Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The impact of protective perfectionistic self-presentation on group psychotherapy process and outcome Flynn, Carol Ann 2001

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

Item Metadata

Download

Media
831-ubc_2001-610896.pdf [ 9.2MB ]
Metadata
JSON: 831-1.0090569.json
JSON-LD: 831-1.0090569-ld.json
RDF/XML (Pretty): 831-1.0090569-rdf.xml
RDF/JSON: 831-1.0090569-rdf.json
Turtle: 831-1.0090569-turtle.txt
N-Triples: 831-1.0090569-rdf-ntriples.txt
Original Record: 831-1.0090569-source.json
Full Text
831-1.0090569-fulltext.txt
Citation
831-1.0090569.ris

Full Text

THE I M P A C T OFPROTECTIVE PERFECTIONISTICSELF-PRESENTATION O N GROUP PSYCHOTHERAPY PROCESS A N D O U T C O M E  by CAROL ANN FLYNN B.Sc, The University of Toronto, 1994 M.A., University of British Columbia, 1996  A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE STUDIES DEPARTMENT OF PSYCHOLOGY CLINICAL PROGRAMME  We accept this thesis as conforming to the required standard  THE UNIVERSITY OF BRITISH COLUMBIA April 2001 © Carol Ann Flynn, 2001  In presenting degree  this  at the  thesis  in  partial fulfilment  of  University of  British Columbia,  I agree  freely available for reference copying  of  department publication  this or of  and study.  thesis for scholarly by  this  his  or  her  f^rJv^kn^.  The University of British Columbia Vancouver, Canada  Date  DE-6 (2/88)  requirements that the  I further agree  purposes  representatives.  may be It  thesis for financial gain shall not  permission.  Department of  the  that  advanced  Library shall make it  by the  understood be  an  permission for extensive  granted  is  for  that  allowed without  head  of  my  copying  or  my written  PSP in Group Psychotherapy/page 11  ABSTRACT Recent studies reported that perfectionism was linked to negative outcome for all interventions in the NIMH Treatment of Depression Collaborative Research Program. Specifically, perfectionism impeded progress following the eighth session of psychotherapy. The present study investigated the impact of perfectionistic self-presentation (the need to appear to be perfect) on members' experience of a short-term interpersonal therapy group. Individuals who score highly on protective forms of perfectionistic self-presentation avoid being seen as imperfect or foolish and avoid admitting to any mistakes or shortcomings. Therefore, it was anticipated that protective forms of perfectionistic self-presentation would be associated with greater therapy-related stress as assessed by self-report of anxiety, salivary Cortisol levels (a physiological indicator of stress), and behavioural indicators such as tardiness, missed sessions, and lack of self-disclosure during therapy. In addition, a mediational model was proposed in which protective forms of perfectionistic self-presentation may lead to discomfort during the process of therapy, which, in turn, results in poor therapy outcome. Therapy outcome was evaluated using measures of trait perfectionism, psychological symptomatology, stress reactivity, and ratings of benefit by group members, therapists, and friends or family members of each group member. Results tended to support initial hypotheses, as well as past research findings suggesting that perfectionists experience greater difficulties later in the course of therapy. In particular, concerns about disclosing imperfections were associated with increased post-session anxiety late in therapy, missing sessions without calling to provide an excuse, and for women, a relative decrease in disclosure over the course of therapy compared to other participants. Nondisclosure of imperfections also predicted less improvement in depression scores and skin conductance stress reactivity following treatment. Finally, the impact of nondisclosure of imperfections on negative therapy outcome was partially mediated by elevated anxiety late in therapy, and for women, lack of increase in disclosure over the course of therapy. Thus, this study provides evidence of the adverse impact of perfectionistic self-presentation on clients' affect and engagement within the therapy context, suggesting two mechanisms that may lead perfectionists to experience poor therapy outcome.  PSP in Group Psychotherapy/page 111 TABLE OF CONTENTS ABSTRACT  ii  LIST OF T A B L E S  v  LIST OF FIGURES DEDICATION ACKNOWLEDGEMENTS INTRODUCTION  vii  viii ix 1  A N O V E R V I E W OF S E L F - P R E S E N T A T I O N  3  PERFECTIONISTIC S E L F - P R E S E N T A T I O N  5  PERFECTIONISM IN T H E INTERPERSONAL C O N T E X T  7  Interpersonal Styles Intimate Relationships Perfectionism in the Therapy Context  7 11 13  W H Y GROUP PSYCHOTHERAPY?  22  T H E R A P E U T I C F A C T O R S IN G R O U P P S Y C H O T H E R A P Y  25  Interpersonal Learning/Self-Disclosure Catharsis and Emotional Experience Attendance and Attrition  27 31 32  O V E R V I E W OF P R I M A R Y OBJECTIVES  34  HYPOTHESES  36  METHOD  38  PARTICIPANTS MEASURES  Predictor Variables Dependent Variables Control Variable  38 40  40 41 47  THERAPY FORMAT  47  PROCEDURE  49  Assessment Therapy Post-therapy Assessment STATISTICAL A N A L Y S E S RESULTS  49 51 51 52 54  HYPOTHESIS 1: DISTRESS D U R I N G T H E R A P Y  55  HYPOTHESIS 2 : T H E R A P Y O U T C O M E  62  G E N D E R DIFFERENCES  66  TESTING THE MEDIATIONAL M O D E L  69  DISCUSSION  72  I M P A C T O N T H E R A P Y PROCESS  73  IMPACT ON THERAPY OUTCOME  83  THE MEDIATIONAL M O D E L  94  IMPLICATIONS FOR T H E PERFECTIONISM C O N S T R U