for the study were placed i n local newspapers. Clients were invited to participate in this study in exchange for free counselling. Interested participants were asked to contact the primary researcher directly. This client population was selected because there is evidence that therapists experience a variety o f emotional reactions when working with clients who suffer from serious health concerns (e.g., Hayes & Gelso, 1993). Because countertransference is a nebulous event to capture, particularly in its subtle forms, increasing the chances o f it occurring by including client groups and issues identified as challenging for therapists seemed important. T w o individuals were referred to the study from local health care clinics serving patients with A I D S and cancer. Nine individuals with a variety o f personal and health concerns responded to the advertisement in the local newspaper. The researcher conducted a preliminary screening interview with potential clients over the phone to ensure suitability for the study. The inclusion criteria were as follows: clients were currently experiencing serious health concerns; clients were not currently seeing another therapist; clients were emotionally stable enough to attend the eight counselling sessions (clients with active substance abuse problems, on-going psychotic episodes, or who were actively suicidal were not included); clients were willing to have the counselling sessions video-taped for the study; and, clients were available during the period in which the study was being conducted. O f the 11 respondents, three met the criteria for the study. Clients who were not selected for this study were provided with referrals to suitable free or low-cost counselling resources. After initial contact over the phone, a meeting was arranged in the Faculty o f Education at the University o f British Columbia with the potential clients. The meeting served several functions, first, it was an opportunity to screen potential clients "face-to-face" for suitability for the study; second, the general purpose o f the study, along with the  45  requirements o f the study, were explained; third, i f acceptable to the client and the researcher, the client signed an informed consent form (see Appendix C ) and completed the demographic questionnaire (see Appendix D ) , the Brief Symptom Inventory (BSI; Derogatis, 1993) (see Appendix H ) ; and lastly, the client was given a tour o f the counselling office and an appointment was made to begin the counselling sessions the following week. Clients' demographics. Client One was a 36 year old, White female. She was single and had no children. She identified her ethnic background as Canadian. Her highest level o f education was a Bachelor o f Arts degree. She identified her main concerns as "the unpredictable and potentially progressive nature o f the chronic illnesses affecting me ( M S and Chron's Disease). Emotionally, this leads to fear, depression, and lack o f confidence." Client T w o was a 50 year old, White female. She was single and had no children. She identified her ethnic background as Celtic and was born in Scotland. Her highest level o f education was a college degree. She identified her main concerns as "having rheumatoid arthritis and increasingly have degenerative joints. I worry about not remaining independent and all that that would entail. I had a hysterectomy in February and developed pulmonary embolism in my lungs." Clinical Research Supervisor: Recruitment and Demographics A n experienced clinical supervisor was enlisted to provide supervision to the counsellors during the study. This was a different supervisor from the supervisors who completed the initial evaluations o f the counsellors and was unacquainted with them prior to the study. The requirements for the clinical research supervisor were that: 1) he or she be a registered psychologist; and 2) he or she have at least three years experience supervising counsellors. A brief general description o f the study was provided to the supervisor along with her responsibilities (see Appendix C ) . The clinical research supervisor was a 45 year old, White female with 12 years experience as a Ph.D. level psychologist. She was working in a private therapy practice parttime and conducting supervision with therapists working at a mental health clinic part-time. Judges Three judges who were trained for the G study (Judges A , B , and C ) served as judges for the current (D) study. The first two judges participated in the G study. The third judge participated in a portion o f that training and was a full member o f the judging team for the D  46  study. Judges' demographics. A l l judges were White females in their middle thirties. The first judge (Judge A ) , also the researcher, was a Master's level clinical counsellor with seven years counselling experience. In addition to completing the requirements for a doctoral degree in counselling psychology, she worked part-time in a private therapy practice. The second judge (Judge B ) was a Ph.D. level psychologist with eight years counselling experience. She worked in a Government facility conducting risk assessments for juvenile offenders as well as in a private therapy practice. The third judge (Judge C ) was a Master's level clinical counsellor with seven years counselling experience. She worked as a therapist in a drug and alcohol clinic. Step 3: Treatment Implementation and Data Collection Figure 3 provides an overview o f the order o f treatment and data collection. A s depicted in the diagram, several sources o f data were collected from the counsellor (e.g., session notes, post-session self-report ratings), the counsellor's previous supervisor (e.g., presession ratings o f counsellor) and the research supervisor (e.g., supervision notes, post-session ratings o f counsellor). Various measures were also completed by the counsellors (e.g., demographic information), clients (e.g., demographic information, B S I ) , and supervisors (e.g., C F I - R ) for descriptive purposes. Treatment The counselling sessions took place from July to September 2001. The client/counsellor pairings were randomly assigned. The researcher met the client and therapist together prior to their first session to facilitate introductions and respond to any questions. Therapy sessions were generally booked the same time every week for each counsellor-client dyad. The researcher and a university supervisor were always available on campus while the therapy sessions were being conducted in the event o f any complications. Treatment consisted o f eight, 50-minute counselling sessions that were provided free o f charge to the clients. Manualized treatment protocols were not used in this study, instead, therapists were permitted to conduct therapy as they deemed appropriate for their client. The counselling sessions were audio and videotaped. The therapists collected all materials after each session (i.e., video and audio-tapes, completed forms) and returned them to the researcher in the Department o f Counselling Psychology at the University o f British  47  o 10  m  c o  u _©  o o (0  Q (0 ©  o  3  o  o  co m  CO TJ  c  (0  c  £  o  i-  .O' f> co o <D CO  m  z  o  1—  3  to  o  n>2  CO  D. co CO  • o  » O  Q — . .  CO CO  CD «> c ."= -S £ • CL — 33 CO 3  O  5  ~  o  «> Z CO  c  o  to CO CD CO  Columbia. Session Notes After each session the counsellors completed notes outlining the content and the process o f the therapy session. In addition, each counsellor was asked to respond to the following questions: 1) "From your perspective, what went well during the session and what was difficult during the session?" and 2) "What was most significant for you from this session?" This form is listed in Appendix I. Supervision Sessions T w o individual supervision sessions were also provided to each therapist. The therapists were instructed to contact the clinical research supervisor and arrange a meeting within a week o f their fourth and final counselling sessions. The supervision sessions took place at the supervisor's office. The clinical research supervisor was also available for additional sessions and phone contact throughout the study i f deemed necessary. The supervision sessions served two purposes: one, to fulfill ethical obligations to the client and the therapist; and two, to provide an opportunity to explore challenges and conflictual feelings aroused in the therapist by the counselling sessions. The supervisor's ratings o f the therapists' reactions were incorporated in the data analysis to help identify potential countertransference triggers. The clinical research supervisor was directed to keep the supervision sessions relatively unstructured, but to address the following general questions: 1) H o w do you find the sessions to date?; 2) What seems to be going well during the sessions?; and 3) A r e there any issues that are posing a challenge for you? The final supervision session was similar in nature, with only minor revisions to the questions: 1) H o w did you find the sessions overall?; 2) What worked well between you and the client?; and 3) Were there any client issues that were challenging for you? These questions and the directions to the supervisor are contained in Appendix J. Counsellors' and Supervisors' Ratings Identifying Potential Triggers A list o f items identifying a variety o f client issues, interpersonal styles, and emotions were compiled and used to elicit ratings from the counsellor, his or her previous supervisor, and the research supervisor regarding potential countertransference triggers for each counsellor. The items pertaining to "client issues" were organized under five areas:  49  relationships, mental health issues, physical health issues, developmental period client was discussing (e.g., past versus present issue), and miscellaneous. The content areas were thought to represent common issues clients address in therapy. The list was given to several colleagues o f the researcher to review and make recommendations. Each content area contained the item "Other" to include new information not covered in the list. The items depicting the client's interpersonal style were derived from the interpersonal circumplex literature, mainly Wiggins (1995) Interpersonal Adjectives Scales. The items describing the client's emotional tone were based on a list o f feeling words condensed to six broad categories o f feelings: mad, sad, peaceful, powerful, joyful, and scared. The counsellor, his or her previous supervisor, and the research supervisor, were asked to rate the items using a five point Likert scale according to level o f difficulty they believed the counsellor had with these issues (ranging from 0 = Not at all to 4 = Extremely). The counsellor and the research supervisor completed the ratings once all the therapy sessions were completed. The previous supervisors completed the ratings at the beginning o f the study. Both the previous and the research supervisors were also asked to respond to the following question: " G i v e n what you know about the counsellor's personal history and counselling skill, are there any other issues not already identified that he or she may be sensitive to in a counselling setting? Please explain." This question was attached at the end o f the list o f items described above. The counsellor was also asked to respond to a series o f questions, such as, "Was there anything about this client's experience that you found difficult to work with?" The purpose o f these questions was to encourage the counsellor to selfidentify potential countertransference triggers. Appendix E contains the list o f items and questions to the supervisors and counsellors. Descriptive Measures This study also included a variety o f data collecting measures that yielded descriptive information about both clients and counsellors. These measures included: 1) a demographic questionnaire; 2) The B r i e f Symptom Inventory (Derogatis, 1993); and 3) The Countertransference Factors Inventory - Revised (Latts & Gelso, 1996). Demographic questionnaire. Basic background information, such as age and gender, was collected from both the counsellors and clients for descriptive purposes. The counsellors' questionnaires also included questions concerning theoretical orientation and counselling  50 experience. Appendix D includes samples o f these questionnaires. Brief Symptom Inventory. Although the focus o f this study was not to assess treatment outcome, the Brief Symptom Inventory (BSI; Derogatis, 1993) was included to identify client concerns for descriptive purposes. The BSI is a 53-item self-report inventory developed to assess psychological symptomology (see Appendix H). A five-point scale is used to rate the client's level o f distress for each item (ranging from 0 = not at all to 4 = extremely). In addition to the 53 individual symptoms, the B S I contains nine primary symptom dimensions and three global indices. When testing the applicability of the B S I with college students, Hayes (cited in Rosenberger & Hayes, 1998) found evidence for six factors (Depression, Somatization, Hostility, Social Comfort, Obsessive-Compulsive, and Phobic Anxiety). The alphas ranged from .70 (Phobic Anxiety) to .89 (Depression) and the convergent validity correlations with a problem checklist ranged from .40 to .69 (Hayes, cited in Rosenberger & Hayes, 1998). In addition to completing the BSI in the initial meeting, the clients were asked to repeat this measure on two other occasions, after the fourth session and after the last session. The counsellors were provided with copies o f the B S I in envelopes. U p o n completion o f the fourth and the last session, the counsellors gave the clients the envelope containing the B S I and asked them to complete the measure in the counselling room. The clients were left alone to complete the measure. When the clients were finished, the counsellors collected the BSI and returned it to the researcher in the Department of Counselling Psychology at the University o f British Columbia. The Countertransference Factors Inventory - Revised. In order to assess the counsellors' ability to manage potential countertransference reactions, the research supervisor and previous supervisors were asked to complete The Countertransference Factors Inventory - Revised (CFI-R; Latts & Gelso, 1996). A s with the other ratings, the previous supervisor completed the measure at the beginning of the study and the research supervisor completed it after the supervision sessions were finished. The C F I - R contains 40 items reflecting five qualities: empathy, anxiety management, conceptualizing ability, self-insight, and self-integration (see Appendix F). Experts in the field of countertransference hypothesize that these qualities are important in the management of countertransference feelings (Van Wagoner, Gelso, Hayes, & Diemer, 1991). Supervisors  51 rate therapist-trainees on each item on a five-point Likert scale (ranging from 1 = strongly disagree to 5 = strongly agree). Higher scores on the C F I - R are thought to suggest greater countertransference management ability. Latts and Gelso (1996) report coefficient alpha estimates ranging from .85 to .96 for the subscales. When compared with therapists in general, excellent female and male therapists were thought to possess more o f the characteristics, as defined by the CFI, required to manage countertransference feelings (Van Wagoner et al., 1991). This information was used for descriptive purposes to more fully understand the counsellors in the study. Step 4: The D Study: Research Question One The first research question is stated as follows: Can countertransference behavioural manifestations, defined as therapist over-involvement and under-involvement, be reliably identified during counselling sessions with clients? A s described in step one, this research question had to be answered first before the second research question could be addressed. This section describes the data preparation, basis for judges' ratings, judges' training, and analysis used in the D study to answer the first research question. Data Preparation The 16 video-tapes of the therapy sessions (8 from dyad One and 8 from dyad Two) were transcribed verbatim. In the transcripts, each statement made by the counsellor and client was numbered according to talking turns. For example, the clients' first statements were numbered l a and the counsellors' first statements, or responses, were numbered lb. This numbering system was used throughout the transcript (e.g., 2a, 2b, 3a, 3b, etcetera). The video tapes o f the counselling sessions and the corresponding transcripts were employed to rate countertransference behavioural manifestations. The judges viewed all the sessions in sequence (e.g., 1 through 8) and rated the counsellor's behaviour for over/underinvolvement. Rating Countertransference Behavioural Manifestations A s described earlier in the G study, countertransference behavioural manifestations, operationalized as therapist over-involvement and under-involvement, were assessed by three trained judges observing video-tapes o f counselling sessions, paired with verbatim transcripts, and rating each o f the responses made by the counsellors using a 7-point Likert scale (i.e., +3 = over-involved, +2 = somewhat over-involved, +1 = possibly over-involved, 0  = empathic connection, -1 = possibly under-involved, -2 = somewhat under-involved, and -3 = under-involved). The judges' agreement for over or under-involvement constituted a rating o f countertransference behaviour. The same manual described in Appendix B was employed for the D study. W h i l e watching the videotapes o f the counselling sessions and reading the corresponding transcripts, the trained judges used the seven point Likert scale to rate each counsellor response according to his or her level o f empathic involvement or over/under-involvement. Judges' training. The same two judges who participated in the G study were used in the D study. The two G study judges ( A and B ) were joined by a third judge (C), as the G study indicated a need for three judges. The third judge also participated in a portion o f the G study training but did not receive the same degree o f training as the other two judges. The training procedures for the G study were described above. Four counselling sessions were rated in the G study by the three judges (two from Counsellor A and two from Counsellor B ) . The intraclass coefficients for the three judges' ratings for the four sessions was .72 and .76 for Counsellor A and .63 and .70 for Counsellor B . A l l three judges participated in further training for the D study (see below). A week prior to commencing the ratings for the D study, the three judges met for approximately eight hours to discuss the rating system and compare ratings from the G study, including those that had included the third judge. The judges were given a copy o f their ratings to review as the four videotapes from the G study were played. The tapes were stopped at five to ten minute intervals to discuss ratings. Approximately three hours o f videotape were observed. Anchoring ratings to the descriptors contained in the I C B helped to clarify discrepancies. A t times judges experienced difficulty assigning a rating as overinvolved or under-involved, i f aspects o f both dimensions were present in the counsellor's response. For example, the counsellor's response could be rated as under-involved (e.g., "was critical o f the client") and over-involved (e.g., "talked too much). In those instances, judges were directed to assign the rating that they felt most represented the response. B y the end o f training, the judges had discussed and resolved differences in ratings o f two points or greater. T o help the third judge gain additional training in conducting ratings, the three judges watched the first two sessions for each counsellor in the D study together. They made their  53  ratings independently, but at five to ten minute intervals the video was stopped so they could compare their ratings. Discrepancies in ratings o f two points or more were discussed. A t this point, judges could either keep their original rating or alter their rating if they found the discussion compelling. The altered rating was circled so that two reliabilities could be calculated, one for the "original rating" and one for the "revised rating." If reliability were above .75, the judges would cease to compare ratings in subsequent sessions. B y the third session, judges' ratings converged so no additional consultation was necessary. Data Analysis Quantitative methods were used in the D study to analyze the first research question. First, descriptive statistics were calculated to provide an overview o f the counsellors' response patterns (e.g., mean and standard deviation of judges' ratings, frequency o f counsellor's over-involved and under-involved behaviour for each session). Second, interrater reliability for the judges rating the counsellors' behaviour was calculated using the intra-class correlation. Third, the judges' ratings were graphed for each session using the moving averages to depict the therapists' over-involved and under-involved behaviour. Each data point on the graph represented the mean of five counsellor responses, shifting forward by increments o f one response (e.g., mean of responses 1-5, 2-6, 3-7, etcetera). The use of this graphing procedure allowed the researcher to detect subtle changes in the counsellors' response patterns. The graphs charted the therapists' departures from an empathic stance. The therapists' responses above one standard deviation from their mean were considered examples of countertransference behaviour and identified as "over-involved or underinvolved episodes." These episodes were analyzed further in the second research question. Step 5: The Q Study: Research Question T w o The second research question is stated as follows: Is there evidence to support the contention that