C T  95  LIMITATIONS A N D F U T U R E DIRECTIONS IMPLICATIONS  96 102  PSP in Group Psychotherapy/page TABLE OF CONTENTS  (Cont'd)  REFERENCES  1  TABLES  1  FIGURES  1  APPENDIX A: PARTICIPANT RECRUITMENT POSTER  1  APPENDIX B: STUDY MEASURES  1  PERFECTIONISTIC S E L F - P R E S E N T A T I O N S C A L E  1  M U L T I D I M E N S I O N A L PERFECTIONISM S C A L E  1  M O O D SCALES  1  DISRUPTIVE T H E R A P Y B E H A V I O U R C H E C K L I S T  1  PSYCHOTHERAPY SELF-DISCLOSURE CODING S Y S T E M  1  T H E R A P Y PROGRESS R A T I N G - PATIENT, THERAPIST & O T H E R V E R S I O N S  1  D A V I C O N I Q QUESTIONS  1  INTERACTION A N X I O U S N E S S S C A L E  1  APPENDIX C: SOBEL'S (1982) FORMULA FOR MEDIATIONAL MODELS  PSP in Group Psychotherapy/page V  LIST OF TABLES Table 1:  Means, standard deviations and alphas of predictor variables and dependent variables  118  Correlations between protective forms of perfectionistic selfpresentation, trait perfectionism and other control variables  121  Correlations between protective forms of perfectionistic self-presentation and mood ratings variables  121  Multiple regression analyses predicting post-session anxiety at various points in treatment  122  Multiple regression analyses predicting post-session anxiety following all sessions controlling for trait perfectionism and related variables  123  Multiple regression analyses predicting post-session anxiety following sessions late in therapy, controlling for trait perfectionism and related variables  124  Table 7:  Correlations between perfectionism and salivary Cortisol  125  Table 8:  Multiple regression analyses predicting salivary Cortisol concentrations during the assessment session, controlling for trait perfectionism and related variables  125  Correlations between perfectionism and resistant therapy behaviours  126  Table 2: Table 3: Table 4: Table 5: Table 6:  Table 9:  Table 10: Multiple regression analyses predicting not showing up to sessions without providing an excuse, controlling for trait perfectionism and related variables  127  Table 11: Correlations between perfectionism and observational data from an early session  128  Table 12: Correlations between perfectionism and observational data from a late session  129  Table 13: Correlations between perfectionistic self-presentation, self-disclosure totals, valence, and focus, and observed sadness scores across all sessions  130  Table 14: Regression analysis of perfectionistic self-presentation predicting disclosure late in therapy, after controlling for level of disclosure early in therapy, trait perfectionism and other related variables  131  Table 15: Regression analysis of perfectionistic self-presentation predicting negative self-disclosures late in therapy, controlling for disclosures of this valence early in therapy, trait perfectionism and other related variables  132  Table 16: Regression analysis of perfectionistic self-presentation predicting ingroup and combined ingroup/outgroup self-disclosures, controlling for trait perfectionism and other related variables  133  Table 17: Correlations between perfectionism and ratings of treatment satisfaction and benefit  134  PSP in Group Psychotherapy/page vi  LIST OF TABLES (cont'd) Table 18: Multiple regression analyses of therapist ratings of patient benefit from treatment, controlling for trait perfectionism and other related variables  135  Table 19: Correlations between perfectionistic self-presentation and post-treatment outcome measures  136  Table 20: Partial correlations between perfectionistic self-presentation and posttreatment outcome measures, controlling for pre-treatment levels of these variables  137  Table 21: Multiple regression analyses predicting post-treatment depression, controlling for pre-treatment depression, trait perfectionism and other related variables  138  Table 22: Correlations between perfectionistic self-presentation and others' ratings of perfectionism pre- and post-treatment  139  Table 23: Correlations between perfectionistic self-presentation and autonomic reactivity in response to a stress test  140  Table 24: Partial correlations between perfectionistic self-presentation and posttreatment autonomic reactivity, controlling for pretreatment reactivity  141  Table 25: Multiple regression analyses predicting post-treatment SCL reactivity in response to a stress test, controlling for pre-treatment SCL reactivity, trait perfectionism and other related variables  142  Table 26: Differing ratings for male and female group members  143  Table 27: Interactions between gender and perfectionistic self-presentation in predicting disclosures late in therapy, controlling for the frequency of disclosures early in therapy  144  Table 28: Interactions between gender and perfectionistic self-presentation in predicting negative disclosures late in therapy, controlling for the frequency of negative disclosures early in therapy  144  Table 29: Interactions between gender and perfectionistic self-presentation in predicting positive disclosures late in therapy, controlling for the frequency of positive disclosures early in therapy  145  Table 30: Interactions between gender and perfectionistic self-presentation in predicting ingroup disclosures late in therapy, controlling for the frequency of ingroup disclosures early in therapy  145  Table 31: Interactions between gender and perfectionistic self-presentation in predicting ratings of perceived benefit by close others  146  Table 32: Interactions between gender and perfectionistic self-presentation in predicting post-treatment skin conductance return to baseline following a stress test, controlling for post-stress return to baseline pre-treatment  146  Table 33: Interactions between gender and perfectionistic self-presentation in predicting change in nondisclosure of imperfections as rated by close others  147  Table 34: Tests of the mediational model  148  PSP in Group Psychotherapy/page Vll  LIST OF FIGURES Figure 1: Gender differences in the relationship between nondisclosure of imperfections and level of disclosure late in therapy, after controlling for the level of early disclosure  149  Figure 2: Gender differences in the relationship between nondisclosure of imperfections and negative self-disclosures late in therapy, controlling for early negative disclosures  150  Figure 3: Gender differences in the impact of nondisclosure of imperfections on the frequency of positive self-disclosures late in therapy, controlling for positive disclosures early in therapy  151  Figure 4: Gender differences in the impact of nondisclosure of imperfections on ingroup disclosure late in the course of therapy, controlling for ingroup disclosure early in therapy  152  Figure 5: Gender differences in nondisplay predicting others' ratings of benefit from the group treatment  153  Figure 6: Gender differences in nondisplay of imperfections predicting SCL return to baseline post-treatment, controlling for this decline pre-treatment  154  Figure 7: Gender differences in nondisclosure of imperfection predicting SCL return to baseline post-treatment, controlling for this decline pre-treatment  155  Figure 8: Gender differences in post-treatment other-ratings of nondisclosure of imperfections controlling for pre-treatment levels of this variable  156  Figure 9: Illustrations of mediational models for nondisclosure of imperfections predicting problems in group therapy process and negative treatment outcome  157  PSP in Group Psychotherapy/page Vlll  DEDICATION  To my cousin, Andrea Herzer, who passed away shortly before the final defense of this thesis. Her courage and perseverance in the face of adversity have inspired many. Andrea's valiant battle for life demonstrated her commitment to her goals and dreams, and helped me find the strength to pursue my own.  PSP  in Group  Psychotherapy/page  IX  ACKNOWLEDGEMENTS  A project of this magnitude can only be accomplished with the commitment of a large and wellfunctioning team. As a group with limited financial resources, we relied on our team's enthusiasm and dedication to the development and study of a promising group treatment for trait perfectionism. Sincere thanks must be offered to the entire team including: my research supervisor, Dr. Paul Hewitt, who offered his support and excitement for this enormous undertaking, and who dedicated many of his research resources to this study; Dr. Sam Mikail who provided insightful and knowledgeable training and supervision of the therapy groups; the group therapists, Mike Papsdorf, Brenda Hogan, Bruce McMurtry, and Tanna Mellings who donated time and emotional energy to these often challenging groups; Michelle Haring and Melanie Parkin who co-lead several of the pre-group preparation sessions; and our large team of volunteers - Nicholette Beckett, Linda Chung, Hyo-Jung Chwa, Shauna Doersam, Revital Hayoun, Greenly Ho, Jeffery, Kathryn Koehler, Faye Lee, Gail Low, Marena, Katherine McKenney, Patrick Myers, Paula, David Paul, Grant Sigaty, Andrea Wardrop, and Victor Wong - who administered tests to study participants, worked on data entry and organization of study materials, and assisted with  Cortisol  analyses. I am particularly grateful t o my "right-hand" women, Melanie Parkin and Salome Mui, who are the most dedicated, loyal and supportive research assistants imaginable. On many occasions, they far surpassed their required work hours, yet remained enthusiastic  about  the  project  and eager t o assist me.  The passion and dedication inspired by this project is largely due to our desire to make a difference in the lives of the perfectionistic individuals that presented to our program. I must express my admiration and appreciation for the participants in our program who were courageous in identifying their difficulties, seeking help, and remaining in group treatment despite their fears and the challenges inherent in such a program. In addition, these participants were honest and forthright in offering feedback about the program which provides us with important insight, and the opportunity to make modifications in future treatment designs. In the design, analyses and writing of this dissertation, several other individuals played an integral role. I would like to thank all members of my various dissertation committees for their feedback and helpful suggestions. In particular, I am grateful for the support and input of my core committee of Dr. Paul Hewitt, Dr. Charlotte Johnston, and Dr. Dan Perlman. Thanks must also go to Dr. Marie Habke, Carmen Caelian, Simon Sherry and Jeneva Ohan for their feedback on earlier drafts, to Mike Papsdorf for his assistance in interpreting the methodology for mediational analyses, to Dr. John Ogrodniczuk for materials relevant to random regression models, and to Dr. George Iwama and members of his lab for providing the use of their lab equipment and training in Cortisol assay techniques. Finally, I would like to thank my grandparents, parents, and husband, Andrew, for their support of my goals for advanced education, and Jennifer, Mariana, and other friends and family members for their encouragement and support.  