counsellor over-involvement and under-involvement are valid indicators o f countertransference behaviour? Research suggests that therapists may react to a variety o f triggers, such as the content of client material, the client's emotional expression, and the client's presentation style that may lead to countertransference behaviour (e.g., Hayes et al, 1998; Latts & Gelso, 1995). It is theorized that therapists are more vulnerable to material (content, emotions, and/or styles) that is related to unresolved conflict within their own lives (e.g., Hayes et. al., 1998). This section describes the data preparation and analysis used in the  54 Q Study to answer the second research question. Data Preparation Countertransference triggers were assessed using qualitative research methods, mainly content analysis o f the counsellor/client dialogue associated with the over/underinvolved episodes identified as departures from the counsellor's typical pattern in the first research question, and by triangulation o f other sources o f data (depicted earlier in Figure 3). The qualitative analysis was conducted by the author. A grid was developed to organize the qualitative data into five columns: 1) client stimulus, 2) counsellor verbalizations, 3) session notes, 4) supervision session, and 5) ratings by counsellor, previous supervisor, and research supervisor regarding the counsellor's reactivity to a variety o f client issues. The purpose o f organizing the data in this manner was to attempt to identify potential triggers leading up to the counsellors' over/under-involved responses. Consistency in the various data sources would be considered evidence o f a potential trigger. Client stimulus. In order to identify the client stimulus, each over/under-involved episode reliably identified by the judges was highlighted on the corresponding transcript for each session. The dialogue preceding each over/under-involved episode was also reviewed to understand the factors leading up to the episode. This preceding dialogue, roughly 10 counsellor-client exchanges, was read several times by the author and the client's dialogue was analyzed for content and emotional tone using both the transcript and the corresponding videotape for the episode. The videotape was also reviewed to ensure that the emotional tone o f the dialogue was accurately perceived. A brief descriptive statement was written to capture what was occurring in the session preceding the counsellor's over/under-involved reaction. This descriptive statement was labelled "Client Stimulus" and was written in the first column o f the grid. Counsellor verbalizations. A n illustrative sample o f the counsellors' reactions during the episode were recorded verbatim in column two o f the grid and labelled "counsellor verbalization." The counsellors' reactions had already been identified as either over-involved or under-involved during the first research question. The complete transcription o f each episode from the sessions is included in Appendices K and L . Session notes. The corresponding session notes for the sessions containing the episodes were reviewed to identify potential triggers. Material from the session notes was  55 thought to be relevant i f it related to the content o f the dialogue recorded in the client stimulus and counsellor verbalizations for the episodes in that session. Particular attention was given to issues the counsellor self-reported as challenging. Material was considered a "direct reference" i f the counsellor commented in the session notes about an experience during the session that could be clearly linked to the content o f the client stimulus and counsellor verbalizations for the episode in question. Material was considered an "indirect reference" i f the counsellor commented in the notes about a broader therapeutic challenge related to the episode in question, but did not have as strong a connection as a direct link (e.g., "I felt challenged when the client discussed her mother's suicide attempt," is a direct reference to a specific therapy event, whereas, "I sometimes feel helpless about the client's illness," is an indirect reference to a broader therapy issue). Direct and indirect references were recorded in column three o f the grid labelled "Session Notes." Supervision. A s with the session notes, a similar procedure was followed with the supervision notes to identify potential triggers for the counsellors. The supervisor's notes from the two supervision sessions were reviewed for relevant materiel. Material from the supervision notes was thought to be relevant i f it related to the content o f the dialogue recorded in the client stimulus and counsellor verbalizations for the episodes in the counselling sessions preceding the supervision session. Particular attention was given to issues the counsellor found challenging. Again, material was considered a "direct reference" i f the counsellor commented in the supervision session about an experience during a session that could be clearly linked to the content o f the client stimulus and counsellor verbalizations for the episode the session. Material was considered an "indirect reference" i f the counsellor commented during supervision about a broader therapeutic challenge related to the episode in question, but did not have as strong a connection as a direct reference. Direct and indirect references were recorded in column four o f the grid labelled "Supervision." Counsellors' and Supervisors' Ratings Identifying Potential Triggers. The list o f items rated by the counsellor, his or her previous supervisor, and the research supervisor regarding potential countertransference triggers was reviewed by the primary researcher. The responses included client issues, interpersonal styles, and emotions that may serve as triggers for the counsellor. A s with the session notes and supervision notes, items and comments related to the content o f the client stimulus and counsellor verbalizations were deemed  relevant. These ratings were listed in the fifth column o f the grid labelled "Ratings." Although this data source is furthest removed from the actual therapy event or episode, the ratings were examined for additional support to the direct and indirect references identified in the counsellor's session notes and the supervision notes. Data Analysis A s noted above, the Q Study addressed the second research question by employing qualitative methods to investigate the validity o f the behavioural measure o f countertransference. First, the content and emotional tone o f the clients' dialogue preceding the counsellors' countertransference manifestations were analyzed and compiled in a grid along with the other data sources (e.g., session notes, supervision session, and ratings by counsellors and supervisors). The over/under-involvement ratings served as the behavioural dimension o f countertransference, whereas the other sources o f data provided the cognitive and emotional elements. Second, by organizing the separate data sources into a grid, the data could be reviewed and compared, making it possible to identify common themes or potential triggers leading up to the counsellors' over/under-involved responses. I f potential triggers in the client dialogue preceding the counsellors' over and under-involved behaviour received support from multiple sources o f data in the grid, the validity o f operationalizing countertransference manifestations in this manner would be supported. Summary To summarize, the primary goal o f this research was to determine whether countertransference manifestations could be identified within therapy sessions in a reliable and valid manner. A generalizability (G) study was conducted to assess the dependability o f the behavioural measure o f countertransference and to design the decision (D) study (e.g., how many dyads, sessions, judges would be necessary to achieve a dependable measure o f over/under-involvement). The G study also served to train the judges on rating the construct under investigation. The D study addressed the first research question; the issue o f reliability was explored by assessing judges' agreement on rating the counsellors' behaviour. The issue o f validity was more complex because countertransference is a challenging construct to operationalize. The descriptors utilized to conceptualize countertransference were derived from the theoretical literature and limited empirical research. Over-involvement and under-  involvement only captures part o f this complex construct. Further evidence supporting the validity o f operationalizing countertransference in this manner was derived from the second research question. The Q study addressed the second research question; evidence that over-involvement and under-involvement were valid indicators o f countertransference behaviour was gathered by converging multiple data sources to identify potential triggers in the dialogue preceding the behaviour (e.g., content o f client dialogue, counsellor session notes, supervision notes, counsellor and supervisors' ratings). I f potential triggers could be found to precede countertransference behaviour, additional support for the construct could be established. In other words, i f the supervisors and/or the therapist identified certain issues as "difficult" for the therapist, and the analysis o f the therapy sessions indicated that when those same issues were addressed by the client the therapist responded in an over-involved or under-involved manner, then this finding may offer preliminary support that the behavioural observations o f countertransference manifestations are in fact countertransference.  Chapter I V Results This chapter is organized as follows: first, the results from the client and counsellor measures are presented; second, the results from the quantitative analysis o f research question one are described in the D Study; and third, the findings from the qualitative analysis o f research question two are presented in the Q Study. Counsellor and Client Measures Countertransference Factors Inventory-Revised The Countertransference Factors Inventory-Revised (CFI-R; Latts & Gelso, 1996) was administered to measure countertransference management. The ratings on the C F I - R range from 1 to 5, (e.g., 1 = strongly disagree to 5 = strongly agree), with higher scores reflecting better ability to manage countertransference reactions. The mean ratings for Counsellor One by the previous supervisor and the research supervisor were respectively as follows: self-insight 4.36 and 4.82; self-integration 4.45 and 5.00; anxiety management 3.75 and 4.75; empathy 4.45 and 4.82; and conceptual skills 4.00 and 4.22. The mean ratings for Counsellor T w o by the previous supervisor and the research supervisor were respectively as follows: self-insight 4.00 and 3.82; self-integration 4.00 and 3.91; anxiety management 3.00 and 2.86; empathy 4.00 and 3.45; and conceptual skills 3.44 and 3.56. Anxiety management was the lowest rating for both Counsellor One (rating by previous supervisor) and Counsellor T w o (ratings by both supervisors). B r i e f Symptom Inventory The B r i e f Symptom Inventory (BSI; Derogatis, 1993) was utilized to assess client functioning for descriptive purposes. The B S I was administered on three occasions (i.e., prior to commencing therapy, at the mid-point, and after completion o f therapy). Scores were calculated for each o f the nine factors, as well as the Global Severity Index (GSI). Client One's mean scores for the three test occasions were as follows (range = 0-4): Somatization, 0.14, 0.0, 0.29; Obsessive-Compulsive, 1.33, 0,5, 1.17, Interpersonal Sensitivity, 0.5, 0.25, 0.5, Depression, 1.00, 1.33, 1.33, Anxiety, 0.67, 1.33, 0.5, Hostility, 0.2, 0.2, 0.2; Phobic Anxiety, 0.0, 0.0, 0.0; Paranoid Ideation, 0.2, 0.0, 0.0; Psychoticism, 0.6, 0.2, 0.2. Using adult female non-patient norms, her T scores on the G S I were 58, 43, 58,  59 respectively. Scores within this range suggest that Client One was not reporting any major symptomology. Client Two's mean scores for the three test occasions were as follows (range = 0-4): Somatization, 2.29, 2.85, 3.43; Obsessive - Compulsive, 3.00, 3.83, 3.5; Interpersonal Sensitivity, 1.25, 2.0, 1.25; Depression, 1.33, 2.16, 2.16; Anxiety, 1.33, 2.00, 1.5; Hostility, 1.00, 1.00, 0.6; Phobic Anxiety, 0.2, 1.4, 1.6; Paranoid Ideation, 0.6, 2.4, 1.4; Psychoticism, 0.00, 1.4, 0.8. Using adult female non-patient norms, her T scores on the GSI were 69, 74, 71, respectively. Scores within this range suggest that Client Two was reporting some major symptomology. The factors that appeared to be contributing to the elevated GSI score were Somatization and Obsessive -Compulsive. There did not appear to be change in pre and post measures of the BSI for either client. Client One's GSI score was lower than Client Two's across all three testing occasions. Client One's GSI score went down at the mid-point, indicating a decrease in the number and the intensity o f symptoms, whereas Client Two's GSI score increased at the mid-point. The D Study: Research Question One The first research question queried whether countertransference behavioural manifestations could be reliably identified during counselling sessions. Judges' agreement for over or under-involvement by counsellors constituted a rating o f countertransference behaviour. This section presents the following results: descriptive statistics, interrater reliability, and identification of over-involved and under-involved episodes. Descriptive Statistics Descriptive statistics, including frequency distributions, means, and standard deviations were used to describe the judges' ratings o f over-involvement and underinvolvement. Table 2 reports each judges' distribution of ratings when all the therapy sessions were combined. Frequency distributions for the judges' ratings for each counsellor across sessions one through eight are presented in Appendix K. The majority o f the ratings for all judges clustered around "empathically involved" (0) and "possibly over or underinvolved" (+1 or -1). A small percentage o f ratings were at "somewhat over or underinvolved (+2 or -2) and ratings at "over or under-involved" (+3 or -3) were negligible for all judges. It appears that Judge One assigned fewer ratings at "empathically involved" and more  60 ratings at "somewhat over-involved" for both Counsellor One and T w o in comparison to Judges T w o and Three. The mean over-involvement/under-involvement rating by each judge rating both counsellors across the eight sessions is presented in Table 3. The Likert scale used in this study ranged from +3 (over-involvement) through to - 3 (under-involvement). Empathic involvement received a rating o f zero and was anchored in the middle o f the scale. The Table 2 Frequency Distribution: Percentage of Judges' Ratings Across Therapy Sessions* Ratings  Under-Involved (-3) Somewhat Under-Involved (-2) Possibly Under-Involved (-1)  Judge 1  Judge 2  Counsellor 1/2  Counsellor 1/2  .1 / .15  0/.85  Judge 3 Counsellor 1/2  .1/.45  2.03 / 6.02  2.45/7.59  2.85/6.08  17.27 / 25.30  14.00/23.17  10.24 / 24.43  Empathically Involved (0)  44.50/29.50  56.70/44.10  65.20 / 48.30  Possibly Over-Involved (+1)  29.64/33.20  20.00/18.43  19.16/17.77  6.47/5.78  6.77/5.71  2.45 / 2.82  0/.15  .1 / .15  0/.15  Somewhat Over-Involved (+2) Over-Involved (+3)  * Sessions 1, 2, and 8 were eliminated before calculating the percentages because of low inter-rater reliability.  mean represents judges' average assessment o f the counsellor's movement away from empathic involvement, assigned a rating o f zero. In order to calculate the mean rating for each session, the data was re-coded to eliminate the positive and negative signs, otherwise the ratings would cancel each other out and result in an artificially smaller mean (i.e., closer to zero). Because the counsellors' movement away from empathic involvement was central to this study, the rationale for dropping the negative and positive signs was appropriate. The direction o f the ratings was determined by referring to the raw data. Although the means are positive, the scores reflect the counsellor's average distance from empathic involvement (0) in either direction (i.e., +over-involvement or -under-involvement). The three judges' ratings were averaged to derive a combined score (mean) for each session. In addition, a grand mean was calculated for each judge and for the combined score across all the sessions. Overall, the mean ratings for all three judges were higher for Counsellor T w o than Counsellor One. There also seemed to be a pattern in the grand mean ratings: Judge One's  61 ratings for both counsellors were higher than Judge T w o and Three's ratings and Judge T w o ' s ratings were higher than Judge Three's. Table 3 Mean Ratings of Over/Under-Involvement for Counsellor 1 and 2* Counsellor 1  Judge 1  Judge 2  Judge 3  n  M/SD  M/SD  M/SD  M/SD  Session 1  101  .52/.54  .46/.67  .32A51  .43A42  Session 2  204  .73/.60  .47/.69  .47A63  .56A48  Session 3  128  .86A68  .161.16  .69A76  .77A58  Session 4  122  .74A65  391.51  .52A63  .55A52  Session 5  167  .52/.56  .39A62  .35A54  .42A47  Session 6  144  .61/.67  .47A67  .22A45  .43A52  Session 7  134  .54/.60  .67A62  .30A48  .50A47  Session 8  227  .48/.51  .33A53  .20A45  .33A37  Grand Mean**  695  .64/.64  .53A67  .40A60  .53A52  Judge 1  Judge 2  Judge 3  Single Score  n  M/SD  M/SD  M/SD  M/SD  Session 1  383  .88/.63  .78A58  .55/.61  .74A49  Session 2  432  .88Z.46  .73A59  .49A67  .70/.44  Session 3  362  .93A62  .91/.76  .111.12  .87A57  Session 4  229  .71/.60  .67A68  .101.69  .69A52  Session 5  295  .111.51  .68A65  .47A65  .64A49  Session 6  261  1.10/.63  .87A82  .72A62  .90A58  Session 7  237  .57A59  .31/.51  .37A56  .41/.44  Session 8  212  1.06/.67  1.01/.68  .52A66  .86/.51  Grand Mean**  1384  .83/.63  .71/.73  .62A67  .72A55  Counsellor 2  Combined Score  *The 7-point Likert scale used for rating the counsellors' behaviour ranged from -3 (under-involvement) to +3 (over-involvement). Empathic connection (0) was anchored in the middle of the scale. The positive and negative signs were eliminated to calculate the means. The mean represents the counsellor's average "movement away" from an empathic connection, in either an under-involved or over-involved direction. ** Sessions 1, 2, and 8 were eliminated before calculating the grand mean because of low inter-rater reliability. Reliability A s reported earlier in the methods section, each counsellor response or talking-turn during each session was rated for over-involvement, empathic involvement, and under-  62 involvement using a seven-point Likert scale. T o assess interrater reliability, intraclass correlations were performed for the three judges' ratings o f the counsellors' behaviour. The intraclass correlations for the three judges are reported in Table 4. They ranged from .54 to .84 for Counsellor One and .63 to .84 for Counsellor Two. The reliability o f the three judges' ratings across all eight sessions was also assessed using the intraclass correlation. Coefficients o f .76 and .79 were obtained for Counsellor One and Two, respectively. Research employing Bandura et al.'s (1960) coding system for approach-avoidance behaviour reported similar interrater reliability coefficients (this coding system operationalized countertransference behaviour as avoidance reactions and facilitative counsellor behaviour as approach reactions). Hayes and Gelso (1993) reported coefficients o f .60, .66, and .79 for pairs o f raters in their analogue research design. B y the end o f training, Rosenberger and Hayes (2002) reported interrater agreement for the approach-avoidance classification as .75, .80, and .80 for pairs o f judges. The authors'  Table 4 Inter-rater Reliability: Intraclass Correlation for Three Judges 7-point Likert Scale Counsellor 1  Counsellor 2  Session 1  .54  .63  Session 2  .61  .66  Session 3  .81  .83  Session 4  .77  .84  Session 5  .82  .81  Session 6  .84  .82  Session 7  .84  .81  Session 8  .66  .75  case study obtained a generalizability coefficient o f .62 for judges' agreement o f avoidance behaviour. Although employing a different coding system, these studiesprovide a basis for comparing the ratings o f the current study. Overall, the judges' ratings o f counsellor over-involvement and under-involvement were deemed reliable, supporting the first research question. T o ensure that a reliable sample o f countertransference behaviour was utilized for the second research question, only sessions above .75 were analyzed further.  A s described in the methods section, sessions one and two were used to recalibrate the judges' ratings given the lapse in time between the training sessions for the generalizability study, addition o f a third judge, and commencing the decision study. The three judges watched the sessions together, making their ratings independently. The videotape was stopped approximately every five minutes so the judges could compare and discuss their ratings. Only the judges' original responses were included to assess reliability, however. After their discussions regarding the ratings, the judges' recorded any changes to their ratings beside their first rating. The "revised" ratings were circled to differentiate them from the "initial" ratings. The initial reliability coefficients for sessions one and two were .54 and .61 for Counsellor One and .63 and .66 for Counsellor T w o . The revised reliability coefficients were .78 and .83 for Counsellor One and .87 and .85 for Counsellor T w o . Because the initial reliability coefficients for sessions one and two were low, the judges watched sessions three and four together. Again, the ratings were made independently, but the judges compared their responses and discussed ratings two or more points apart (e.g., 0 vs. 2, -2 vs. +2). The intra-class correlation for sessions three and four were deemed adequate so the judges rated the remaining sessions without comparison o f ratings. The reliability coefficient dropped for session 8 for both counsellor one and counsellor two. The drop in the reliability coefficient could not be accounted for by reduced variability in the judges' ratings (e.g., reliability decreases i f the range o f ratings is restricted). It is possible that the drop in the reliability coefficient in the eighth session could indicate a need to repeat calibration o f judges' ratings with this method o f assessing countertransference manifestations. Judges reported difficulty at times assigning directional ratings (e.g., over-involvement or under-involvement) when the counsellor's response included aspects o f both dimensions. In addition, it is possible that the judges' ratings may have "drifted" by the eighth session. F o r example, without the opportunity to discuss the ratings, the judges' subjectivity may have influenced their ratings more strongly over time. Anchoring ratings to the behavioural criteria seemed to help judges during the training sessions. Thus, re-calibrating ratings more frequently may be necessary to maintain high reliability across numerous sessions.  