PSP in Group Psychotherapy/page 1  INTRODUCTION The devastation of perfectionism appears not to end with a long list of associations to various forms of psychopathology. Instead, recent studies have started to indicate that perfectionism also leads to difficulties accessing help, remaining in treatment and achieving positive therapy outcome (Blatt, Quinlan, Pilkonis & Shea, 1995; Habke & Hewitt, 2000; Nielsen, Hewitt, Han, Habke, Cockell, Stager, G., & Flett, 1997). Given the serious and even life-threatening conditions tied to various forms of perfectionism (e.g., Hewitt, Flett & Turnbull-Donovan, 1992; Hewitt, Flett & Weber, 1994), it is essential that barriers to accessing and making use of treatment be identified and removed so that perfectionists can be successfully treated. This investigation seeks to clarify the associations between several forms of perfectionism and key therapeutic factors in group psychotherapy, as well as therapy outcome. Early conceptualizations of perfectionism described it primarily as a cognitive style (e.g., Burns, 1980). In the last decade, two research teams have recognized that perfectionism is multidimensional and incorporates important interpersonal facets (Frost, Marten, Lahart & Rosenblate, 1990; Hewitt & Flett, 1991b). For example, Hewitt and Flett (1991b) first focused on trait dimensions of perfectionism, which they distinguish from early cognitive models of perfectionism by emphasizing strong motivational elements. They identified three forms of trait perfectionism: Self-oriented perfectionism is the requirement for oneself to be perfect and includes behaviours such as setting and maintaining unrealistic standards for oneself and selfcritical evaluations of behaviour to see if it meets those standards. Other-oriented perfectionism involves the same characteristics as self-oriented perfectionism, but the standards and evaluations are directed towards other people; that is, other-oriented perfectionists require others to be perfect. The last type of trait perfectionism is socially-prescribed perfectionism. This subtype refers to the belief that one cannot meet the perfectionistic expectations that others have for oneself. Having created this taxonomy, Hewitt and Flett then suggested that each dimension is associated uniquely with various forms of psychopathology and maladjustment. For example, whereas self-oriented perfectionism is associated more strongly with depression, particularly in the face of achievement stress (Hewitt & Flett, 1991a; Hewitt & Flett, 1993; Hewitt, Flett &  PSP in Group Psychotherapy/page 2  Ediger, 1996), other oriented perfectionism is linked to poor dyadic adjustment (Habke, Hewitt, Fehr, Callander & Flett, 1997; Hewitt, Flett & Mikail, 1995). Repeated studies have demonstrated strong associations between trait perfectionism dimensions and various psychological and physical problems. Depression, eating disorders, anxiety disorders, and personality disorders are just a few of the clinical diagnoses associated with elevations in perfectionism (Flett, Hewitt, Endler & Tassone, 1995; Hewitt, Flett & Ediger, 1995; Hewitt et al., 1992; Hewitt, Flett, Ediger, Norton & Flynn, 1998). In addition, type A behaviour, self-control, workaholism, suicide ideation and attempting, physiological and emotional stress reactivity, and intensity of migraine headaches appear to be linked to specific perfectionism dimensions (Flett, Hewitt, Blankstein & Dynin, 1994; Flett, Hewitt, Blankstein & O'Brien, 1991; Flynn, Hewitt, Flett & Weinberg, 2001; Hewitt, Newton, Flett & Callander, 1997; Hewitt, Norton, Flett, Callander & Cowan, 1998; Kowal & Pritchard, 1990; Spence & Robbins, 1992). Evidently, this personality trait has far-reaching implications in both the psychological and physiological realm. As awareness for interpersonal aspects of perfectionism grew, greater attention also was paid to the interpersonal implications of this personality style (Habke & Flynn, in press). Evidence is accumulating to support the notion that many forms of perfectionism are associated with problematic interpersonal functioning in casual, intimate and therapy relationships. Along with this knowledge has come a better conceptualization of how perfectionism operates in interpersonal settings. Hewitt and his colleagues (2000) have developed a multidimensional scale that assesses three forms of perfectionistic self-presentation. Preliminary studies have shown that perfectionistic self-presentation is an important predictor of relationship functioning even beyond trait perfectionism dimensions (Habke & Hewitt, 2000; Hewitt et al., 2000). Although relatively new to the field, these self-presentational forms of perfectionism will provide the primary focus for this paper. As a result, subsequent sections will explore the general theory of self-presentation as well as Hewitt et al.'s (2000) conceptualizations of the various dimensions of perfectionistic self-presentation. Following this introduction, data will be presented on associations between trait and self-presentational perfectionism and interpersonal  PSP in Group Psychotherapy/page 3  problems. Particular attention is paid to the literature on perfectionism in the therapy context. With that background, hypotheses then can be formulated concerning the role of perfectionism in group psychotherapy. An Overview of Self-Presentation  Much has been written about the related concepts of impression management and