64 Intraclass correlations were also performed for pairs of judges. Table 5 contains the correlations between Judges One and Two, One and Three, and T w o and Three. Judges One and T w o generally appear to have slightly higher reliability coefficients for both Counsellor One and T w o compared to the other combination o f judges. Table 5 Inter-rater Reliability: Intraclass Correlation for Pairs of Judges Judges 1 & 2  Judges 1 & 3  Judges 2 & 3  Counsellor  Counsellor  Counsellor  1/2  1/2  1/2  Session 1  .40/.59  .54A57  .38/.41  Session 2  .38A66  .67/.50  .45/.53  Session 3  .83/.78  .71/.78  .68A73  Session 4  .63A87  .69/76  .77A72  Session 5  .80/.79  .71/.71  .74/.7I  Session 6  .85A84  .73A63  .73/73  Session 7  .85/.71  .74A79  .69A73  Session 8  .65/.77  .46A58  .62/60  Identification o f Countertransference Episodes In order to track the counsellors' movement away from empathic involvement towards over or under-involvement, the judges' ratings for each session were depicted in linear graphs using moving averages. Each data point on the graph represented the moving average for the judges' ratings o f five o f the counsellor responses or talking turns. For example, the first data point represents the mean o f ratings one to five, the second data point represents the mean of ratings two to six, the third data point represents the mean o f ratings three to seven, etcetera. The Y axis represented the judges' ratings (from empathic involvement (0) to over or under-involvement (3)) and the X axis represented the counsellors' talking-turns during the session. A s with calculating the means, the data was re-coded to eliminate the positive and negative signs before constructing the graphs. A greater degree o f movement from the X axis indicated that the counsellors' behaviour was rated as a departure from empathic involvement, into either over or under-involvement. Whether a peak on the graph represented  65 over or under-involvement was determined by referring to the raw data. The moving averages for the judges' combined score for Counsellor One and T w o is represented in Graphs 1 through 10. Appendix L contains Graphs 11 to 20 depicting the judges' separate ratings for Counsellor One and T w o , respectively. The grand mean and standard deviation for the judge's combined score was used to determine a cutoff point for each counsellor. Only those ratings one standard deviation above the grand mean for a specific counsellor were analyzed further and were labelled as "significant episodes" for that counsellor. The counsellors' grand mean and standard deviation were employed for the cutoff (rather than the session mean and standard deviation or a uniform cutoff point), because it was thought to more accurately represent a divergence from his or her "typical" response pattern. The line drawn across the graph denotes one standard deviation above the counsellors' grand mean. For Counsellor One the cutoff point was 1.05 and for Counsellor T w o the cutoff point was 1.27. Across sessions three to seven, a total o f 13 episodes were above the cutoff point for Counsellor One and 20 episodes were above the cutoff for Counsellor Two. The Q Study: Research Question T w o Research question two addressed whether there is evidence to support the contention that counsellor over and under-involvement were valid indicators o f countertransference behaviour. Research question two was analyzed qualitatively: first, this section will discuss how support for over and under-involvement as countertransference manifestations was established; and second, common themes or potential triggers that emerged from the data will be identified. Convergence o f Data Sources: Establishing Support for Episodes as Indicators o f Countertransference The 13 episodes for Counsellor One and 20 episodes for Counsellor T w o identified in research question one as departures from the counsellor's typical pattern were analyzed qualitatively in order to illustrate counsellor over and under-involved behaviour and investigate potential triggers for the counsellor. A s described earlier in the methods (see Figure 3), multiple sources o f data (i.e., session notes, supervision notes, counsellors' ratings, and previous supervisors' and research supervisor's ratings) were converged to investigate the second research question. This data was compiled in a grid containing five columns:  66  Graph 1 | g  Moving Averages: Judges' Combined Ratings 1  6  3.00  21.00 12.00  39.00  30.00  Talking Turn Counsellor 1 Session 3 Cutoff = 1.05  57.00  48.00  75.00  66.00  93.00  84.00  111.00  102.00  120.00  Graph 2 Moving Averages: Judges' Combined Ratings 2.0,  Talking Turn Counsellor 1 Session 4 Cutoff = 1.05  68  Graph 3 | o  o >  Moving Averages: Judges' Combined Ratings 2  .o  1  12.00  30.00  48.00  Talking Turn Counsellor 1 Session 5 Cutoff = 1.05  66.00  84.00 102.00 120!00 138.00 156.00  69  Graph 4 c E  Moving Averages: Judges' Combined Ratings  >  O >  c. I  l_  <u  TJ C  *—  CD >  o  c CD  3.00  21.00 12.00  39.00  30.00  48.00  Talking Turn Counsellor 1 Session 6 Cutoff = 1.05  57.00  75.00  66.00  93.00  84.00  111.00  102.00  129.00  120.00  138.00  70  Graph 5 Moving Averages: Judges' Combined Ratings |  2.0.  75 >  •-47  ,—i-J-,  ,  3.00  21.00 12.00  ,  39.00  30.00  48.00  Talking Turn Counsellor 1 Session 7 Cutoff = 1.05  H=  ,  57.00  , — — ,  75.00  66.00  .  1  93.00  84.00  .  •  111.00  102.00  129.00  120.00  71  Graph 6 c  Inv Ive  E  o  iL <D "O  c  > o c  CO CD 2  I• 3.0  -r  2.52.01.51.0.50.0 3.00  39.00  21.00  75.00 111.00 147.00 183.00 219.00 255.00 291.00 327.00  57.00  93.00 129.00 165.00 201.00 237.00 273.00 309.00 345.00  Talking Turn Counsellor 2 Session 3 Cutoff = 1.27  Graph 7 M o v i n g A v e r a g e s : J u d g e s ' C o m b i n e d  -  1  R a t i n g s  AAA  rt  1  1  X  39.00  21.00  — , — , — , — i — i — i — • — < 75.00 111.00 147.00 183.00  57.00  Talking Turn Counsellor 2 Session 4 Cutoff = 1.27  93.00  129.00  165.00  ' 219.00  201.00  73  Graph 8 |  M o v i n g A v e r a g e s : J u d g e s ' C o m b i n e d R a t i n g s  "4 3.00  ,  " !  I i.  39.00  21.00  ,  ,  75.00  57.00  r - "  111.00  93.00  Talking Turn Counsellor 2 Session 5 Cutoff = 1.27  r  1  -.  147.00  129.00  .  .  183.00  165.00  .  .  219.00  201.00  >  >  255.00  237.00  291.00  273.00  74  Graph 9 c E  M o v i n g A v e r a g e s : J u d g e s ' C o m b i n e d  R a t i n g s  >  O > XJ  a  Q3  >  O c  (TJ  3.00  39.00 21.00  75.00  57.00  Talking Turn Counsellor 2 Session 6 Cutoff = 1.27  111.00  93.00  147.00  129.00  183.00  165.00  219.00  201.00  255.00  237.00  75  Graph 10 C <D  E cu  M o v i n g A v e r a g e s : J u d g e s ' C o m b i n e d  R a t i n g s  >  o > CD  -a c 3  i  (D >  o c  CO CD  3.00  39.00 21.00  75.00 57.00  Talking Turn Counsellor 2 Session 7 Cutoff = 1.27  111.00 93.00  147.00  129.00  183.00  165.00  219.00  201.00  1) client stimulus (e.g., potential triggers), 2) counsellor verbalizations (e.g., over and/or under-involved response), 3) session notes, 4) supervision notes, and 5) counsellors' and supervisors' ratings. Tables 6 and 7 depict this grid for Counsellor One and T w o , respectively. The tables appear at the end o f this section starting on page 118. A description o f the dialogue between the counsellor and client preceding the counsellor over and under-involved response is included in column one (client stimulus). Direct quotations from the over and under-involved episodes are reported in column two (counsellor verbalizations). Relevant data from the session notes, supervision sessions, and ratings by the counsellor and supervisors were included in the remaining columns o f the grid. The full transcription o f this information was not included to protect the counsellors' and clients' confidentiality. Evidence that over and/or under-involvement were valid indicators o f countertransference behaviour was established by gathering support for potential triggers in the client stimulus. Direct confirmation for a therapy event or trigger was determined to exist i f the event in the client stimulus was directly referred to by the counsellor as problematic in either the session notes, the supervision session, or the post-session ratings. Indirect confirmation for a trigger was determined to exist i f the event or content and emotional tone o f the client stimulus was identified by the counsellor or supervisors as a difficult issue for the counsellor, but was not specifically referred to by the counsellor. In other words, although the counsellor may not have directly identified an event during the episode as "difficult," i f the content and emotional tone o f the client dialogue (client stimulus) contained issues identified in the supporting data as "challenging" for the counsellor, then the episode was deemed to have indirect confirmation. The qualitative analysis o f the data suggested that there was preliminary support for identifying potential triggers - or therapy events that elicited a reaction in the therapist - for both Counsellor One and T w o . O f the 13 episodes identified for Counsellor One, four episodes had direct confirmation from the supporting data regarding the triggers identified in the client stimulus (session 3: episodes 1 and 3; session 5: episode 1; and, session 7: episode 1.) The remaining 9 episodes had indirect confirmation from the supporting data for the triggers identified in the client stimulus. O f the 20 episodes identified for Counsellor T w o , six episodes had direct confirmation from the supporting data regarding the triggers  77 identified in the client stimulus (session 3: episode 6, 7; session 4: episodes 1, 2, and 3; and, session 5: 2). The remaining 14 episodes had indirect support for the triggers identified in the supporting data. Interestingly, some episodes were identified by the counsellors as "helpful" in the session notes, but were rated by the judges as over-involvement and/or under-involvement. For example, judges rated Counsellor One's behaviour as under and over-involved for episodes 2 and 3 in session 5, respectively, whereas Counsellor One identified those episodes as "what went w e l l " during the session. Similarly, the judges rated Counsellor T w o ' s behaviour as either over and/or under-involved for session 4 episode 4 and session 6 episodes 2, 3, and 5, whereas Counsellor T w o identified those episodes as "what went w e l l " during the session. Direct and Indirect Confirmation: Counsellor One Based on the data converged in Table 6, the episodes that received direct and indirect confirmation for the triggers are summarized below for Counsellor One. Episodes with direct confirmation ( D C ) are presented first, followed by those that received indirect confirmation (IC). For the episodes that received direct confirmation (e.g., D C ) , only the datum source that included direct support for the potential trigger(s) in the client stimulus is reported below. For the episodes that received indirect confirmation (e.g., IC), all the data sources that "build a case" for the potential trigger(s) in the client stimulus are described below. Session 3: episode 1 ( D C ) . The client stimulus involved discussion o f loss o f bowel control. The counsellor reported feeling challenged by the discussion o f this health issue in her session notes. Session 3: episode 3 ( D C ) . The client stimulus involved discussion o f mother and aunt's suicide attempts. The counsellor reported feeling challenged by the discussion o f these particular family traumas during the supervision session. Session 5: episode 1 ( D C ) . The client stimulus involved discussion o f client's health concerns. The client vehemently corrected the counsellor when she described M S as "your disease." The counsellor reported this incident o f experiencing correction by the client as challenging in her session notes. Session 7: episode 1 ( D C ) . The client stimulus involved discussion o f the client's dream about death earlier in the session. The counsellor did not process the dream. Later in  78  the session the client discussed feeling sad and depressed. The counsellor noted they had not processed the dream. The counsellor reported feeling challenged by the dream in her session notes and during supervision. Session 3: episode 2 (ICY The client stimulus involved the following themes: depression, emotional arousal (crying, sad, angry), and helplessness. The counsellor reported themes o f helplessness in her session notes. During supervision she reported that she found sitting with the client's feelings o f helplessness and depression difficult. Ratings by the counsellor and the supervisors o f the counsellor's reactivity to various client issues and presentation styles identified depression and emotional arousal (sadness, anger) as "slightly to quite difficult" and helplessness and lack o f control as "somewhat to quite difficult". Session 3: episode 4 (IC). The client stimulus involved the following themes: personal stress regarding health concerns, helplessness, emotional arousal (crying, sad, and powerful). In her session notes, the counsellor reported that she lacked a framework to support the client in managing her disease. During supervision she reported a sense o f helplessness regarding the client's medical issues. Ratings by the counsellor and the supervisors o f the counsellor's reactivity to various client issues and presentation styles identified personal stress, health concerns, and emotional arousal (sad) as "slightly to somewhat difficult" and helplessness and lack o f control as "somewhat to quite difficult". Session 4: episode 1 (ICY The client stimulus included discussion o f the client's health history and feelings o f helplessness. The counsellor referred to perfectionism in her response. In the session notes, the counsellor reported that she felt she supported and validated the client throughout the session. During supervision the counsellor reported feeling challenged by the client's feelings o f helplessness. She also identified perfectionism as a challenge for her working with this client. Ratings by the counsellor and the supervisors o f the counsellor's reactivity to various client issues and presentation styles identified physical health as "slightly difficult" and perfectionism and helplessness as "somewhat difficult". Session 4: episode 2 (IC). The client stimulus contained the following themes: family o f origin (mother), physical health, sadness, crying, lack o f control, and helplessness. In the session notes, the counsellor reported that she felt she supported and validated the client throughout the session. During supervision the counsellor reported feeling challenged by the  client's feelings o f helplessness regarding medical concerns. She found it difficult to stay with client's emotional expression and not move into a problem-solving mode. She also noted she identified with the client's mother issues. Ratings by the counsellor and the supervisors o f the counsellor's reactivity to various client issues and presentation styles identified parents, physical health, helplessness, lack o f control, and sadness as "slightly to somewhat difficult". Session 5: episode 2 (IC). The client stimulus included discussion o f the client's health concerns and how they have forced her to face her reality. She wondered whether she had "manifested" help when she needed it most (e.g., in reference to the counsellor). In the session notes, the counsellor noted that this part o f the session "went w e l l " whereas the judges rated the counsellor's responses as under-involved. During supervision the counsellor noted she often felt the urge to "fix" the client's health concerns. The counsellor reported that loss o f control and helplessness was challenging for her to deal with. Ratings by the counsellor and the supervisors o f the counsellor's reactivity to various client issues and presentation styles identified physical health as "slightly difficult" and helplessness and lack o f control as "somewhat to quite difficult". Session 5: episode 3 (IC). The client stimulus included themes o f personal stress and references to death. The client was powerful in her presentation style. The counsellor reported that this section o f the session "went w e l l " in her session notes, whereas the judges rated the counsellor's response as over-involved. During supervision she related that she found sitting with depression, the client's feelings o f hopelessness, helplessness, and loss o f control challenging. Ratings by the counsellor and the supervisors o f the counsellor's reactivity to various client issues and presentation styles identified personal stress and death "slightly to somewhat difficult" and helplessness, lack o f control, and powerful interpersonal style as "somewhat to quite difficult". Session 6: episode 1 (IC). Client stimulus involved discussion o f a medical terminology course that the client found upsetting because o f its reference to M S as a progressive degenerative disease. Themes o f physical health, loss or control, and emotional arousal (sad, scared, mad, crying) were addressed. The counsellor described the incident regarding the course as upsetting for the client in her session notes. During supervision the counsellor reported feeling challenged to stay with the client's feelings and not move into a  80  problem-solving mode. She also reported loss o f control and helplessness challenging. Ratings by the counsellor and the supervisors o f the counsellor's reactivity to various client issues and presentation styles identified anger, sadness, and fear as "slightly to quite difficult" and lack o f control and helplessness as "slightly to somewhat difficult". Session 6: episode 2 (IC). The client stimulus involved a continuation o f the discussion regarding the medical terminology course. Themes o f physical health, loss o f control, and emotional arousal (fear, anger, and power) were addressed. The counsellor described the incident regarding the course as upsetting for the client in her session notes. During supervision the counsellor reported feeling challenged to stay with the client's feelings and not move into a problem-solving mode. She also reported that loss o f control and helplessness was challenging for her to deal with. Ratings by the counsellor and the supervisors o f the counsellor's reactivity to various client issues and presentation styles identified anger, fear, and power as "slightly to quite difficult" and lack o f control and helplessness as "slightly to somewhat difficult". Session 6: episode 3 (IC). The client stimulus included discussion o f a course that the client was facilitating. She described working to make the class positive for students, unlike her own experiences as a student (reference to the medical terminology course in episodes 1 and 2). The counsellor described the incident regarding the medical course as upsetting for the client in her session notes. During supervision the counsellor reported feeling challenged to stay with the client's feelings and not move into a problem-solving mode to "fix it". She also reported that the client's lack o f control and helplessness was challenging. Ratings by the counsellor and the supervisors o f the counsellor's reactivity to various client issues and presentation styles identified lack o f control and helplessness as "slightly to somewhat difficult". Direct and Indirect Confirmation: Counsellor T w o Based on the data converged in Table 7 the episodes that received direct and indirect confirmation for the triggers are summarized for Counsellor T w o . Episodes with direct confirmation ( D C ) are presented first, followed by those that received indirect confirmation (IC). For the episodes that received direct confirmation (e.g., D C ) , only the datum source that included direct support for the potential trigger(s) in the client stimulus is reported below.  For the episodes that received indirect confirmation (e.g., IC), all the data sources that supported the potential trigger(s) in the client stimulus are included below. Session 3: episode 6 ( D C ) . The client stimulus involved the client self-disclosing that one o f her older brothers had sexually abused her as a child. The counsellor reported in the session notes that the issue o f the client's abuse was challenging to deal with. H e noted that he felt as i f he wanted revenge on the client's behalf. Session 3: episode 7 ( D C ) . The client stimulus involved discussion o f the client's sexual abuse and younger brother's accidental death. The counsellor and client were completing a genogram. The client presented as sad and tearful. The counsellor reported in the session notes that he found the discussion about the client's brothers challenging. H e felt drawn into problem-solving and wanting revenge. Session 4: episode 1 ( D C ) . The client stimulus included the counsellor asking the client how she felt they were doing in terms o f developing strategies to deal with her anxiety. The client related that she did not feel they had not made any progress. The counsellor identified the client's denial o f progress as challenging in his session notes, during supervision, and in his post-session ratings. Session 4: episode 2 ( D C ) . The client stimulus involved discussion about how the client has dealt with anxiety in the past. The client described an extremely difficult period o f her life. The counsellor reported in his session notes that he felt "clumsy" in working with how the client previously dealt with anxiety. This episode followed shortly after the client's denial o f progress in therapy (see session 4: episode 1). Session 4: episode 3 ( D C ) . The client stimulus included a continuation o f the discussion on how the client dealt with anxiety. The client had reported an experience in which her ex-boyfriend had threatened her life at knife point. A s referenced above (see session 4: episode 2), the counsellor reported in his session notes that he felt "clumsy" in working with how the client previously dealt with anxiety. This episode followed shortly after the client's denial o f progress in therapy (see session 4: episode 1). Session 5: episode 2 ( D C ) . The client stimulus involved discussion o f the client's experience o f sexual abuse by her older brother. The counsellor reported the discussion o f sexual abuse as challenging in his session notes.  Session 3: episode 1 (IC), The client stimulus included themes o f physical health, self-esteem, and family o f origin. The client related how previous therapy had helped her work on assertiveness. In the session notes the counsellor stated that he decided to write notes during the session to keep focused because he found his "mind wandering" in the previous sessions trying to problem-solve. H e noted the client's conflictual family relationships in his session notes and in supervision. During supervision the counsellor reported that he liked to use notes for structure. The clinical research supervisor noted that the counsellor did not seem comfortable sitting with the client's process without a clear direction. Ratings by the counsellor and the supervisors o f the counsellor's reactivity to various client issues and presentation styles identified self-esteem, parents, and siblings as "slightly to somewhat difficult" and physical health as "somewhat to quite difficult". Session 3: episode 2 (IC). The client stimulus included the following themes: family o f origin, self-esteem, depression, and alcoholic ex-partner. The references from the session notes and supervision notes are the same as above (see session 3: episode 1). Ratings by the counsellor and the supervisors o f the counsellor's reactivity to various client issues and presentation styles identified self-esteem, parents, and siblings as "slightly to somewhat difficult," depression and substance abuse as "somewhat difficult," and ex-partner as "slightly to quite difficult". Session 3: episode 3 (IC). The client stimulus involved discussion o f the client's chaotic experiences in her family including a brief reference o f sexual abuse by her brother. She reported it was difficult for her to get mad in her family and her experiences resulted in her feeling unworthy. The counsellor referred to the client's conflictual family relationships in his session notes and during supervision. H e also noted that it was a challenge for him to deal with anger in his own life. Ratings by the counsellor and the supervisors o f the counsellor's reactivity to various client issues and presentation styles identified parents, siblings, and self-esteem as "slightly to somewhat difficult" and sexual abuse and anger as "somewhat to quite difficult". Session 3: episode 4 (IC). The client stimulus involved themes o f physical health concerns and self-esteem. The counsellor had asked the client how they were doing in terms o f the initial goals she came to counselling with. The references from the session notes and supervision notes focus on the counsellor's desire for structure and problem solving (see  83  session 3: episode 1). Ratings by the counsellor and supervisors o f the counsellor's reactivity to various client issues and presentation styles identified self-esteem and physical health concerns as "somewhat to quite difficult". Session 3: episode 5 (IC). The client stimulus included discussion about the client's recent complications from her hysterectomy operation that almost resulted in her death. Themes o f personal stress and anger towards her niece also emerged. In the session notes the counsellor noted the client's anger towards her niece and reported that the client has experienced two life-threatening episodes - an abusive relationship and blood clots from the operation. During supervision the counsellor reported the client's chaotic family o f origin. H e also related that it is a challenge for him to deal with anger in his own life. Ratings by the counsellor and supervisors o f the counsellor's reactivity to various client issues and presentation styles identified personal stress as "slightly to somewhat difficult," death as "somewhat difficult," and physical health concerns and anger as "somewhat to quite difficult". Session 3: episode 8 (IC). The client stimulus included the following themes: family o f origin, sexual abuse, and emotional arousal (sad and angry). In the session notes the counsellor reported that the client was visibly upset. H e felt like he wanted revenge against the client's brother (who had abused her). The counsellor went overtime in the session to deal with the sensitive issue the client brought up. During the supervision session, the counsellor reported that he usually goes over-time in the session. H e noted the theme o f anger in relationships for the client. The counsellor acknowledged it was a challenge for him to deal with anger in his own life. Ratings by the counsellor and supervisors o f the counsellor's reactivity to various client issues and presentation styles identified sadness and sibling issues as "slightly to somewhat difficult" and sexual abuse and anger as "somewhat to quite difficult". Session 4: episode 4 (IC). The client stimulus involved discussion o f the client's level o f anxiety. The counsellor asked the client the "miracle question." In the session notes the counsellor reported he thought the "miracle question" was successful. The judges rated this section as over/under-involved. Earlier in the session he reported feeling surprised when the client stated that they had not made much progress in terms o f dealing with anxiety (see session 4: episode 1). During supervision the counsellor reported that he liked to use notes  84  for structure. The clinical research supervisor noted that the counsellor liked to have structure available - he did not seem comfortable sitting with the client's process without a clear direction. Ratings by the counsellor and the supervisors o f the counsellor's reactivity to various client issues and presentation styles identified anxiety as "slightly to somewhat difficult". Session 5: episode 1 (ICY The client stimulus involved discussion o f physical health issues and medication. The client presented as worried and sad. The counsellor asked the client i f she had completed the homework (e.g., miracle question). The client related that she had not completed the exercise because she felt i l l . The counsellor switched the focus to gathering further information regarding the client's family o f origin. The counsellor reported in the session notes that he felt awkward during the session because he was uncertain about the amount o f progress they were making. The reference from the supervision notes is the same as above (see session 4: episode 4) - the counsellor and the supervisor reported the counsellor's preference for structure. In addition, the counsellor noted that he felt he should know more about medications, but was staying out o f that "trap." Ratings by the counsellor and the supervisors o f the counsellor's reactivity to various client issues and presentation styles identified emotional arousal (sad and worried/scared) as "slightly to somewhat difficult" and physical health and medication as "somewhat to quite difficult". Session 5: episode 3 (ICY The client stimulus included themes o f family o f origin and physical health. The counsellor was taking notes about the client's family when he noticed she was rubbing her arm. They returned to information gathering after the client confirmed that she was not cold, but her arms were sore from her arthritis and surgery. The reference from the session notes is the same as above (session 5: episode 1) - the counsellor reported feeling awkward during the session. The reference from the supervision notes is also the same as in other episodes (e.g., session 4: episode 4) - the counsellor and supervisor reported the counsellor's preference for structure. Ratings by the counsellor and the supervisors o f the counsellor's reactivity to various client issues and presentation styles identified parents and siblings as "slightly to somewhat difficult" and physical health as "somewhat to quite difficult". Session 6: episode 1 (IC). The client stimulus included the following themes: physical health, medication, emotional arousal (scared and sad), and personal stress. The client was  85  describing how her doctor had to change her medication because o f damage to her heart muscle. The counsellor was trying to write down all the medication names. In the session notes the counsellor reported that the client was experiencing difficulty with her heart. H e had prepared an agenda prior to the session and had hoped for some "positive breakthroughs" during the session. During supervision the counsellor and supervisor noted his preference for structured activities during session. The counsellor was less comfortable with a non-directive process. Also, the counsellor related that he felt he should know more about medications. Ratings by the counsellor and the supervisors o f the counsellor's reactivity to various client issues and presentation styles identified personal stress and emotional arousal (sadness and fear) as "slightly to somewhat difficult" and physical health and medication as "somewhat to quite difficult". Session 6: episode 2 (IC). The client stimulus involved personal stress. The client felt overwhelmed by the clutter in her home and lacked the energy to do anything about it. The counsellor was trying to encourage her to think o f ideas to address the clutter. The counsellor referred to the discussion o f "clutter" in his session notes. H e felt the exploration "went w e l l , " but the judges rated the episode as over/under-involved. The counsellor was hoping to discover strategies to help the client deal with her anxiety. During supervision the counsellor and supervisor noted his preference for structured activities during sessions. The counsellor was less comfortable with a non-directive process. Ratings by the counsellor and the supervisors o f the counsellor's reactivity to various client issues and presentation styles identified personal stress and anxiety as "slightly to somewhat difficult". Session 6: episode 3 (IC). The client stimulus involved discussion o f family o f origin issues (client's sister) and physical health concerns. The client was frustrated and angry about her sister's failure to ask the client how she is doing or whether she requires any assistance. I f her sister promises to help with a task, she never follows through. The counsellor recommended that she find someone else to help her. In the session notes, the counsellor reported that he felt he and the client understood each other well regarding her sister. The judges rated this episode as over-involved. During supervision the counsellor and supervisor noted his preference for structured activities during sessions. The counsellor was less comfortable with a non-directive process. Ratings by the counsellor and the supervisors o f the counsellor's reactivity to various client issues and presentation styles identified personal  86 stress and siblings as "slightly to somewhat difficult" and physical health and anger as "somewhat to quite difficult". Session 6: episode 4 (IC). The client stimulus involved discussion o f the client's difficulty "letting go o f people and things." The counsellor was trying to get the client to think o f exceptions to the rule. H e pointed out that she had "let go" o f her ex-partner. The client stated that her ex-partner had died so she had had no choice. In the session notes and during supervision, the counsellor emphasized his desire for a positive outcome during sessions. He preferred a structured style and did not seem comfortable sitting in the process. The counsellor reported during supervision that he frequently asked, "how would you recognize strategies". Ratings by the counsellor and the supervisors o f the counsellor's reactivity to various client issues and presentation styles identified personal stress and selfesteem as "slightly to somewhat difficult" and ex-partner as "quite difficult". Session 6: episode 5 (IC). The client stimulus involved a continuation o f the discussion about the client's relationship with her sister, who lives in the same apartment building. The client expressed feeling angry and fed up with her sister's selfish attitude. In the session notes, the counsellor reported that he felt he and the client understood each other well regarding her sister. The judges rated this episode as over-involved. Again, in the session and supervision notes the counsellor focused on trying to identify strategies to deal with the client's anger. The counsellor reported that dealing with anger was a challenge in his own life in the supervision session. Ratings by the counsellor and the supervisors o f the counsellor's reactivity to various client issues and presentation styles identified personal stress and siblings as "slightly to somewhat difficult" and anger as "somewhat to quite difficult".  87  il  ! o  o -a  fi oo  u O ©  o l , --"  CO lH  00 ?o  CN  .1 4>  l-i O  oo o  *—'  03 J I  O a O e  (A CO  o  o o  CJ co  1  s  T3  CO  ••a  ITS  a § 1 &  tin <o  "  O <L> S3 , G  3 .3  a "ft ft  O '55  1  1 8 i s  s o  1-2  00  cn O  e > e cu  *3  T3 1 S  •O  I  e [2 cs CU >>  CU  an  M  w \o 1 ©j  W '«» HJ  S  9 '§ S I a  •<!  ©  o>  0  a  2  3  •i a  o  |  |  •8 5  en  Q  «g 1 g G  •  1  I  2o"s  oo OO „ tu &< o  o a g  5- >H  i) S  ^ w S  I I  11  lH  o o  o co  o o •3 •8 2 o  3  X)  <4H  o o  ft c  0  h S  o ft  O  O  co *0 O  o  oo o 5  as  o  s o  8  c £ £ e £2 ^  Q 9>  si =g «  o  s  o  1/3  3  4)  O to rv <D  *  o  8  oo Pi  wQ  ~-»  PH OO "-~  W  i-H  O  "T  I —<  88  60  .3  II  -a  8«  '5 • 1—1  3  &0  CN  ~ 8  O  J  8  o o 8 1  OO i  PH  I o  co  V) <U 3  -a  O  O  o O o  to CO 4>  T3  M  co CO  CL, O  LH  0)  PH  < L > CN <s>  - ^  0)  ^sOO  oo" 00 cd  ^  CN  CN  603  o 2 _  o u  73 8 o  5? 8  OH Q  8  CO  3  4>  co 4>  -H cn« ,aa -rzj — S-S toco f? ?? *-i o-  o CO o a  1-. . 3  3  5  T3  60  5  CO  I« d o  oo  2 <u  3  ;G  '§ 1 S o  B  SI'S  1^  TJ  2  •a  2  d o  •e  G  s  .1°  1 d CO o < L >  I  o  g  i3  6  §6  r'5  o d O  M  13  ll  in CO  d  in  Q.  JH  <S  (U  3 >  o ex  •a  •a  a o l  '3  o  •8  CJ '55  c n c n cj C cn ^ cn  O  «  •a  CJ 8  O S3  o o  ^ -a  O  i>  O  55  o ^* «y §c cS  Ig  111! O  u  CM  OJ ^3  •a  3 E  "  O H  to S co y8  "S «  2o,o VH  T!  CJ *ta  *3 § 1  A  D  6  D  •a  o  •8 CJ o  a  S  60  a* "I  § «Q § c£5  Oo ?  1I O =§ S3  O CJ 6 0 xi  CJ 5  cn  7j  ^3 cj . 3  c -a  8-s *H  1* *" *2  — <u 2 CJ cn  s o  3* 9- 8  5  •2 »  s  •a o i i '5 tt> cn is cn o 3 HI 5° £ 81  aJ  •I  f  cucncn .£  •§ s  u cn cn V fc> P .  .9  s S3  3  5  ai  o 6a § o o cS £ 03  e  < L > O  5 5  g CJ  90  .1°  II  o o o  1 o u  o  o^  o  CO  SB  a o i  I IS 3  o 00  "8  o co  a- 1  C3  II  6  I d  M  I  6 "51  •a I o u  u  5-•s e O  co JS  3  2  o y O S  o  < L >  e  •a  6  O  91  o o u  I  O ca o> 43 CJ  og  13 s  O  S  U  O, I  ol  ^5  is CN  o  ,0}  _ OO"  PH ~—'  s  .5 .2  y  ti  U  00  G O  o  !S  «J ro C O c£ ffi rt J I C N ^—' I  a u  s  Mini 3  *3  o  O ^  CX32  O  cn  (JO  I - S i &"S 8\3 •3 S3 «H  *i c§ -*=£  a ™ 53 a «  T!  ol O O  3  43  •1 1 i<3 60  -3  ll s CD  •all Q  _ f  tO  b 8 o §  13  -3 8 cn —  S 13  .2  I  (D  M  | O  5 § |  -3  w  92  8 "  ?0  a  Si  ~1  O eu <N Q O  CO CO  O O o  *  I O  .i  U  o ri  co CO  " -a to 60  CO  _CJ  00  rt B Q co "3 — ^ 2P  (L)  T3  Q  <l>  H  O 1> 60 XS  CJ  rt S3 o  G O  rt a m  rt <L> 6 0  I  rt o  1  rt  rt o 60 M  I  Sp  —' co 4> co  ~ ^ g « g ao p c « | ^ » O S T3 ~  S i  I rt co  I  lo  co  fli QJ 1 ^ s..  PS c/3  Si-  P.  p  w  ^ ^  •3*  1  xi  u  CJ  o < -e * xi  2 -S  — co  1  S in  CJ X ]  Io  o  °  o o  CO  1 o  CO CO CO  4>  o  o  IslllILilslIsi 4>  I,  o  S  O TJ  CJH  O  I  e3 CO  bp  CO  - i f O 41  CJ TJ O  8 €8<a|| xi  H  S  (J £  M X )  <L>  o  rr> co  1  Io o o  «(-(  5  CJ TJ O  u  CO _co . 4> ?P<2. JU 3 4> f o- "_-43 a. 35b co -tS TJ  00  s4)  J•a a  o  2 60  a 73  o  € S<a § I 4> & 43 S ° CJ 73 6 -O CJ H CJ  2  UH e3  S  CO  s  Xi  box:  CO  o  CJ -rl *> 33  ^ w  73  •a  •8  8 8  «•§ «-> ^  O TJ 4>  I|  4) 3 S O 43 .S •5 •tf bp bp ^ CJ 43 3 4) co cd p bO CO  w  CO  o3  4> -J3 4> TJ 33 3 o 8 CJ  CJ  c3 19  o TI  if  3P t3 S 4> XW. 3 3 43 3 cj co  co cd P  <->  Q>  g 5  o  O  8 -  CJ co O  3  s 4)  4>  a S  CJ  3 © cj  ^  5  1-  TJ  3  CJ  s  XI  o  CO  fi  g  5  ill  B 3  cd TJ 4> « 3 •3  +J  I  TJ  CO  cd  1 bo „  o ,2  J g  94  *  (3 O  I-  3  ,5  8  -SI ?  £ 8" 2  00  i "8  "5 •&  I* 8  o  1 •e  2 a  ia !2 6fl M  a o  •c  et) 60  75 -C a  8 | % js  io  a & 5 .a  .s a  60  e «  s s •a 3 a3 3  8 .a 5  S  -a  £ .3T&  u 73  «z  8  rt to T«3 rF  <L>  O Q>  25  «>  2  U  <u  73 -3  1  95  I i  s •PH  96  I  O  o  CJ  o  rt £ 3  oo W D oo x,rtoo  oo  rt fe<*>^  oo 73 @£ a  rt  rt  •S  ga  2 .2 tj °P 3 c  rt  3  <u  o  S flS I i2 d S B  fi =* ,s M  I*  " I I  o 00  to  rt  *5 cd  S  S S op  p  2  "8  n  a  O Qg fi o  o  1 .-81  (A  (A  O  CA  to  O O  "8 © ja  ss §  •2  2 .2  So  73 S£  cr"a  « 1  CO  (3 O  a  1  1  i3 6C  et  •a 3  73  3  o  -g »g g  5  s i Irt §>  Sb-B " g.73 «  ••& § 1 ^ I  00  73  , „cS8.9 u  73 3  is co  o  O v»J  J  ;j <!  60  CJ  00  < L >  o  cj  o  CJ  oo <N e£ •s OO <s oo SOn  97  .i  iS  I  1  o y -Aco CN^ m  C3  CJ  ^  § 2 tS bb a 3 g u CO  •i § | a  II11  .1  l<8  o  •1  43 3  CJ  s 8 a  < L >  M  •c  1  cj  s  I  •a a D  •as  a  a (j « «j •g  IO  CD >  o co  3  CJ  «.6  o g fe . g u-f a o g ,« a w 3  tc acj P-s?s tor 0 S  3  3  .3 o  CO  d  98 I  II  S&9 8  o  O xi  u  CN  35  o  .3 S3 cf  •c 1° £r i i l v a o  "5  •g o  i—H l l  a o  oo  cd CJ  I  •c l l  IC  JJ  a •S  T3 T J  ^  13 §  to  U  h  g  CD  £  i l * bp  i £  > 2 3 p X i o •S -a T 3 O X> a a* •§ 3 3 8 .3 a 3 x i c2 •3 o s  8.  5  eah, bO bC  bo  o  lag  i C  i—>  o  j '  u  d  o  yeah  {  QJ  s  b  cj  SIS  |3  O  X) X!  X>  cj  CD  99  art  o  I 23  rt  u  8 ^ oo y i!-' n N jp  O o  « ^  ^  T3  •a  oo  bo CM  S£ XI  rt  «  o cn 73 xi i - <srt i> XI rt O 13* S a  00  O  .a < + = •  -a ^  Till SS g  sB O  O  « a cd  rt 8  rt c 5 2 <a  P  s * is .a a a <§ S a | * b | o P -Si .b 5 > 3 8 8 " p.73 -3 2 ^.y .2  M i l at!J  ~  2, B * 5 3 c° H  73 a «> *  a l §>  a if if S |  XI  8 53  o p. a i  rt  l el  o  oo  s  •i-H  O  1  (A  a  1 1 2  ( L >  a  -3  60  CD  I  60  <i>  a  a  60  3  i<3  O p  O  u  cn  O rt p  •a  O  a <u rt o  60  ill »-s CO  l l  s  IS  .  ( L >  ™ a 60 M a to a> rg c£ w  cn  75  3  rt  XI  S3 rt  o  S •£ xi o 3 a £ ja ^ -d o 5 , ii  u rt  oc  n^1 g 3  a  w8 w  a  O  I  5  oo <J  100  6 CJ  Si  « OS  o-| CO  51  oo  I*  a o  a o  a  .1  CJ  •as  4>  73  i I § s2  lJ!  * £ -2  _„  1  SI*  PI  i § •  a s a g l?l s o ^ o CJ .a x, >  XI  8  > X> T J X>  m I-H  si  CJ  5 "° R  U28  TJ  3  a a  Ia S xi  I a o  1 XI  ca  a  M  a  1 •a  CO  CI CJ  O co  S  CN  CO  XI  co  101  "8 ,2 g « |  fi  Si  ol  |  Ilia H  ^  c/3  s  JTQ P y J3 <N fe^ rt  a  so  60  e o  It III .s .a £ -  rt  te §  J  1H5 S -S « J 5  w  s "a aa a O  I  lll-a  ~ I £  .a  P  2  o  75 JO  (U  d>  o  „  I  p  1 l>  lH  o (->> to  i3  <u rt  O.Srt  « rto  l l  |60  HI*  10  0  "ct «  p  xi  .a a 75  so  h^g 8 o c3  2P  o  102  60  C/5  e o  60X7" £ « g 2 t3 5 -r w  I I  n  ll  o o  lH  60  •e O  s  i3  3 •3 •a  .9  I d  a  J .8  d«  a ^  d  1  § to  .a 3 .  as  s  Q,  d 3ss  o  < L > 60  ;1 | CO  10  TJ  v.  3  O O  1  •s 73  3  CJ  CJ ' oo  O  .53  si  co CO cS  o O  CO >> <N (M CO $5  "I "8  a  If  §  T3  I  It  CJ  I55  CO  O  CJ  O 73 <L>  <L>  co  —  CJ  S ft |> S  uri this tho he uch is he or felt  CO  CO  i  to  •g a xi  rre  o  o  •a  1  a  O .g  The  CJ  73  73 CO  g  .3  43 43 cS  Lit  «  5  co CJ  O CJ  1"§  ca - .  g 1 o  a  P-p  s £ s?  co XI 6 cd F  M f l  s  S  a  s  3  o  JJ  e ^  I'll-S3 •a  sess pref coul  §  ga  73 CO  i  a  CJ  •as ^ b xi-  xs  43  8  00  S •a5 -S co  35 • 73  ) B O ™ T3 ! 73 « & 3 3 u fc o Q < eso co a , cj  5  CJ  5  S4?  a  I, a  XI CJ  £  3  73  1  X i X ? 43  cn  12  w  cS &  3  CJ  3  .1 i 5  J •3  3  cj  CJ.  3  8 7a  cc  CJ  ^  a aa  .3 oa a 433 cn* O  a a £73cn a  ^  O 3 3 s 8 ga ^  O tn  <J  SSI  CJ  43  a «  e S as X)  CJ  O v-  •5 3 73  "O J i l  fj  104  •If ll  |g c  I  00  3 .3 lH  if  1 l l  o  00  T3  1  i-e  OH  o c/i  i3 60 M  •n  is  10  ri £  d  •s rt  o  4  d  1(  OJD  .5  I  § <L>  a o CO  'I  II  |5 60  1 |oo  |8  WW  -  106  *  .1 CO  fi  II  o  .If 33 bo i  '5  I  160  i * i T3  «§ I g 73 G « O |  ^  i%  B %  i  i  CO  N  ?  #  N  ^ "2 ^  >/J  "Is  o OO ft,  i l l  I § I a .s  o to  i?  1  ! I  co  V  <S  co  S JB  o i  o  00  -a >  u  -a  c  If ^ Jr s jj s21 i „ J  cn 60 to  "3 P 73 "3  T3 jj  rt >  il  B  o  1.8  1  2 60 S SJ  •>-» oj  | & s  O  IS co  ca  CO  •8  1 I  •8 60  i  a § s  rt a  >.  |5 60 M  QP j' a,  73 2  o  •1 —4  CO  3 oo  i  -a  in  d  I  a  s §2 rt 2  —  Ip i^  O  CJ  CJ  •HH  CO  5 45l l  rt  §  73  s rts e  rt  < L >  1  a  CO  is  2  _  S o o o  ass 1-1  t  60  .3 33 bo  107  *  tn  I I  ao  60  I  3  CL)  4<  3  .3  o  " o3 *  .... 1)  si  lH  60  to CA CO  1 4J  CO  oO  4>  CJ  cn  •a  I* 3  $> X 3  o  GO s  g  8 £ •a  eo  3  >>.s  •a  O  •g o  to  60  if  3  I  3  Q. O  0  •e  3  I  u  1 ic5  «s ^1  8ll  tn  •a 00  d  3 CD  1 73 8 O  S o2  »  2  • «  1  XJ  JS I  ^ S* o ^ 3  o3  13  •g  CO  T3  I  o o  60  O 3 60 O O 3  £?|  CJ  43  5 a S3  cd  73 §  oCD  3  &  f  3  J©  es  CD  c/f  6  u  00  ii  .. g s  e  CD  3  §  TO  ger  3  8  60 g  X)  if  'fi o o 60  O «> o >  ,  fa 3  s CA  d  CD  TJ  2 00  d  60  M Q il  3 CD  >  I I  i-H  cd CA  CA |_  CD  O  a « o" g  OO  „  . . <n E  J  cn  108  m  o  ll  m  o  ^ § ^  fe-V  rt  O  ll  ll  A O  c  1 •e A  to  •a  I to  to  J-: to w -A RT <u ca  •a oo  a  <u  — .A rt to  & «8 A  1 O  to  <u (X  I 0)  33 00  109 Thematic Analysis: Identifying Common Triggers The 13 episodes for Counsellor One and the 20 episodes for Counsellor T w o were analyzed further to identify common themes or triggers associated with counsellors' countertransference manifestations. The list o f items used by the counsellors and supervisors to rate problematic client issues, interpersonal styles, and emotions (see Appendix E ) , was employed by the author to reduce the content o f the client stimulus to one or two word descriptors (e.g., physical health, suicide, sexual abuse). This process allowed common triggers to be identified among the episodes. Each episode could have as many descriptors as necessary to explicate the content o f the client stimulus. F o r example, three content themes were identified for session 3 - episode 1 for Counsellor Two: Physical Health, Self-Esteem, and Progress in Therapy - Anxiety. The number o f times a theme or trigger was identified in an episode ranged from one to ten. Table 8 contains the list o f the triggers that emerged for Counsellor One and T w o , along with the corresponding session and episode from which the trigger was identified. Findings o f The O Study The findings o f the Q Study supported the contention that counsellor over and underinvolvement were valid indicators o f countertransference manifestations. Every counsellor response that was rated as either over-involved or under-involved by the judges was preceded by client stimulus (e.g., potential triggers) that were deemed challenging for the counsellors based on multiple sources o f data. Although all the triggers in Table 8 received "indirect confirmation" from the multiple sources o f data, the potential triggers that received "direct confirmation" by the counsellors as problematic were most convincing because they could be clearly linked to events in the client stimulus. Thus, these findings provided preliminary support for a temporal link between countertransference behavioural manifestations and countertransference triggers. Common themes or triggers emerged for both counsellors. These themes were as follows: physical health, mental health, family o f origin, emotional arousal, and death. F o r Counsellor One, the triggers that emerged most frequently were physical health, emotional arousal, and helplessness. F o r Counsellor Two, the triggers that emerged most frequently were physical and mental health, family o f origin, and emotional arousal (particularly anger).  A s w i l l be addressed further in the next chapter, these themes were consistent with the literature describing client issues that counsellors find challenging. Table 8 Potential Triggers Identified in Client Stimulus Counsellor One  Session  Episode(s)  Physical Health  3 4 5 6  1*,2 1 1*,2 1,2  Mental Health - Depression  3 7  2,3* 1*  Mental Health - Suicide  3  3*  Relationships - Family of Origin  3 4 5  3* 2 2  Loss of Control  3 4 5  1* 1 1*  Helplessness  3 4  1*, 2,4 2  Emotional Arousal - Anger  3 5 6  2,3* 1* 1,2  Emotional Arousal - Sad  3 4 6 7  1*, 2,4 2 1 1*  Emotional Arousal - Crying  3 4 6 7  1*, 2,4 2 1 1*  Emotional Arousal - Powerful  3 4 5  4 3 3  Perfectionism  4  1  Death  5 6  3 1*  Counsellor Two  Session  Episode(s)  Physical Health  3 4 5 6 3 4 6  1,4,5 2* 1,3 1 1, 3, 4, 5 2*, 3* 4  Mental Health: personal stress  4 6  2*, 3* 1,2  Relationships - family of origin  3 5 6  1, 2, 3, 5, 6*, 7*, 8 3 3,5  Relationships - boyfriend  3 4 5  2 2*, 3* 4  Emotional Arousal - anger  3 4 5 6  3, 5, 6*, 7* 8 3* 2* 3,5  Emotional Arousal - sad  3 4 6  7*, 8 1* 1  Sexual Abuse  3 5  3, 6*, 7* 2  Medication  4 5 6  1* 1 1  Death  3  5, 7*  Progress in Therapy - anxiety  3 4  1,4 1*  Mental Health: self-esteem  * Triggers that received direct confirmation for each counsellor.  112  Chapter V Discussion Over the last century, theorists, writers, and researchers on countertransference have struggled, debated, and even lamented over how to define and study such a complex, elusive construct. Most psychodynamic writers and therapists do not question the influence of countertransference on the therapeutic process; however, researchers attempting to study the phenomena have been stymied on how to operationalize the construct in a meaningful, measurable manner. Singer and Luborsky (1977) stated: How does one systematically study the core unconscious conflicts of the therapist and the extent to which they are aroused and influence his behaviour in psychotherapy? Taken to an extreme, it would almost have to require an investigator lying hidden under the couch of the patient and the analyst (under the analyst's couch during his own treatment) in order to attempt to analyze the phenomenon in systematic detail. (Singer & Luborsky, 1977, p. 49) Although much has been written in the clinical literature about countertransference, the paucity of empirical research is notable. In particular, empirical research in naturalistic settings operationalizing countertransference in a reliable and valid manner has been lacking. Given the daunting task of trying to "systematically study the core unconscious conflicts of therapists," it is understandable that empirical research has lagged behind in this area. In spite of limited research, the impact of the counsellor's countertransference reactions on the therapeutic process has been widely acknowledged beyond the psychoanalytic community (e.g., Gelso & Hayes, 1998). Overview of Study The primary goal of this research was to determine whether countertransference manifestations, operationalized as over-involvement and under-involvement, could be identified within therapy sessions in a reliable and valid manner. This study focused on behavioural manifestations of countertransference because they were potentially observable and they have been identified as the more problematic aspect of countertransference in the therapeutic relationship. A multiple case study design was employed to intensely analyze the counsellors' behaviour in a naturalistic setting. A generalizability (G) study was first conducted to assess the dependability of the  113  behavioural measure of countertransference and to design the decision (D) study (e.g., how many dyads, sessions, judges would be necessary to achieve a dependable measure of over/under-involvement). The G study also served to train the judges on rating the construct under investigation. The D study implemented the design specifications from the G study to address the issue of reliability (e.g.,firstresearch question). Judges' agreement constituted a reliable rating of over and under-involvement. The issue of validity was more complex, reflecting the fact that countertransference has been a challenging construct to operationalize. The conceptualization of countertransference behavioural manifestations as overinvolvement and under-involvement was derivedfromthe theoretical and empirical literature (e.g., Friedman & Gelso, 2000; Hayes et al., 1998; Wilson & Lindy, 1994). Previous research focused solely on therapists' avoidance behaviours (e.g., Rosenberger & Hayes, 1997) and overlooked negative aspects of therapists' seemingly facilitative behaviours, such as oversupporting or colluding with clients. While this broadened definition was thought to have both clinical and empirical value, over-involvement and under-involvement still only captured part of this complex construct. The Q study investigated whether there was further evidence to support operationalizing countertransference in this manner (e.g., second research question). The Q study gathered evidence that over-involvement and under-involvement were valid indicators of countertransference behaviour by triangulating multiple data sources to identify potential countertransference triggers in the dialogue preceding the behaviour (e.g., content of client dialogue, counsellor session notes, supervision notes, counsellor and supervisors' ratings). Whereas previous research (e.g., Rosenberger & Hayes, 1998) had attempted to assess countertransference originsfirst(e.g., unresolved intrapsychic conflicts), then to identify countertransference triggers based on the origins, and then to predict countertransference manifestations, this study adopted Hayes et al.'s (1998) suggestion to "work backwards" byfirstidentifying countertransference manifestations, then trying to uncover countertransference triggers and origins. My logic was that, if potential triggers were found to precede countertransference behaviour, additional support for the construct could be established. In other words, if the supervisors and/or the therapist identified certain issues as "difficult" for the therapist ~ and the analysis of the therapy sessions indicated that, when the therapist responded in an over-  114 involved or under-involved manner those same issues were addressed by the client ~ then this finding may offer preliminary support that the behavioural observations of countertransference manifestations are, in fact, countertransference. Key Findings: D Study The D study found support for the first research question. Behavioural manifestations of countertransference, operationalized as over-involvement and under-involvement, were able to be identified by independent judges. Before the second research question could be addressed, it was essential tofirstestablish a reliable measure of countertransference manifestations. The results of the G study found that a dependable measure of countertransference behavioural manifestations could be achieved through designing a study that included having three trained judges rate two counsellor-client dyads across eight therapy sessions. The generalizability coefficient for this combination of judges, dyads, and session was .79. The decision (D) study assessed interrater reliability using the intraclass correlation. The intraclass correlations for the three judges rating the counsellors' behaviour in the D study rangedfrom.54 to .84 for counsellor one, and .63 to .84 for counsellor two. The reliability coefficient for all eight sessions was .76 and .79 for Counsellor One and Two, respectively, lending credibility to the utility of the G study. Several factors could have contributed to the low reliability coefficients for the first two sessions of the D study. For example, the training provided to the judges, particularly the third judge, during the G study and prior to commencing to D study could have been insufficient. It is also possible that judges required a few sessions to become familiar with a counsellor's range of response. Interestingly, the judges' agreement dropped again in session eight for both counsellors. Given that the judges rated sessionsfivethrough eight without discussing their ratings with one another, it is possible that consensus was more difficult to maintain withoutfrequentre-calibration. The judges may have been more likely to rate the counsellors' behaviour based on their own subjective responses to the counsellors rather than stringently applying the coding scheme. Overall, the findings suggested that countertransference manifestations could be reliably identified by independent judges. The reliability coefficients were within the range of those found in other studies employing similar rating procedures (e.g., Hayes & Gelso, 1993; and Rosenberger & Hayes, 2002). In order to focus the Q study on those samples of  115  counsellor behaviour that represented over-involvement and under-involvement most reliably, only those sessions with a correlation coefficient above .75 were selected for further analysis in the Q study. For both counsellors, sessions three, four,five,six, and seven were above the cutoff. Significant over and under-involvement episodes in sessions three through seven were identified using a graphing procedure that plotted the counsellors' moving average. This procedure allowed statistically defined deviations from counsellors' "typical" responses to be identified. These deviations represented a reliably indexed movement awayfroman empathic stance into either over or under-involvement. Key Findings: Q Study The Q study found support for the second research question. Evidence that overinvolvement and under-involvement were valid indicators of behavioural manifestations of countertransference was gained through establishing a temporal link between potential countertransference triggers during the session and subsequent countertransference behaviour (e.g., counsellor over or under-involvement.) Multiple sources of data were employed to identify triggers (e.g., content analysis of over/under-involvement episodes, session notes, supervision notes, and counsellor and supervisors' ratings). There appeared to be direct and indirect support for countertransference triggers leading up to counsellor over-involvement and under-involvement. The strongest support for the validity of countertransference triggers were the triggers that received direct support from the various sources of data. Direct confirmation for a therapy event or trigger was determined if the content of the client dialogue (e.g., client stimulus) was directly referred to as problematic by the counsellor in either the session notes, the supervision sessions, or the final ratings by the counsellors and supervisors. Potential countertransference triggers that received indirect support were also worth exploring because they were similar to those that received direct support. Indirect confirmation for a trigger was determined if the content of the client stimulus was identified by the counsellor or supervisors as a difficult issue for the counsellor "in general", but the specific event in the client stimulus was not directly referred to by the counsellor. Although the meaning of the indirect triggers was less clear than the direct triggers, both types were identified in the client stimulus preceding the counsellors' countertransference behaviour. It  116 is possible that the triggers receiving direct support were within the counsellors' conscious awareness, whereas the triggers receiving general support may have been outside of their awareness and may be examples of the counsellor's "blind spots." In addition, both Counsellor One and Two identified moments in the sessions that they described as "going well," whereas the judges rated their behaviour as either over-involved or under-involved and the other data sources provided some evidence that the material was challenging for the counsellor. This discrepancy in perception by the judges and the counsellors regarding "helpful" moments during sessions may also have been examples of counsellors' "blind spot." As Luborsky and Spence (1971) noted, third parties (e.g. the judges) may be more attune to detecting countertransference reactions than the counsellor. For Counsellor One, the themes or potential triggers that emergedfromthe various data sources were as follows: physical health, mental health (depression and suicide), relationships-family of origin, loss of control, helplessness, perfectionism, death, and emotional arousal in the client. The themes that emerged for Counsellor Two included the following: physical health, relationships-family of origin, relationships-ex-partner, mental health (self-esteem, personal stress, anxiety, and sexual abuse), medication, death, progress in therapy-anxiety, and emotional arousal in the client. Study Links to the Literature Recent theory and research has conceptualized countertransference behaviour as therapists' behaviour that deviates from their baseline or general tendencies (e.g., Holmqvist, 2001; Kiesler, 2001). The current study selected episodes of counsellor over-involvement and under-involvement applying this principle. The moving averages graphing technique graphed therapists' response patterns for each session reflecting his or her departuresfroman empathic stance. A cutoff point was determined for each counsellor using one standard deviation above his or her mean. Thus, for each counsellor, deviations above the cutoff point were detected and were proposed as examples of over-involved or under-involved behaviour. The findings of this research provide preliminary support for conceptualizing countertransference behaviour in this manner. Although the method applied in Holmqvist's (2002) study did not allow him to identify reasons for the deviating responses, the current study was able to link deviations in counsellors' responses (e.g., over-involvement and under-involvement) to countertransference triggers.  117  Many of the triggers identified by the author for both counsellors in the current study received support from the literature. Hayes et al. (1998) identified various categories of countertransference origins, triggers, and manifestations. Countertransference origins were defined as areas of unresolved intrapsychic conflict for the therapist that may serve as "blind spots" for the therapist and which can impact the therapeutic relationship if they are triggered. They cited examples of origins as including family issues, therapist's needs and values such as need to help or need for control, therapy issues such as termination, and cultural issues. Countertransference triggers included therapy events that evoked the therapist's unresolved issues. Examples of triggers included the content of the client's material (e.g., death, family of origin, parenting, partner issues), changes in the therapy structure, therapist's perception of the client (e.g., as dependent or hostile), and the emotions (e.g., client expressing negative emotion). Countertransference manifestations included therapists' behaviours, thoughts, or feelings that were a consequence of unresolved issues being triggered for the therapist. Examples of manifestations included "approach" responses by the therapist (e.g., nurturing, identification, positive feelings towards the client), "avoidance" responses by the therapist (e.g., distancing self from the client, boredom or fatigue, disappointment with the client), and negative feelings by the therapist. Many of the instances in which the counsellors' responses were rated as over or under-involvement were preceded by the client discussing issues pertaining to their illness. For example, the themes of physical health, helplessness and lack of control emerged for Counsellor One and those of physical health and medication emerged for Counsellor Two. The clinical and research literature on health care providers working with clients with serious health concerns has clearly documented the stress that these care-givers experience (e.g., Martin & Julian, 1987; Weisman, 1981). "A descriptive interview study revealed that among the more common psychological indications of caregiver stress are depression, grief and guilt, anger, irritability, frustration, over-investment and over-involvement, anxiety and difficulty with decision making helplessness and insecurity (Vachon, cited in Farber, 1994, p. 715). Similarly, the theme "death" emerged as a trigger for both counsellors. Dunkel and Hatfield (1986) stated, "working with a person who is dying challenges unresolved feelings of one's own mortality" (p. 115). Although neither of the clients were facing imminent death,  118  the topic of death arose in the context of the clients' personal experiences (e.g., death dream for Client One and boyfriend's attempted killing and physical deterioration for Client Two), and in the context of significant others' experiences (e.g., mother and aunt's attempted suicide for Client One and brother's accidental death for Client Two). It may have been that the counsellors' helplessness in the face of physical deterioration and/or death was sufficiently anxiety provoking that it aroused countertransferential behaviour in both these counsellors. As Counsellor One noted, the client's feelings of helplessness triggered her own feelings of helplessness and desire to "fix it." Counsellor Two seemed to focus on strategies to assess progress in therapy pertaining to anxiety and was disheartened by the client's "lack of progress." Another trigger that emerged for both counsellors was clients' emotional arousal. Both counsellors appeared to have difficulty sitting with their clients when strong emotions such as sadness or anger were being expressed. Singer and Luborsky (1977) noted that "therapists who have less anxiety and less conflict about their own feelings are not as personally affected by the patient's expression of emotions and are able to deal with their patients more therapeutically, allowing the patient to continue to explore his threatening feelings." (p. 440). Interestingly, Counsellor One's previous supervisor and Counsellor Two's previous and research supervisor rated "anxiety management" lowest on the Countertransference Factor's Inventory (CFI-R; Latts & Gelso, 1996). This subscale contains items such as, "is comfortable in the presence of strong feelingsfromothers"; "is comfortable with him/herself; and "tends not to be trouble by anxiety." It is possible that anxiety management was an area of personal vulnerability for both counsellors that was more easily activated in the presence of strong emotions by