selfpresentation. Schlenker and Wei gold (1992) suggested that impression management is an active attempt to control information about an object, event, or one's self that is conveyed to a real or imagined other. Self-presentation is a subtype of impression management and specifically relates to information about the self (Leary, 1993). Some authors suggested that selfpresentation is pervasive in all forms of social interaction (e.g., Kelly, 2000a; Schlenker & Weigold, 1992), whereas others took the restrictive view that it is associated with interpersonal motives such as gaining power or approval (e.g., Jones & Pittman, 1993). Specifically, Jones and Pittman (1993) proposed five classes of power augmentation motive: ingratiation which attempts to gain affection by increasing one's likeabilitiy, intimidation which seeks to increase others' fear by coming across as dangerous, self-promotion which seeks respect by appearing attractive or competent, exemplification which induces guilt in others by the presentation of integrity and moral worthiness, and supplication which uses displays of weakness and dependence to encourage nurturance by others. Other researchers have built upon this model by clarifying how emotional expression can serve these motives (Clark, Pataki & Carver, 1996). For example, one might ingratiate oneself through a display of happiness or use expressions of anger to achieve intimidation. Two general styles of self-presentation are discussed in the literature (e.g., Baumeister, Tice & Hutton, 1989; Lee, Quigley, Nesler, Corbett & Tedeschi, 1999; Roth, Snyder & Pace, 1986; Schlenker & Weigold, 1992; Wolfe, Lennox & Cutler, 1986). The acquisitive style attempts to achieve a valued outcome by presenting self-flattering aspects of the self. This style is associated with more enthusiastic participation in social exchanges (Schlenker & Weigold, 1992), and a tendency to call attention to the self and one's good qualities (Baumeister et al., 1989). In contrast, protective forms of self-presentation strive to avoid negative or feared outcomes by  PSP in Group Psychotherapy/page 4  limiting social interactions and attempting not to draw others' attention (Schlenker & Weigold, 1992). As noted by Baumeister and his colleagues (1989), the protective style avoids the risks involved with self-promotion, particularly the risks of being perceived as boastful and of raising others' expectations. Several studies have explored the differences between these two styles. Frequently, high self-esteem is linked to self-promoting or acquisitive self-presentation strategies and low self-esteem, self-consciousness and social anxiety to protective strategies (Baumeister et al, 1989; Lee et al., 1999; Paulhus & Reid, 1991, Roth et al., 1986). In addition, different motives may be associated with each. Wolf et al. (1986) suggested that avoiding disapproval allows people to "get along" whereas promoting one's accomplishments and strengths permits us to "get ahead" (see also Bornstein, Riggs, Hill & Calabrese, 1996). Consistent with this, Assor (1996) identified both secure and insecure forms of recognition motivation, which differed in that secure forms led to active seeking of recognition and insecure forms were tied to strong needs for intimacy and fears of power. Normally, self-presentation serves several important purposes. It can maintain the individual's self-esteem, help with self-construction by shaping one's behaviour in keeping with visions of an ideal self, compensate for areas of weakness by focusing on and developing strengths, and help decrease evaluative scrutiny and elevated expectations by others through the use of self-protective strategies (Schlenker & Weigold, 1992). However, when taken to the extreme, these styles can lead to significant difficulties. Schlenker and Weigold (1992) noted that "indiscriminant self-glorification... can generate unrealistically high expectations for performance that doom the actor to failure; it can generate anxiety caused by concerns about having to maintain an unrealistic facade; and it can mark the actor as an egotist, an irritant, and an uncooperative social participant". Leary (1993) also commented on the risks of trying to appear better than one really is. He suggested that this behaviour further elevates the individual's motivation to self-present as a consequence of the growing discrepancy between their desired and real public image. Similarly, protective styles also can have adverse effects when taken to an extreme. There are times when it is important and even necessary to show weakness to others. For example, Clark et al. (1996) demonstrated that expressions of sadness can elicit helping  PSP in Group Psychotherapy/page 5  behaviour from others. Thus, an individual who chronically uses protective styles may fail to evoke such nurturance due to an inability to demonstrate their negative affect. From another perspective, these authors noted the impact of self-presentation on the target person. The recipient of self-presentational strategies may feel less able to trust the presenter and may attend more closely to their behaviour in order to discern their "true" selves. This response may increase levels of tension for both the self-presenter and the target. Perfectionistic Self-Presentation  The concept of perfectionistic self-presentation stems from the above literature. Initial studies on perfectionism focused on trait components of the construct, which are long-term, motivational facets of perfectionism (Hewitt & Flett, 1991a). However, it also is essential to consider stylistic aspects of personality traits which.explain "the how, not the what, of behaviour" (Buss & Finn, 1987, p. 438-439). Perfectionistic self-presentation considers how people attempt to create an image of perfection for others around them (Hewitt et al., 2000). As with self-presentation in general, both acquisitive and protective styles exist. These have been labeled perfectionistic self-promotion, nondisplay of imperfection and nondisclosure of imperfection. For the perfectionistic self-promoter, an image of perfection is obtained by overcommunicating information about strengths and successes. Such an individual may brag about his or her test scores or invite friends and family members to watch their athletic triumphs. In essence, this strategy actively draws the audience's attention to vibrant displays of accomplishment in hopes that signs of weakness and imperfection will be forgotten and ignored. In contrast, individuals who employ protective strategies seek to hide all signs of imperfection and flaws by attempting to become 'invisible' in some way. Hewitt et al.'s studies demonstrate that two protective forms of perfectionistic self-presentation exist and have different impacts. Individuals scoring highly on the nondisplay of imperfection avoid being seen making errors or being perceived in a flawed light. Those with concerns about the nondisclosure of imperfections avoid admitting to others, or talking about, anything that might be perceived as a flaw or a shortcoming. Thus, the nondisplay of imperfections may be most relevant in public contexts in which one might be observed making a mistake, whereas the nondisclosure of imperfections may  PSP in Group Psychotherapy/page 6  be most important in close personal relationships in which such intimate disclosure may be the norm (Hewitt et al., 2000). As noted in the previous section, several authors have recognized the aversive impact of excessive attempts at self-presentation (Clark et al., 1996; Leary, 1993; Schlenker & Weigold, 1992). Clearly, attempts at presenting the self as perfect qualify as excessive. Hewitt and his colleagues (2000) investigated the relative associations of each form of perfectionistic selfpresentation and a variety of psychological symptoms. Consistent with past work on selfpresentation and self-esteem (Baumeister et al, 1989; Roth et al., 1986), both protective forms of perfectionistic self-presentation, and nondisplay of imperfection in particular, were linked with lower self-esteem. In general, nondisplay of imperfections was associated with greater psychopathology in the form of anxiety and depression symptoms, social anxiety or social phobia, and reports of interpersonal problems such as difficulties with intimacy, submission, assertiveness, and sociability. Similarly, nondisclosure of imperfection was linked to anxiety and social phobia symptoms, interpersonal difficulties and especially, problems with dyadic adjustment. Although perfectionistic self-promotion also was correlated with many of these problem categories, the magnitude of these correlations was much less than that for the protective styles. In addition, it rarely explained any unique variance in psychological symptomatology. It is also important to compare trait and self-presentational forms of perfectionism. As noted earlier, the trait dimensions of perfectionism reflect underlying motivations to be perfect. In contrast, perfectionistic self-presentation reflects the need to appear to be perfect to others (Hewitt et al., 2000). In other words, perfectionistic self-presentation is the interpersonal manifestation of perfectionism. All subtypes of trait and self-presentational perfectionism are correlated (Hewitt et al., 2000) and correlations vary from .20 to .59. Other-oriented perfectionism demonstrates the smallest links to perfectionistic self-presentation, perhaps because this form of perfectionism focuses more on others' imperfections. Self-oriented perfectionism is highly associated with perfectionistic self-promotion and nondisplay of imperfection and socially-prescribed perfectionism shows large correlations to all three forms of  PSP in Group Psychotherapy/page 7  perfectionistic self-presentation. However, trait and self-presentational measures are not redundant. The correlations reported above are farfromperfect. Thus, individuals may score highly on trait but not self-presentational forms of perfectionism and vice versa. A good example is the subscale of nondisclosure of imperfections. One individual may score high on self-oriented perfectionism and believe that admitting to faults is a sign of weakness that they cannot tolerate. Another self-oriented perfectionist may perceive it as a flaw not to own up to one's mistakes. Many combinations or profiles of scores are possible. From a research perspective, several studies have shown that perfectionistic self-presentation can help predict scores on indices of psychopathology, over and above the knowledge gainedfromtrait perfectionism. This was true for predictions of several facets of self-esteem, anxiety, depression, social anxiety, and interpersonal problems (Hewitt et al., 2000). As would be expected theoretically, perfectionistic self-presentation appears to be most relevant in predictions of social difficulties. Perfectionism in the Interpersonal Context  Of particular import to this study is research on the role of perfectionism and perfectionistic self-presentation in interpersonal relationships. Only