others. As Hayes et al (1998) noted, it is not only the client's emotional arousal that serves as a trigger for countertransference behavioural manifestations, but also the counsellor's emotional arousal. For example, Counsellor One reported difficulty processing a dream with her client because she had also woken up that morningfroman intense dream. Counsellor Two experienced difficulty working with the client's issue of sexual abuse. He reported feeling very sad to the point that he wanted to weep and wanting revenge on the brother who abused the client. He also felt out of his element regarding working with the issue of "sexual abuse." Also, Counsellor Two was noticeably thrown off balance by the client's negative  119 response to his question regarding progress in dealing with anxiety in the sessions. He reported his surprise in his session notes, his supervision session, and his final ratings. These examples seem to support that counsellor's emotional arousal may also be a factor contributing to countertransference behaviour. Thefindingsfromthis study provided support for the broader, "totalistic perspective" which defines countertransference as all the therapist's reactions to his or her client, both conscious and unconscious (e.g., Kernberg, 1965). For example, those episodes during the session in which the counsellor reacted as either over-involved or under-involved, and subsequently reported problematic events (e.g., direct triggers) within the session provided support for the "conscious" reactions. The unconscious reactions were less clear. Although countertransference manifestations were identified by independent judges (e.g., not dependent on the counsellors' level of awareness to detect reactions), support for countertransference triggers camefromthe counsellor's self-evaluation from session notes and supervision (e.g., events within the counsellors' awareness). It may be fair to say that, because the counsellors evaluated their behaviour after the session(s), they may have become aware of their countertransferential reactions "after the fact" but not while they were engaging in countertransference behaviour. This idea is supported by the abundance of clinical literature on the topic (e.g. Singer & Luborsky, 1977). It was impossible to ascertain whether the countertransference behaviour and triggers originatedfromunresolved intrapsychic conflicts in the counsellor (e.g., subjective countertransference) or whether their reactions were due to client characteristics (e.g., objective countertransference). For example, Counsellor One acknowledged certain identifications with the client regarding "mother" issues and perfectionism. Similarly, Counsellor Two identified difficulties dealing with anger in his life and difficulties working with health concerns. It may be possible that these triggers originatedfromunresolved intrapsychic conflicts for the counsellors; however, this study was unable to explore this possibility in sufficient depth to warrant such a conclusion. Methodological Contributions of the Study The methodology employed to investigate countertransference manifestations was a major contribution of this research. This current study was able to move beyond the theoretical conceptualizations, anecdotal reports, and analogue research designs that  120  characterize the literature on countertransference, to observe countertransference manifestations as they emerged within actual counselling sessions. The multiple case study research design allowed for intense analysis o f the moment to moment interactions between the counsellors and the clients in a naturalistic setting. In order to have confidence in research results, one requires confidence in how the concepts have been operationalized (e.g., construct validity); as well as confidence in the methods used to study the phenomenon. This study focused on countertransference behaviour because it was potentially observable by trained judges. Building on recent conceptualizations o f countertransference (e.g., Friedman & Gelso, 2000; Wilson & Lindy, 1994), the construct was operationalized as counsellor over-involvement and underinvolvement. A s noted earlier, previous research focused on the counsellor's avoidance behaviour, hence over-looking negative aspects o f counsellors' seemingly "positive" behaviour (e.g., colluding, over-supporting). The broader conceptualization allowed for a wider range o f therapist behaviours to be investigated. Another methodological contribution o f this study was the attention paid to establishing a reliable index o f countertransference. A s noted earlier, a generalizability study was performed to increase the dependability o f the behavioural measure. To date, no other research on countertransference has employed generalizability theory. Consequently, researchers interested in this topic may be able to apply the recommendations from this G study to their own research designs. In addition, the reliability data established in this study w i l l serve as a basis o f comparison for future researchers. Perhaps more importantly, this study also attempted to determine whether the behavioural observations were a valid index o f countertransference. The Q study employed multiple sources o f data to assess countertransference manifestations. Again, previous research has not incorporated multiple sources o f data, such as session notes and supervision sessions, into the research design. A major finding o f this research was that it was able to link countertransference behaviour to potential triggering events during the session The methods employed in this study allowed for this temporal relationship between triggers and behaviour to be investigated, and hence, strengthened the validity o f operationalizing countertransference behaviour as overinvolvement and under-involvement.  121  Finally, the participants in this study (e.g., counsellors, clients, and supervisors) were unaware of the focus of the research until after all the data had been collected. One previous case study requested that the counsellor identify unresolved intrapsychic conflicts prior to commencing the counselling sessions and it is possible that collecting this data pre-session may have influenced the counsellor's behaviour in sessions. By setting up the current study so that the participants were unaware of the topic of interest, as unbiased a sample of countertransference behaviour as possible could be gathered. Individual debriefing sessions with the counsellors and clients were conducted after all the data was collected. Limits of the Study Single-case research designs have received both praise and criticism as a method of conducting counselling research (e.g., Galassi & Gersh, 1993; Heppner, Kivlighan, Wampold, 1992; Kazdin, 1981). On one hand, they are viewed as an excellent means of observing the counselling process as it unfolds; on the other hand, they are criticized for lacking generalizability and having problems with internal validity. The limitations of the current multiple-case research study will be discussed under the heading of "external" and "internal validity." External Validity. A frequent criticism of single-case designs concerns their lack of generalizability. Galassi and Gersh (1992) note that it is commonly assumed that generalizability of research findings is more feasible in large-sample studies than in singlesubject designs; however, this criticism is fallacious. In single-case designs the issue of generalizability or external validity is addressed through the replication of additional cases (Galassi & Gersh, 1992, Hilliard, 1993, Yin, 1989). In other words, single-case research is a category of within subject, or intrasubject, research in which the accumulation of data across cases is avoided and generalization of finding occurs through replication of subsequent cases (Hilliard, 1993). Hence the emphasis is on programmatic research, rather than "one shot deals." Consequently, explicit description of all aspects of the research process is crucial so that accurate replication can occur. Barlow and Hersen (cited in Galassi & Gersh, 1993) described three types of replication: direct, systematic, and clinical. In direct replication the same procedures are replicated with several additional clients; with systematic replication the variable of interest, such as settings or disorders, is altered in following studies; and in clinical replication  122 treatment packages are tested with clients presenting similar behavioural-emotional problems. "Whereas direct replication addresses the reliability o f findings or internal validity, systematic and clinical replications are concerned with generalizability or external validity" (Galassi & Gersh, 1993, p. 527). Although this multiple case study design allowed for the intensive study o f two cases, generalizations to the population o f counsellors in general is limited. A s Galassi and Gersh (1993) noted, systematic and clinical replications are required to increase external validity. In this study, direct replication o f findings across two cases provided initial support for external and internal validity. The research presented in this dissertation was a multiple case study aimed at reliably measuring behavioural countertransference manifestations and investigating their validity in a naturalistic counselling setting. A s a result, the development o f construct validity was important to the generalizability o f this study. Y i n (1989) noted that the development o f construct validation required the use o f measures that truly captured the concepts under investigation. Because this issue o f valid measurement has been challenging in the field o f countertransference, the current research attempted to develop rigourous measures o f behavioural manifestations based on the literature. Consistent with others conceptualizations, countertransference behaviour was operationalized as counsellor over-involvement and under-involvement (e.g., I C B ; Friedman & Gelso, 2001; Wilson & Lindy, 1994). Additional support for operationalizing countertransference manifestations is this manner was established by identifying other factors (e.g., triggers) thought to be related to the phenomenon. This study applied Hayes et al.'s (1998) theory o f countertransference origins, triggers, and manifestations to identify potential countertransference triggers thought to precede the behavioural manifestations. Thus, external validity was not only increased by conducting this research in a naturalistic setting, but also by grounding the study in relevant theory. Internal Validity. Another major criticism o f single-case designs includes threats to internal validity. In order to draw valid inferences from case studies one must be able to rule out alternative rival hypotheses o f the results. Although single-case experimental designs (e.g., A B A B designs) permit superior control and manipulation o f the treatment variable(s), such designs are often not applicable to clinical settings due to ethical and methodological concerns (Kazdin, 1981). For example, withdrawing treatment to return to baseline levels o f  functioning can be extremely distressful for participants. Thus, non-experimental designs are often adopted. However, this concession does not mean that scientific rigour is sacrificed. Consequently, intensive single-case designs should be constructed to have high internal validity (e.g., B o r g & Gall, 1989). Because random assignment and control groups are not feasible, internal validity must be achieved though other design techniques. For example, making reliable observations and repeated measurements are vital components to improve internal validity in single-case designs (Borg & Gall, 1989). In the first situation, reliability can be achieved through careful training o f observers, operationally defining the behaviours to be observed, periodic checks o f observer reliability, and control for observer bias. In the latter instance, it is important to standardize the measurement procedure to minimize the contamination o f treatment effects with measurement effects. In terms o f internal validity, the current study was primary concerned with establishing a reliable measure o f countertransference behaviour. The following steps were taken to ensure reliable observations: one, conducting a generalizability study; two, training judges; three, operationally defining countertransference behaviour as over-involvement and under-involvement; four checking interrater reliability; and five, obtaining independent ratings to minimize rater bias. The current study systematically collected a great deal o f data from multiple sources (counsellor, supervisors, judges) across time. A s Heppner et al. (1992) stated, " . . the objective and subjective data collected from various perspectives allowed for comparisons to be made, and subsequently, for conclusions to be developed based on the convergence o f a wide range o f information rather than a single data point." (p. 172). In other words, by reducing threats to internal validity, this research design allows greater confidence in the research findings. A threat to internal validity arises from operationalizing countertransference as counsellor over-involvement and under-involvement because it reduces the construct to behavioural criteria that are potentially observable, but loses some of the richness o f the phenomena. A s Hayes and Gelso (1991) pointed out: A n inherent difficulty in any piece o f research that deals with countertransference lies in operationalizing the term In this and previous studies, countertransference was  124 operationalized usefully as the therapist's withdrawal o f personal involvement in responding to a client. However, this definition o f countertransference does not capture all possible manifestations o f the construct nor does it share complete agreement with alternative empirical definitions o f countertransference. (p. 290) Although the authors applied a different definition than that used in the current study (e.g., withdrawal o f involvement versus over and under-involvement), the sentiment still applies. It could be argued that countertransference manifestations may best be studied by asking the counsellors to identify their own reactions as they emerge, rather than inferring their reactions from judges' observations. Although counsellors were asked to write, "what went w e l l " and "what was challenging" after each session, and to rate what client issues they were reactive too at the end o f the study, they were not directly asked to reflect on their countertransferential reactions. Because there is an abundance o f anecdotal research describing therapists' self-reported reactions during sessions, the current study focused on "observed counsellor behaviour" in order to obtain as unbiased sample o f countertransference behaviour as possible and to expand methods used to study countertransference. It is acknowledged that by studying counsellor's behaviour, some o f the complexity and nuances o f the construct are potentially over-looked. Future Research A strength o f this research was that it moved beyond analogue research designs to employ actual counselling sessions. In addition, this research incorporated supervision sessions in the research design to include a valuable source o f information regarding countertransference manifestations. It is recommended that future research continue to employ intensive case study research, including the supervision component, to establish a solid base o f knowledge on this topic before developing experimental designs. A s Hayes et al. (1998) noted, "it appears that researchers have jumped the gun somewhat in hypothesizing about and examining factors believed to relate to countertransference without having those hypotheses informed by empirical data from actual therapy" (p. 469). Developing creative ways to access therapist's unresolved unconscious conflicts continues to be a challenge for future research. It is possible that this challenge is unsolvable. The current methodology focused on countertransference behavioural manifestations and  125 potential triggers, but did not fully explore countertransference origins. A n attempt was made to incorporate ratings by others (e.g., previous supervisor and research supervisor) regarding the counsellor's reactivity to various client issues. Because the research supervisor only had contact with the counsellors on two occasions, her insight into their unresolved unconscious conflicts was necessarily limited, but still useful. Rosenberger and Hayes (1998) included ratings by the counsellor's therapist and partner, people who have longer-term contact with the counsellor and may have deeper insight into his or her blind spots. It is recommended that future research conduct in-depth interviews with the counsellor and "significant others" after all the therapy sessions are completed with the explicit purpose o f identifying countertransference origins. B y conducting the interviews at the end o f the study, the data is protected from bias. Although this study did not focus on the effects o f countertransference behaviour on counselling outcome, the B r i e f Symptom Inventory (BSI; Derogatis, 1993) indicated that both clients' pre and post-measure did not change, suggesting they did not improve by the end o f therapy. Interestingly, client one's B S I score dropped at the mid-point measure, whereas client two's score increased. It is difficult to ascertain the degree to which countertransference behaviour impacted treatment outcome in the current study because there is no baseline to compare whether the amount o f countertransference manifested in the sessions was significant. Future research may be able to develop indices o f high, medium, and l o w countertransference using the current methodology to explore the impact o f countertransference manifestations on treatment outcome. This research focused on identifying a reliable and valid method o f measuring counsellor's countertransference behavioural manifestations. In particular, this study explored whether potential countertransference triggers could be identified in the client's behaviour/dialogue leading to the counsellor's subsequent over-involved or under-involved behaviour. Thus, no attempt was made to track the counsellor's over-involved or underinvolved behaviour on the client's subsequent responses. F o r example, some research suggests that when clients act submissive, counsellors act dominant and vice versa (e.g., Singer & Luborsky, 1977). It would be useful for future research to track the reciprocal influence o f the counsellor-client interaction. In addition, it would be interesting to include client ratings o f counsellor's countertransference behaviour as another datum source. For  126 example, the clients could have been asked to identify moments during the session that they felt misunderstood by, distant from, or annoyed with the counsellor. Other exciting and challenging methods to study countertransference include incorporating measures to capture physiological changes in therapist during the session. Some research has monitored therapist's autonomic responses during sessions with interesting results (e.g., Redington & Reidbord, 1992; Reidbord & Redington, 1993). Conducting this kind o f research requires sensitive, sometimes costly equipment, making it out o f reach for many researchers. Implications F o r Practice and Training Although countertransference has been discussed extensively in psychoanalytic journals, the construct has implications for therapists regardless o f their theoretical orientation. The introduction o f an intervention at a particular moment during the therapeutic process may, at times, be influenced by the therapist's reactions to the client during the session. F o r example, the timing o f an analytical interpretation or the introduction o f a behavioural strategy can both be influenced by the therapist's countertransferential reactions to the client. Helping both experienced counsellors and counsellors-in-training pay attention to their reactions to their clients is a vital ingredient in counsellor development and supervision. Consequently, the methodology and results o f this study have direct implications for counsellor training and practice. Operationalizing countertransference as counsellor overinvolvement and under-involvement provides a framework for which supervision could be conducted. Counsellors, along with their supervisors, could review their therapy session videotapes to identify and explore moments o f over-involvement and under-involvement. Because over and under-involvement include primarily behavioural descriptors, the phenomena are more likely to be detected compared with other dimensions o f countertransference. This process could be a valuable springboard to increase counsellor selfawareness regarding unresolved intrapsychic conflicts, as well as toidentify client issues that are particularly challenging for that counsellor that may require further training and continuing education. Summary The primary purpose o f this research was to determine, one, whether independent judges could rate countertransference manifestations, operationalized as counsellor over-  involvement and under-involvement during actual counselling sessions, and two, whether there was evidence to support the contention that over and under-involvement are valid indicators o f countertransference manifestations. A multiple case research design employing methodology to maximize measurement o f countertransference manifestations in a reliable and valid manner was used. The results o f the D study and Q study indicated preliminary support for the research questions. Inter-rater reliability was sufficiently high to support the dependability o f the judges' ratings. In other words, the judges were able to reliably rate the counsellors' behaviour for empathic breaks into over or under-involvement. In addition, potential triggers leading up to counsellor over-involvement or under-involvement were identified for both counsellors in this study. These results provide preliminary support linking countertransference triggers and subsequent behavioural manifestations. Although these findings require replication, operationalizing countertransference behavioural manifestations as over-involvement and under-involvement received validation.  