a few studies have been conducted in this area, yet each has pointed to a key role for perfectionism. In the following paragraphs, work on perfectionism and interpersonal styles will be considered first. Next, a closer inspection will be made of perfectionism in specific interpersonal contexts with a particular focus on psychotherapy. A greater volume of work has been published on trait forms of perfectionism, however there is growing evidence that the newly identified perfectionistic self-presentation is especially relevant in intimate relationships. Interpersonal Styles Habke and Flynn (in press) reported that trait perfectionism can have indirect and direct effects on interpersonal relationships. They suggest that indirect effects occur through perfectionism's associations with depression, social anxiety and personality disorders. Consider, for example, links with personality disorders. Other-oriented perfectionism is associated with dramatic cluster personality traits, especially narcissism, which could lead to more aggressive  PSP in Group Psychotherapy/page 8  conflicts with others in which the perfectionistic individual plays a domineering and exploitive role (Hewitt & Flett, 1991b; Hewitt, Flett & Turnbull, 1992). In contrast, the socially-prescribed perfectionist is characterized by both odd/eccentric and anxious/fearful cluster traits which make it likely that they will avoid social contact and conflict (Hewitt & Flett, 1991b; Hewitt, Flett & Turnbull, 1992). However, this perfectionism dimension can also be linked to the intense emotions and interpersonal disarray of the borderline personality (Hewitt, Flett & Turnbull, 1994). While self-oriented perfectionism is not as clearly connected to the rigid styles of any particular personality disorder (Hewitt, Flett & Turnbull, 1992), there is some indication that the interpersonal behaviour of these individuals can create conflict through the display of mildly narcissistic characteristics (Hewitt & Flett, 1991b). Other studies have explored a more direct link between perfectionism and interpersonal problems. For example, various forms of trait perfectionism have been mapped onto an interpersonal circumplex derived from the Inventory of Interpersonal Problems (IIP; Horowitz et al., 1988; circumplex scalesfromAlden, Wiggins & Pincus, 1990). The interpersonal circumplex has been derived from neo-Sullivanian theory, which focuses on interpersonal exchanges of love and status (Wiggins, Trapnell & Phillips, 1988). The circumplex is divided into quadrants by the axes of love or nurturance, and status or dominance. More defined groupings can be made by dividing the circumplex into octants. Hill, Zrull and Turlington (1997) concluded that in males, all three trait forms of perfectionism were tied to dominant and hostile personality traits, specifically the octants of 'domineering' and 'vindictive'. These sections reflect problems with control, manipulation, suspicion, and lack of empathy. In women, a different pattern emerged in which each form of perfectionism was associated with a distinct set of problem areas. Self-oriented was not correlated strongly with any particular octant although it was most related to 'overly nurturant'. The authors concluded that this form of trait perfectionism was not significantly tied to interpersonal distress. Other-oriented perfectionism in women demonstrated similar relationships to those displayed by male subjects, that is, strong links to 'domineering' and 'vindictive' problem areas. Finally, socially-prescribed perfectionism in women was positively  PSP in Group Psychotherapy/page 9  correlated with all interpersonal problem domains and reflected the greatest degree of interpersonal distress. A more recent study extended Hill et al.'s (1997) work by mapping interpersonal problems associated with trait and self-presentational forms of perfectionism as well as perfectionism measures developed by other research teams (Flynn, Hewitt, Broughton & Flett, 1998). Their findings for trait perfectionism were quite similar to those of Hill et al. and emphasized ties to domineering behaviour for both genders and coldness in women. Results for perfectionistic selfpresentation suggested that these styles are linked to dominance and coldness in men and difficulties with coldness in women. Other important findings included the fact that at least one perfectionism subscale by some research group was located in each octant of the circumplex providing further support for the multidimensional conceptualization of perfectionism. It is also important to note that perfectionism as construed by Hewitt and Flett (1991, Hewitt et al., 2000) appeared to access a more hostile interpersonal style, which contrasted with the more agreeable traits tapped into by Frost et al. (1990) with their measure of perfectionism also called the Multidimensional Perfectionism Scale. The fact that so many of Hewitt and Flett's perfectionism dimensions lie in the hostiledominant and hostile-submissive quadrants has important therapeutic implications. Such interpersonal behaviours have been associated with poor therapeutic outcome. Gurtman (1996) found that interpersonal problems associated with hostility predicted less therapeutic progress within sessions. In addition, hostile dominance was related to therapeutic liabilities such as a present orientation, impulsivity, and difficulties with intimacy, as well as poor 'emotional resonance', a sense that patient and therapist were not relating well. Thus, the discovery of