128  Footnotes  ^ t h o u g h the G study determined that only two counsellor-client dyads were required for the D study, an additional dyad was included as a back up in the event that one o f the counsellors or clients withdrew their consent to participate in the study. A l l three dyads completed the eight sessions. One counsellor employed a counselling technique ( E M D R ) in several sessions that did not require much direct dialogue or interaction between the counsellor and the therapist. A decision was made to hold this dyad in reserve as a back up. Because it was not needed as a back up, this case was not analyzed.  129  References  th  Anastasi, A . (1988). Psychological testing. (6  ed). N e w York: Macmillan.  Bandura, A . (1956). Psychotherapist's anxiety level, self-insight, and psychotherapeutic competence. Journal o f Abnormal and Social Psychology. 52, 333-337. Bandura, A . , Lipsher, D . , & Miller, P. (1960). Psychotherapists' approach-avoidance reactions to patients' expressions o f hostility. Journal o f Consulting Psychology. 24(1), 1-8. Blanck, G . , & Blanck, R . (1970). E g o psychology: II. Psychoanalytic developmental psychology. N e w Y o r k : Columbia University Press. Bouchard, M . A . , Normandin, L . , & Seguin, M . H . (1995). Countertransference as instrument and obstacle: A comprehensive and descriptive framework. Psychoanalytic Quarterly. 64, 717-745. Cutler, R. (1958). Countertransference effects in psychotherapy. Journal o f Consulting Psychology. 22, 349-356. Derogatis, L . (1993). B r i e f Symptom Inventory: Administration, scoring, and procedures manual (3rd ed.). Minneapolis, M N : National Computer Systems, Inc. Dube, J. E . , & Normandin, L . (1999). The mental activities o f trainee therapists o f children and adolescents: The impact o f personal psychotherapy on the listening process. Psychotherapy. 36(3), 216-228. Fiedler, F. E . (1951). O n different types o f countertransference. Journal o f Clinical Psychology. 7, 101-107. Freud, S. (1910). The future prospects o f psychoanalytic therapy. Standard edition, i i , 139-151. London: Hogarth Press, 1957. Friedman, S., & Gelso, C . J. (2000). The development o f the Inventory o f Countertransference Behaviour. Journal o f Clinical Psychology. 56. 1221-1235. Galassi, J., & Gersh, T. (1993). Myths, misconceptions, and missed opportunity: Single-case designs and counseling psychology. Gelso, C. J., & Carter, J. (1985). The relationship in counseling and psychotherapy. The Counseling Psychologist, i l , 155-244. Gelso, C . J., & Carter, J. (1994). Components o f the psychotherapy relationship: Their interaction, and unfolding during treatment. Journal o f Counselling Psychology. 41.  130  296-306. Gelso, C. J., Fassinger, R., Gomez, M . , & Latts, M . (1995). Countertransference reactions to lesbian clients: The role o f homophobia, counselor gender, and countertransference management. Journal o f Counseling Psychology. 42(3), 356-364. Gelso, C. J., & Hayes, J. A . (1998). The psychotherapy relationship: Theory, research, and practice. N e w Y o r k : Wiley. Gelso, C . J., & Hayes, J. A . (2002). The management o f countertransference. Manuscript in preparation. Hayes, J. A . (1995). Countertransference in group therapy: Waking a sleeping dog. International Journal o f Group Psychotherapy. 45, 521-535. Hayes, J. A , & Gelso, C . J. (1991). Effects o f therapist-trainees' anxiety and empathy on countertransference behaviour. Journal o f Clinical Psychology. 47, 284-290. Hayes, J. A , & Gelso, C . J. (1993). Counselors' discomfort with gay and H I V infected clients. Journal o f Counseling Psychology. 40, 86-93. Hayes, J. A . , Gelso, C . J., V a n Wagoner, & Diemer, R. (1991). Managing countertransference: What the experts think. Psychological Reports. 69. 139-148. Hayes, J A , McCracken, J., McClanahan, M . , H i l l , C , Harp, J., & Carozzoni, P. (1998). Therapist perspectives on countertransference: Qualitative data in search o f a theory. Journal o f Counseling Psychology. 45, 468-482. Heimann, P. (1950). O n countertransference. International Journal o f Psychoanalysis. 31, 81-84. Hilliard, R . B . (1993). Single-case methodology in psychotherapy process and outcome research. Journal o f Consulting and Clinical Psychology. 61(3), 373-380. Hinshelwood, R. D . (1999). Countertransference. International Journal o f Psychoanalysis. 80, 797-818. Holmqvist, R. (2001). Patterns o f consistency and deviation in therapists' countertransference feelings. Journal o f Psychotherapy Practice Research. 10(2), 104-116. Jones, E . E . (1993). Introduction to special section: Single-case research in psychotherapy. Journal o f Consulting and Clinical Psychology. 61(3), 371-372. Kernberg, O. (1965). Notes on countertransference. Journal o f the American Psychoanalytic Association. 13, 38-56.  131  Kernberg, O. (1999). Acute and chronic countertransference reactions. Kiesler, D. J. (2001). Therapist countertransference: In search of common themes and empirical referents. Psychotherapy in practice. 57(8), 1053-1063. Latts, M., & Gelso, C. (1995). Countertransference behavior and management with survivors of sexual assault. Psychotherapy. 32(3), 405-415. Lindy, J. D. (1989). Transference and post-traumatic stress disorder. Journal of American Academy of Psychoanalysis. 17, 397-403. Little, M. (1951). Countertransference and the patient's response to it. International Journal of Psychoanalysis. 32, 32-40. Maroda, K. J. (1991). The power of countertransference. New York: Wiley. McClure, B. A., & Hodge, R. W. (1987). Measuring countertransference and attitude in therapeutic relationships. Psychotherapy. 24, 325-335. Peabody, S. A., & Gelso, C. J. (1982). Countertransference and empathy: The complex relationship between two divergent concepts in counseling. Journal of Counseling Psychology. 29, 240-245. Reich, A. (1951). On countertransference. International Journal of Psychoanalysis. 32,25-31. Reich, A. (1960). Further remarks on countertransference. International Journal of Psychoanalysis. 41, 389-395. Robbins, S., & Jolkovski, M. (1987). Managing countertransference feelings: An interactional model using awareness of feeling and theoretical framework. Journal of Counseling Psychology. 34(3), 276-282. Rosenberger, E., & Hayes, J. (1998). Through a looking glass somewhat clearly: A case study of countertransference. In J. Hayes (Chair), Exploring the therapist's inner-world: Current research on countertransference. Symposium presented at the 105th convention of the American Psychological Association, Chicago, Illinois, August 1997. Rosenberger, E., & Hayes, J. (2002). Origins, consequences, and management of countertransference: A case study. Journal of counseling psychology. 49(2), 221-232. Shavelson, R. J., & Webb, N. M. (1991). Generalizability theory: A primer. London: Sage Publications. Shavelson, R , Webb, N., & Rowley, G. (1989). Generalizability theory. American  132  Psychologist. 44(6), 922-932. Singer, B., & Luborsky, L. (1977). Countertransference: the status of clinical versus quantitative research. In A. S. Gurman and A. M . Razin (Eds.), Effective psychotherapy: A handbook of research (pp. 433- 451). New York: Pergamon. Slakter, E. (1987). Countertransference. Northvale, NJ: Jason Aronson. Snyder, W. U , & Snyder, B. J. (1961). The psychotherapy relationship. New York: Macmillan. Van Wagoner, S., Gelso, C , Hayes, J., & Diemer, R. (1991). Countertransference and the reputedly excellent therapist. Psychotherapy. 28(3), 411-421. Wilson, J. (1989). Trauma, transformation, and healing: A n integrative approach to theory, research, and post-traumatic therapy. New York: Brunner/Mazel. Wilson, J., & Lindy, J. (1994). Empathic strain and countertransference. In J. Wilson & J. Lindy (Eds), Countertransference in the treatment of PTSD (pp. 5-30). New York: Guilford Press. Winnicott, D. W. (1949). Hate in the countertransference. International Journal of Psychoanalysis. 30, 69-75. Yin, R. K. (1989). Case study research: Design and methods. London: Sage. Yin, R. K. (1994). Case study research. Thousand Oaks, CA: Sage. Yulis, S., & Kiesler, D. (1968). Countertransference response as a function of therapist anxiety and content of patient talk. Journal of Consulting and Clinical Psychology. 32, 414-419.  Appendix A  Generalizability Study  134  G Study The purpose o f the G study was to improve the dependability o f the behavioural measurement o f countertransference manifestations, conceptualized as counsellor overinvolvement and under-involvement, and reduce the amount o f measurement error by using data from the G study to design the decision study ( D study). The G study helped to answer the question, " H o w many counsellor-client dyads, counselling sessions, and judges would be necessary to increase the reliability o f the measure o f countertransference, reduce measurement error and, consequently increase the level o f generalizability?" This task was accomplished by computing variance components for each parameter o f interest in the study (i.e., judge, sessions, therapists, interactions, error). Table 1 presents the estimated variance components from the analysis o f the two therapists using "therapists' countertransference behaviour" as the dependent variable. This table includes the percentage o f total variance attributed by each variable (i.e., therapist dyad, rater, and session.) The variance component for therapists is the universescore variance; it shows the amount o f systematic variability between therapists in their behaviour. The estimated variance component is substantial (o'p = 1 9 4 ; 14% o f total variance). This percentage indicates that therapists differed in their behaviour. The lack o f variance forjudges suggests that raters were in agreement with their ratings o f the counsellors' behaviour. With nested designs it is impossible to separate the session main effect from the interaction between therapists and sessions. However, the negligible variance component for sessions:therapists suggests that neither session produced more behaviour than the other, nor did the relative standing o f the therapists differ from one session to the other. The interaction between judges and sessions, and the three-way interaction between therapists, judges, and sessions, and unmeasured variation are confounded in this design. The large residual component (a js,tjs,e = 1.186; 84% o f the total variance) 2  suggests that the majority o f variation is due to these confounded sources (Shavelson & Webb, 1991). A s noted above, the purpose o f conducting a G study is to identify sources  135  o f unwanted variation so that they can be reduced when designing a D study. "These variance components are the 'stuff out o f which particular measures are constructed for substantive research or decision-making studies." (Shavelson & Webb, 1991, p.83). Table 1 T w o Facet. Partially Nested Random G Study o f Counsellor Behaviour with (s:c) x j Design  Source of  Estimated  Percentage  Mean  Variance  Variance  of Total  Component  Component  Variance  .194  14  Variation  df  Squares  Therapists(t)  1  .04532  a t  JudgesO)  1  121.938  O J  .00003  0  Sessions:Therapists(s:t)  2  9.413  a s,ts  .028  2  tj  1  .02582  0  0  o:tj,e  1200  1.186  1.186  84  2  2 .  a js,tjs,e 2  Table 2 Formulas for Relative and Absolute Error Variance for the Two-Facet. Partially Nested (o:p) x r Design  o Rel = 2  c pr + q o . p o + o ro.pro.e 2  2  n'r  a Abs =  2  n'o  n'r n'o  a r + o pr + a o.po + o ro.pro.e 2  n'r  n'r  2  n'o  2  n'r n'o  T o confirm that you consent to participate in this study and that you have received a copy o f this consent form for your own records, please sign the space provided below.  Participant's Signature  Date  Sincerely,  Elsie De Vita, M A .  Dr. Beth Haverkamp  Appendix B  Counselling Process Training Manual  146  Counselling Process Training Manual  This manual provides standardized guidelines for judges rating counsellor and client interactions collected as part o f a doctoral research project on the counselling process. Constructs to be Rated A s a judge, you w i l l be watching video-tapes and reading corresponding transcripts o f actual counselling sessions. Using the seven-point Likert scale below, you w i l l rate the counsellor's behaviour according to three dimensions: 1 ) counsellor over-  involvement; 2) counsellor empathic connection; and 3) counsellor underinvolvement. These behaviours all pertain to how the counsellor behaves towards the client. The lists o f items on the pages two to four provides descriptors o f the three dimensions to be rated. Please feel free to ask questions or make comments regarding the dimensions.  -3 Underinvolved  -2 -1 0 +1 +2 +3 Somewhat Possibly Empathically Possibly Somewhat Overinvolved Underinvolved Underinvolved Connected Overinvolved Overinvolved Minimal Empathy Minimal Empathy  You will be provided with examples o f these three dimensions from a video-taped counselling session to practice rating. After watching each video segment, rate each therapist response according to the three dimensions using the seven-point Likert scale. Pay attention to the intensity of the therapist's behaviour when making your ratings. I f your ratings are in accordance with the other judge, rate the training tape from start to finish. A s you watch the video, it is easier to first write your ratings on the corresponding transcript next to the therapist's response. Stop the video at five minute intervals (approximately two pages of transcript) to compare your ratings with the other judge. Discrepancies in ratings o f two points or more should be discussed with an attempt to clarify and refine ratings.  147  Once training is completed, proceed to rate the four counselling sessions, two with counsellor A and two with counsellor B . Rating sheets w i l l be provided to record your ratings.  1) Counsellor Under-involvement can be characterized in a variety o f ways: The counsellor.... -treated client in a punitive manner* -was critical o f the client* -spent time complaining* -provided too much structure* -inappropriately questioned the client's motives* -inappropriately took on advising tone* -distanced him/herself from the client* -was apathetic toward the client* -behaved as i f she or he were 'somewhere else'* -behaved as i f she or he were absent* -rejected the client* -redirected client away from emotion and to cognition -body language was "under-attending," seemed avoidant or disinterested (e.g., counsellor looked away often, was constantly fidgeting, or shifting his/her body), -voice tone was "under-attending" (e.g., counsellor's voice tone reflected irritation, boredom, withdrawal).  148  2) Counsellor Empathic Involvement can be characterized in a variety of ways:** The counsellor.... -used facilitating skills such as reflection, summaries, clarification, labelling, and empathy to demonstrate an understanding of the client's experience and willingness to explore issues further, -was warm and caring -appeared genuine in his/her interactions with the client, -shifted the discussion from neutral issues to deeper or "less comfortable" topics, -body language demonstrated respectful attending behaviour and suggested an openness and willingness to "be" with the client (e.g., counsellor maintained appropriate eyecontact, body posture was comfortable and non-distracting), -voice tone was appropriate and audible (e.g, coun(e.g, cocounsellor'or's v voice was pleasant and variable, he/she sounded interested and engaged with the client).  149  3) Counsellor Overinvolvement can be characterized in a variety of ways: The counsellor -seemed to agree too often with the client* -oversupported the client* -befriended the client* -frequently changed the topic* -talked too much* -acted in a submissive way* -inappropriately apologized to the client* -engaged in too much self-disclosure* -colluded with the client* -acted in a dependent manner* -body language was "too much" or was over-attending (e.g., counsellor leaned forward too much, touched the client too soon, etc.). -voice tone his/her voice was "too much" or was over-attending (e.g., counsellor was over-exuberant or sounded overly sympathetic).  *Over-involvement and Under-involvement items takenfromICB (Friedman & Gelso, 2000). **Empathy items are taken from Bandura et al.(1960), Rosenberger and Hayes (1998), and Carl Rogers.  150  Please review each video-tape and corresponding transcript. You are asked to rate whether the counsellor's behaviour demonstrates empathic involvement with the client, which includes a wide range of appropriate counsellor behaviour, or whether the counsellor's behaviour demonstrates departuresfroman empathic stance into areas of under-involvement and/or over-involvement. The dimensions of under-involvement and over-involvement vary in intensity. For example, under-involved behaviour may range from boredom to physically leaving the session. Over-involved behaviour may range from being sympathetic to taking responsibility for the client. Rate the counsellor's behaviour in each talking turn using the 7-point Likert scale below. Please review the definitions of the constructs provided in the manual as a reminder. Record your ratings on the rating sheet provided. As you watch the video, it is easier to first write your ratings on the corresponding transcript next to the counsellor's response. After watching five minutes of the session, approximately two pages of the transcript, stop the video to transfer your ratings to the rating sheet. You can review your ratings at this time and rewind the video if you wish to reconsider ratings. Please feel free to stop the video at any time to review parts of the session.  151  Rate the counsellor's behaviour in each talking-turn using the scale below. -3  Underinvolved  -2 -1 0 +1 +2 +3 Somewhat Possibly Empathically Possibly Somewhat Overinvolved Underinvolved Underinvolved Connected Overinvolved Overinvolved Minimal Empathy Minimal Empathy  1  21  41  61  81  101  121  141  161  2  22  42  62  82  102  122  142  162  3  23  43  63  83  103  123  143  163  4  24  44  64  84  104  124  144  164  5  25  45  65  85  105  125  145  165  6  26  46  66  86  106  126  146  166  7  27  47  67  87  107  127  147  167  8  28  48  68  88  108  128  148  168  9  29  49  69  89  109  129  149  169  10  30  50  70  90  110  130  150  170  11  31  51  71  91  111  131  151  171  12  32  52  72  92  112  132  152  172  13  33  53  73  93  113  133  153  173  14  34  54  74  94  114  134  154  174  15  35  55  75  95  115  135  155  175  16  36  56  76  96  116  136  156  176  17  37  57  77  97  117  137  157  177  18  38  58  78  98  118  138  158  178  19  39  59  79  99  119  139  159  179  20  40  60  80  100  120  140  160  180  Appendix C  Introductory Letters and Informed Consent Documents Letters o f Agreement for Affiliate Professionals and Judges  Appendix D  Demographic Questionnaires  163  Client Background Information Sheet The following questions request some background information. The top half o f the form contains standard information requested o f clients when entering counselling. This information w i l l be viewed solely by the researcher and your counsellor. The bottom half o f the form contains demographic information that w i l l be used for descriptive purposes in the study. Please take a few moments to complete the questions.  Name: Address: Phone: Emergency Contact: (Name & Number) Can a message be left at home?  Yes  No  1. Age:  2. Sex:  M  F  3. Highest level o f education: 4. Ethnic Background: _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 5. Marital Status: 6. D o you have children?  Yes  No  I f Yes, how many:  7. In a few sentences, describe your main concerns, including your health concerns:  164  Counsellor Background Information Sheet The following questions request some background information. The top half of the form requests your phone number and will be used solely by the researcher. The bottom half of the form contains demographic information that will be used for descriptive purposes in the study. Please take a few moments to complete the questions.  Name: Phone:  Can a message be left at home?  Yes  No  1. Age:  2. Sex:  M  F  3. Highest level of education: 4. Specialization: 5. Ethnic Background: 6. Marital Status: 7. Do you have children? Yes  No  If Yes, how many:  8. Years and months counselling experience: 9. What is your theoretical orientation is counselling?  10. What types of client issues/populations do you have the most experience with?  Appendix E Countertransference Triggers Rating Sheets for Counsellors and Supervisors  166 Counsellor Questionnaire To B e Completed After A l l Sessions A r e Finished  Please respond to the following questions based not only on your experience with this client, but also on your experience in general with clients. For example, based on your personal experiences and life history, you may find that you are more sensitive to certain issues but not to others. Please be as candid as possible. Based on your self-assessment, please use the five point Likert scale to rate your level o f difficulty dealing with, or reactivity to, the following client issues, presentation styles, and emotions.  0 N o t at A l l Difficult  A. Relationships 1. Parents 2. Siblings 3. Partner 4. Ex-partner 5. Children 6. Step-children 7. Friends 8. Extra-marital 9. Co-workers 10. Therapists 11. Other  B. Mental Health 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22.  Depression Anxiety Suicidal A / B Sexual Abuse Substance Misuse Sexuality Self-esteem Chronic Pain Anger Management Personal Stress Other I  1 Slightly Difficult  2 Somewhat Difficult  3 Quite Difficult  4 Extremely Difficult  C. Physical Health 23. Cancer 24. HIV/AIDS 25. Diabetes 26. Heart Disease 27. Minor Complications 28. Weight Issues 29. Other D. Devel. Period 30. Infancy 31. Childhood 32. Adolescence 33. Adulthood 34. Present E. Misc. 35. Education 36. Divorce 37. Death 38. Other Client Presentation 39. Assured/Dominant 40. Gregarious/Extravert 41. Warm/Agreeable 42. Unassuming/lngenu. 43. Unassured/Submis. 44. Aloof/Introverted 45. Cold-Hearted 46. Arrogant/Calculating  Client Emot ons 47. Mad 48. S a d 49. Scared 50. Joyful 51. Powerful 52. Peaceful 53. Other [  168  Counsellor Questions To Be Completed After All Sessions Are Finished  Are there any issues you are less effective working with that have not been accounted for in the questionnaire?  Was there anything about this client's experience that you found difficult to work with?  How were you different with this client compared to how you usually are?  169 Supervisor Questionnaire Based on your knowledge o f the counsellor participating in this research study, please use the five point Likert scale to rate the counsellor's level o f difficulty dealing with, or reactivity to, the following client issues, presentation styles, and emotions.  0 Not at A l l Difficult  1 Slightly Difficult  A. Relationships 1. Parents 2. Siblings 3. Partner  '  4. Ex-partner 5. Children 6. Step-children 7. Friends | 8. Extra-marital 9. Co-workers 10. Therapists 11. Other  B. Mental Health 12. Depression 13. Anxiety 14. Suicidal A / B 15. Sexual Abuse 16. Substance Misuse 17. Sexuality 18. Self-esteem 19. Chronic Pain 20. Anger Management 21. Personal Stress 22. Other  I  2 Somewhat Difficult  3 Quite Difficult  4 Extremely Difficult  170 C. Physical Health 23. Cancer 24. HIV/AIDS 25. Diabetes 26. Heart Disease 27. Minor Complications 28. Weight Issues 29. Other D. Devel. Period 30. Infancy 31. Childhood 32. Adolescence 33. Adulthood 34. Present E. Misc. 35. Education 36. Divorce 37. Death 38. Other Client Presentation 39. Assured/Dominant 40. Gregarious/Extravert 41. Warm/Agreeable 42. Unassuming/lngenu. 43. Unassured/Submis. 44. Aloof/Introverted 45. Cold-Hearted 46. Arrogant/Calculating  Client Emotions 47. Mad 48. S a d 49. Scared 50. Joyful 51. Powerfu 52. Peaceful 53. Other G i v e n what you know about the counsellor's personal history and counselling skill, are there any other client issues and characteristics not already identified that he or she may be sensitive or reactive to i n a counselling setting? Please explain.  171  Appendix F  Countertransference Factors Inventory - Revised  Form-T  The therapist:  Strongly Disagree  Disagree  1. recognizes similarities betwee?. ^srme ar;«! former clients. 2. has a stable sense of identity.  2  3. is often aware of personal areas of unresolved conflict.  2  4. usually restrains him/herself from excessively identifying with the client's conflicts. 5. accurately labels client's emotions. 6. is often aware of feelings in him/her elicited by clients. 7. understands the background factors in his/her life that have shaped his/her personality. 8. tends to resolve his/her own emotional conflicts. 9. is usually emotionally "in tune" with clients. 10. at the appropriate times, stands back from a client's emotional experience and tries to understand what is going on with the client. 11. effectively sorts out how his/her feelings relate to client's feelings. 12.  often sees things from the client's point of  view. 13. conceptualizes relationship dynamics in terms of the client's past. 14. is comfortable in the presence of strong feelings from others.  Not Sure  Agree  Strongly Agree  Form-T  The therapist:  Strongly Disagree  Disagree  Not Sure  15. possesses a "gut-level" self-understanding.  3  16. is usually able to conceptualize client dynamics or issues clearly.  3  17. intuitively understands clients.  3  18. is often aware of his/her personal impact on others.  3  19. at the appropriate times, puts aside his/her intellect and "feels" with the client. 20. does not experience a great deal of anxiety while conducting therapy. 21. effectively distinguishes between client's needs and his/her own needs. 22. can usually apply a theoretical orientation to cases. 23. is often aware of fantasies in him/her triggered by client material or affect. 24. usually comprehends how his/her feelings influence him/her in the therapy. 25. can usually identify dynamics of the counseling relationship. 26. recognizes the limits of his/her clinical competencies. 27. feels confident working with most clients. 28. is psychologically balanced. 29. has a sense of autonomy. 30. is usually able to assess the severity of client issues.  Agree  Strongly Agree  Form-T  The therapist:  Strongly Disagree  Disagree  Not Sure  31. conceptualizes cases deeply.  3  32. can usually identify with the client's inner experience.  3  33. will reformulate an initial diagnosis if warranted by client material. 34. has the capacity to stand back from his/her own emotional experience and observe what is going on with him/herself. 35. gets beyond the manifest content to the latent meanings of a client's verbalizations. 36. usually recognizes his/her own negative feelings. 37. is comfortable with him/herself. 38. is comfortable being close to others. 39. often uses his/her past experiences to aid in understanding the client. 40. is willing to consider him/herself as an impediment to client progress. 41. does not become overly anxious in the presence of most client problems. 42. reflects deeply on his/her own feelings. 43. effectively recognizes the boundaries between self and others. 44. possesses self-confidence.  Agree  Stronpl Agrc  Form-T  The therapist:  Strongly Disagree  Disagree  Not Sure  Strongly Agree Agree  45. is perceptive in his/her understanding of clients.  1  2  3  4  5  46. usually manages his/her need for approval.  1  2  3  4  5  47. usually connects strands of the client's material.  1  2  3  4  5  48. effectively judges a client's readiness to explore particular issues.  1  2  3  4  5  49. tends not to be troubled by anxiety.  1  2  3  4  5  50. often conceptualizes his/her role in what transpires in the counseling relationship.  1  2  3  4  5  Appendix G  Study Announcements  Appendix H  B r i e f Symptom Inventory  182  INSTRUCTIONS: O n t h e next p a g e is a list o l p r o b l e m s p e o p l e s o m e t i m e s h a v e . P l e a s e r e a d e a c h o n e c a r e f u l l y , a n d b l a c k e n the c i r c l e t h a t b e s t d e s c r i b e s H O W M U C H THAT P R O B L E M H A S DISTRESSED OR B O T H E R E D Y O U D U R I N G T H E P A S T 7 D A Y S I N C L U D I N G T O D A Y . B l a c k e n the c i r c l e for o n l y o n e n u m b e r l o r e a c h p r o b l e m a n d d o not s k i p a n y i t e m s . If y o u c h a n g e y o u r m i n d , e r a s e y o u r first m a r k c a r e f u l l y . R e a d t h e e x a m p l e b e l o r e b e g i n n i n g , a n d it y o u h a v e a n y q u e s t i o n s p l e a s e a s k thern now.  183  J?/  HOW MUCH WERE YOU DISTRESSED BY:  V  1  0  1  2  3  • 4  2  0  1  2  3  4  N e r v o u s n e s s or s h a k i n e s s inside Faintness or dizziness  4  T h e idea that s o m e o n e else can control your thoughts  4  Feeling others are to blame for most of your troubles Trouble remembering things  3  4  2  3  4  Feeling easily a n n o y e d or irritated Pains in heart or chest  1  2  3  4 '  1  2  3.  4  1  2  3  4'  (J  1  2  3  12  0  1  2  13  0  1  14  0  15  3  0  1  2  3  4  . 0  1  2  3  5  0  1  2  3  6  0  t  2  7  0  1  8  0  9  0  10  0  11  I 1I I  / *?/J?/ <&/ / *V &/ / / / */ V / .  Feeling afraid in o p e n s p a c e s or on the streets Thoughts of ending your life  _ mm mm mm mm mm mm  4  Feeling that most people cannot be trusted Poor appetite  3  4  Suddenly scared for no reason  mm  2  3  4  Temper outbursts that you could not control  mm  1  2  3  4  0  1  2  3  4  16  0  1  2  3  4  Feeling lonely e v e n when you are with people Feeling blocked in getting things done Feeling lonely  17  0  1  2  3  4  Feeling blue  18  0  1  2  3  4  19  0  1  2  3  4  Feeling no interest in things Feeling fearful  20  0  1 "  2  3  4  Your feelings being easily hurt  mm  21  0  1  2  3  4  Feeling that people are unfriendly or dislike you  mm  22  0  1  2  3  •1  Feeling inferior to others  mm  23  0  1  2  3  4  N a u s e a or upset stomach  mm  mm  mm mm mm'  mm mm mm  24  0  1  2  3  4  Feeling that you are watched or talked about by others  mm  25  I)  1  ?  3  4  Trouble falling asleep  mm  26  0  1  *  3  1  mm  27  0  1  2  3  J'"  28  0  1  2  3  4  Having to check a n d double-check what you do Difficulty making decisions Feeling afraid to travel on b u s e s , subways, or trains  29  0  1  2  3  4  Trouble getting your breath  mm  30  0  1  2  3  4  mm  31  0  1  2  3  a  Hot or cold spells Having to avoid certain things, places, or activities b e c a u s e they frighten you  32  0  1  2  3  33  0  1  2  3  34  0  1  2  3  mm mm  —  Your mind going blank 4  N u m b n e s s or tingling in parts of your body T h e idea that you should be punished for your sins  35  0  1  2  3  4  36  0  1  2  3  4  Feeling hopeless about the future Trouble concentrating  37  0  1  2  3  4  Feeling weak in parts of your body  38  0  1  2  3  4  Feeling tense or k e y e d up  —  39  n  1  2  3  4  40  D  1  2  3  4  Thoughts of death or dying Having urges to beat, injure, or harm s o m e o n e  mm  41  0  1  2  3  4  Having urges to break or s m a s h things  mm  42  Q  1  2  3  4  Feeling very self-conscious with others  mm  43  0  1  2  3  4  Feeling uneasy in crowds, s u c h a s shopping or at a movie  mm  44  0  1  ^  3  Never feeling close to another person  mm  45  0  1  2  3  4  Spells of terror or panic  mm  46  0  I  2  3  4  Getting into frequent  1  2  3  2  3  —  arguments  4  Feeling nervous when you are left alone  4  Others not giving you proper credit for your achievements  3  4.  Feeling so reslless you couldn't sit still  mm  3  4  Feelings of worthlessness  —  2  3  4  Feeling that people will take advantage ol you il you let them  2  3  4  Feelings of guilt  2  3  4  T h e idea that something is wrong with your mind  47  0  48  0  49  0  1  2  50  0  1  2  51  0  >  52  0  1  53  0  •  Appendix I  Session Notes  185  Counsellor To Be Completed After Each Session and Faxed to Dr. Scalzo. Return original to Elsie De Vita with video tape. Session Notes: Content and Process  Use other side of page if necessary. From your perspective, what went well during the session and what was difficult during the session? Can you recall any specific moments, exchanges, or client statements that were particularly significant to your feeling effective? To your feeling less effective?  Use other side of page if necessary.  Appendix J  pervision Questions  187  Registered Psychologist Questions To Be Asked During Supervision  During thefirstsupervision session, please address the following general questions: 1) How do youfindthe sessions to date? 2) What seems to be going well during the sessions? 3) Are there any client issues that are posing a challenge for you? During thefinalsupervision session, please address the following general questions: 1) How did youfindthe sessions overall? 2) What worked well between you and the client? 3) Were there any client issues that posed a challenge for you?  Appendix K  Frequency Distributions  189 Frequency of Judges Rating of Over-involvement/Under-involvement Across Counsellors and Therapy Sessions Counsellor 1: Session 1 N= 101  Rater 1  Rater 2  Rater 3  Under-Involved (-3)  -  1  -  Somewhat Under-Involved (-2)  2  3  2  Possible Under-involvement (-1)  39  14  10  Empathic Involvement (0)  50  64  71  Possible Over-Involvement (+1)  10  15  18  Somewhat Over-Involved (+2)  -  4  -  Over-Involved (+3)  —  —  —  Counsellor 1: Session 2 N = 204  Rater 1  Rater 2  Rater 3  Under-Involved (-3)  -  1  2  Somewhat Under-Involved (-2)  8  19  7  Possible Under-Involvement (-1)  32  25  32  Empathic Involvement (0)  71  130  122  Possible Over-Involvement (+1)  85  28  39  Somewhat Over-Involved (+2)  8  1  2  Over-Involved (+3)  —  —  —  Counsellor 1: Session 3 N= 128  Rater 1  Rater 2  Rater 3  Under-Involved (-3)  -  -  1  Somewhat Under-Involved (-2)  10  12  16  Possible Under-involvement (-1)  44  35  28  Empathic Involvement (0)  40  56  62  Possible Over-Involvement (+1)  22  12  17  Somewhat Over-Involved (+2)  12  13  4  Over-Involved (+3)  —  —  —  190 Counsellor 1: Session 4 N = 122  Rater 1  Rater 2  Rater 3  Under-Involved (-3)  -  -  -  Somewhat Under-Involved (-2)  4  1  4  Possible Under-involvement (-1)  27  23  16  Empathic Involvement (0)  46  79  68  Possible Over-Involvement (+1)  35  15  29  Somewhat Over-Involved (+2)  10  4  5  Over-Involved (+3)  —  —  —  Counsellor 1: Session 5 N= 167  Rater 1  Rater 2  Rater 3  Under-Involved (-3)  -  -  -  Somewhat Under-Involved (-2)  -  5  -  Possible Under-involvement (-1)  30  16  18  Empathic Involvement (0)  85  114  113  Possible Over-Involvement (+1)  47  25  31  Somewhat Over-Involved (+2)  5  7  5  Over-Involved (+3)  —  —  —  Counsellor 1: Session 6 N = 144  Rater 1  Rater 2  Rater 3  Under-Involved (-3)  -  -  -  Somewhat Under-Involved (-2)  -  -  -  Possible Under-involvement (-1)  3  1  2  Empathic Involvement (0)  69  90  116  Possible Over-Involvement (+1)  61  41  24  Somewhat Over-Involved (+2)  11  Over-Involved (+3)  —  11  2  1  —  191 Counsellor 1: Session 7 N= 134  Rater 1  Rater 2  Rater 3  -  -  Possible Under-involvement (-1)  16  22  6  Empathic Involvement (0)  69  55  95  Possible Over-Involvement (+1)  42  46  32  Somewhat Over-Involved (+2)  7  10  1  Over-Involved (+3)  —  —  —  Under-Involved (-3) Somewhat Under-Involved (-2)  Counsellor 1: Session 8 N = 227  Rater 1  1  Rater 2  -  Rater 3  -  -  -  8  4  4  Empathic Involvement (0)  120  160  187  Possible Over-Involvement (+1)  98  56  31  1  7  5  —  —  Under-Involved (-3) Somewhat Under-Involved (-2) Possible Under-involvement (-1)  Somewhat Over-Involved (+2) Over-Involved (+3)  —  Frequency of Judges Rating of Over-involvement/Under-involvement Across Counsellors and Therapy Sessions Counsellor 2: Session 1 N = 383  Rater 1  Rater 2  Rater 3  Under-Involved (-3)  -  -  -  Somewhat Under-Involved (-2)  8  10  10  Possible Under-involvement (-1)  43  42  36  Empathic Involvement (0)  99  115  195  Possible Over-Involvement (+1)  190  199  129  Somewhat Over-Involved (+2)  40  15  13  Over-Involved (+3)  3  2  —  Rater 1  Rater 2  Rater 3  Counsellor 2: Session 2 N = 432 Under-Involved (-3)  -  -  1  Somewhat Under-Involved (-2)  8  15  8  Possible Under-Involvement (-1)  76  52  42  Empathic Involvement (0)  76  147  260  Possible Over-Involvement (+1)  259  202  90  Somewhat Over-Involved (+2)  13  15  31  Over-Involved (+3)  —  1  —  Counsellor 2: Session 3 N = 362  Rater 1  Rater 2  Rater 3  -  9  3  Somewhat Under-Involved (-2)  45  49  40  Possible Under-Involvement (-1)  127  119  76  Empathic Involvement (0)  81  112  139  Possible Over-Involvement (+1)  99  63  95  Somewhat Over-Involved (+2)  9  8  7  Over-Involved (+3)  1  2  2  Under-Involved (-3)  193 Counsellor 2: Session 4 N = 22  Rater 1  Rater 2  Rater 3  Under-Involved (-3)  -  -  -  Somewhat Under-Involved (-2)  2  6  15  Possible Under-Involvement (-1)  56  51  53  Empathic Involvement (0)  85  103  94  Possible Over-Involvement (+1)  70  47  57  Somewhat Over-Involved (+2)  16  22  9  Over-Involved (+3)  —  —  1  Counsellor 2: Session 5 N = 295  Rater 1  Rater 2  Rater 3  Under-Involved (-3)  -  -  1  Somewhat Under-Involved (-2)  12  20  16  Possible Under-Involvement (-1)  90  86  77  Empathic Involvement (0)  91  124  180  Possible Over-Involvement (+1)  92  54  15  Somewhat Over-Involved (+2)  10  11  6  Over-Involved (+3)  —  —  —  Counsellor 2: Session 6 N = 261  Rater 1  Rater 2  Rater 3  2  3  -  Somewhat Under-Involved (-2)  19  27  10  Possible Under-Involvement (-1)  23  25  93  Empathic Involvement (0)  37  104  107  Possible Over-Involvement (+1)  142  66  39  Somewhat Over-Involved (+2)  37  36  11  1  —  1  Under-Involved (-3)  Over-Involved (+3)  194 Counsellor 2: Session 7 N = 237  Rater 1  Rater 2  Rater 3  Under-Involved (-3)  -  -  -  Somewhat Under-Involved (-2)  4  3  3  Possible Under-Involvement (-1)  53  38  29  Empathic Involvement (0)  116  169  159  Possible Over-Involvement (+1)  56  25  40  Somewhat Over-Involved (+2)  8  2  6  Over-Involved (+3)  —  —  —  Counsellor 2: Session 8 N = 212  Rater 1  Rater 2  Rater 3  -  -  --  Somewhat Under-Involved (-2)  40  37  12  Possible Under-Involvement (-1)  49  39  51  Empathic Involvement (0)  40  46  120  Possible Over-Involvement (+1)  72  79  23  Somewhat Over-Involved (+2)  10  10  5  1  1  1  Under-Involved (-3)  Over-Involved (+3)  195  Appendix L  M o v i n g Averages Graphs: Three Judges  Graph 11 Moving Averages: Three Judges 2.5  n i >  ,1, nn  \  \  j\  Judge 1  "~ i  Judge 2 -4  3.00 21.00 39.00 57.00 75.00 93.00 111.00 12.00 30.00 48.00 66.00 84.00 102.00120.00 Talking Turn Counsellor 1 Session 3 Cutoff - 1.05  Judge 3  Graph 12 Moving Averages: Three Judges 2.5 2.0  !'\  1.5  Judge 1  1.0-  • /• l-^T'-!' l/> ^ V / ' r \l ft  .50.0  i . o , : i f l : . _ .  3.00  f  \Ji  f  W  V  *  21.00 39.00 57.00 75.00 93.00 111.Oo'  12.00 30.00 48.00 66.00 84.00 102.00120.00 Talking Turn Counsellor 1 Session 4 Cutoff = 1.05  Judge 2 Judge 3  198  Graph 13 c  Moving Averages: Three Judges  CD  E CD > O >  c  "__ CD T3  C  Z) "v_; CD >  o  S c  CO  3.00 27.00 51.00 75.00 99.00 123.00147.00 15.00 39.00 63.00 87.00 111.00135.00159.00  Talking Turn Counsellor 1 Session 5 Cutoff = 1.05  199  Graph 14 Moving Averages: Three Judges  c E fl) > o> c &L  a> •o c  I t  -  A  i  L  Z)  Judge 1 Judge 2  >  hfil \r W v?' \ 1  o  c co  12.00 30.00 48.00 66.00 84.00 102.00 120.00 138.00  Talking Turn Counsellor 1 Session 6 Cutoff = 1.05  Judge 3  200  Graph 15 c CO  E CD >  Moving Averages: Three Judges 2.0  T  O > c *L  CD TJ C  Z>  CD >  o  CO CD  3.00  21.00  12.00  39.00  30.00  57.00  48.00  66.00  Talking Turn Counsellor 1 Session 7 Cutoff = 1.05  75.00  93.00  84.00  111.00 129.00  102.00 120.00  201  Graph 16 c CD  E CD >  Moving Averages: Three Judges  O >  iL CD TJ  c  Z> CD >  o  e to  CD  3.00 53.00 103.00 153.00 203.00 253.00 303.00 353.00 28.00 78.00 128.00 178.00 228.00 278.00 328.00 Talking Turn Counsellor 2 Session 3 Cutoff =1.27  202  Graph 17 C CD  E  CD > 2.5  Moving Averages: Three Judges  n  O >  c  ll CD TJ C D  cu > o {= CO CD  1.0  Judge 1 Judge 2  .5 0.0  /on 21.00  JTW? ill?! 57.00  93.00  Talking Turn Counsellor 2 Session 4 Cutoff = 1.27  129.00  165.00  201.00  Judge 3  203  Graph 18 c  03  E  Moving Averages: Three Judges  Qi > O >  c  vL <D  TJ C Z>  t:  Judge 1  > O c  Judge 2  <D Judge 3 21.00  57.00  93.00 129.00 165.00 201.00 237.00 273.00  Talking Turn Counsellor 2 Session 5 Cutoff = 1.27  204  -  Graph 19  |  Moving Averages: Three Judges  Judge 1 Judge 2 Judge 3 3.00  39.00  21.00  75.00  57.00  111.00 147.00 183.00 219.00 255.00  93.00  129.00 165.00 201.00 237.00  Talking Turn Counsellor 2 Session 6 Cutoff = 1.27  205  -  Graph 20  I  Moving Averages: Three Judges  Judge 1 Judge 2 +-  V  1  3.00  ''  '  '  39.00  21.00  ,  '  75.00  57.00  1,1'  ji  I)  111.00  93.00  147.00  129.00  Talking Turn Counsellor 2 Session 7 Cutoff = 1.27  „  |)  t  )<  1) ,»  183.00  165.00  ,1'. j  219.00  201.00  Judge 3  Appendix M  Ethical Approval Certificates  Appendix N  Debriefing Information  210  Client and Counsellor Debriefing Information  A s outlined in the original consent form, this study is exploring the counsellorclient interaction across brief-term counselling. The focus o f the study is on the counsellor's behaviour. A l l counselling sessions can be characterized by a continuum o f counsellor movement towards and away from clients. I am interested in whether this movement can be paired with session content and whether it may be characterized as counsellor over-involvement or underinvolvement in some instances. Some research suggests that when counsellors are emotionally triggered during sessions, their ability to remain empathically connected to the client is challenged. It is during these moments that they may behave in an over-involved or under-involved manner towards their clients. A variety o f things during sessions can trigger counsellors. For example, clients may be talking about issues that are difficult for counsellors to deal with. Sometimes the clients and/or their issues remind the counsellors o f someone they know. Often clients' pain and suffering may trigger a reaction in the counsellors. These reactions can be mild or strong, or even outside our awareness. Being able to identify these moments when counsellors are "reacting" has important implications for supervision and counsellor development. It is the hope o f this study to contribute to the field o f research that investigates counsellor development. To Counsellor: A r e you surprised that this was the focus o f the study? D o you recall experiencing any strong reactions to your client? Would you like to discuss them?  To Client: A r e you surprised that this was the focus o f the study? D o you recall sensing any strong reactions like this from your counsellor? W o u l d you like to discuss them?  

Cite

Citation Scheme:

    

Usage Statistics

Country Views Downloads
United States 35 0
China 27 1
Germany 15 21
United Kingdom 10 0
Sweden 4 0
Japan 3 0
France 2 0
Taiwan 1 2
South Africa 1 0
Canada 1 1
City Views Downloads
Unknown 24 24
Beijing 19 0
Ashburn 13 0
Shenzhen 8 1
Wilmington 5 0
Newtown 4 0
Washington 4 0
Billingham 3 0
Tokyo 3 0
San Diego 3 0
Coventry 2 0
Redmond 2 0
Headley 2 0

{[{ mDataHeader[type] }]} {[{ month[type] }]} {[{ tData[type] }]}
Download Stats

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0054557/manifest

Comment

Related Items