links between various perfectionism dimensions and both hostile and dominant interpersonal difficulties provides the first hint of the potential for problems within the therapeutic context. This link will be reviewed more directly in later sections. In addition to investigating associations between perfectionism and interpersonal problems, researchers have looked at specific interpersonal behaviours that are known to be problematic. Hobden and Pliner (1995) asked student volunteers to complete a challenging analogy test. After  PSP in Group Psychotherapy/page 10  trying the first test, participants were offered a choice of music tapes labeled as being on a continuum from performance enhancing to performance inhibiting. These tapes supposedly would be played while they took a second version of the test. As an additional manipulation, some participants made their music choice in front of the experimenter and were told to put their name on their test paper so the experimenter could review their scores with them (public condition). Other participants' music choice and second set of test scores were to be kept confidential from the experimenter (private condition). Participants who scored highly on selforiented perfectionism self-handicapped regardless of the publicity of their condition by selecting less enhancing tapes. Presumably, this would allow them to attribute any lack of a perfect performance to an external factor, the tape, a self-protective technique. In contrast, sociallyprescribed perfectionists self-handicapped more than others in the public condition, but not in the private condition. This result implies that socially-prescribed perfectionists are concerned most with the impression that they make on others. Beyond the interesting differences between the trait forms of perfectionism, it is also noteworthy that both types of perfectionism are linked to self-handicapping, a strategy which shields the individual from direct evaluation. In an interpersonal context, this strategy impedes others' ability to really know an individual and their capabilities. A final study has considered the potential impact of such interpersonal styles and strategies on interpersonal relationships. Flett, Hewitt, Garshowitz and Martin (1997) found that sociallyprescribed perfectionism was positively correlated with frequency of negative social interactions in an undergraduate student sample. These negative interactions included overt criticism, lack of recognition and betrayal. Because participants reported on these negative interactions themselves, it is impossible to know whether their reports are veridical or are biased by skewed social perceptions. However, this study does provide some initial support for the notion that perfectionism may lead to problematic social interactions. Unfortunately, much less work has been done with perfectionistic self-presentation, but based on the positive correlations between these styles and socially-prescribed perfectionism (Hewitt et al., 2000), we might expect similar negative consequences.  PSP in Group Psychotherapy/page 11  Intimate Relationships Although not the primary focus of this paper, there is a growing and important literature base pointing to perfectionism's destructive role in intimate relationships. Some of the problems that are experienced in this context can be used to predict comparable difficulties in the therapeutic relationship, which also requires intimacy and self-disclosure. Thefirststudy that investigated perfectionism in intimate relationships assessed pain patients and their spouses (Hewitt, Flett & Mikail, 1995). Patients whose spouses scored more highly on other-oriented perfectionism reported poorer dyadic adjustment and more family difficulties. In addition, they rated their spouses as less supportive. This study suggested that the targets of perfectionistic expectations might be particularly prone to relationship maladjustment. Habke and Flynn (in press) reviewed several more recent findings in this area. In particular, research has focused on relationship adjustment and interpersonal behaviour. Adjustment often is assessed with the Dyadic Adjustment Scale (DAS; Spanier, 1976), which includes a measure of overall adjustment and subscales tapping into affectional expression, a sense of togetherness (cohesion), agreement between partners (consensus), and satisfaction with the relationship. Studies have demonstrated that the perception that one's spouse expects perfection is associated with lower overall adjustment scores and lower scores on each of the subscales, even after controlling for depression. Habke and Flynn (in press) propose two explanations for this finding. First, distressed marriages may be characterized by poor communication patterns permitting the development of assumptions about one's partner's expectations without opportunities for reality testing. Second, given socially-prescribed perfectionism's strong links to protective selfpresentation (Hewitt et al., 2000), these marriages may be less intimate. In contrast to the findings for socially-prescribed perfectionism, perfectionistic expectations for oneself were linked to greater dyadic satisfaction. Again, two explanations of this finding are possible. Selforiented perfectionists may attribute failure to themselves rather than their relationship or they may work harder at being "the perfect spouse" (Habke & Flynn, in press). This work recently has been extended to perfectionistic self-presentation. Hewitt and Flett (1997) report that higher levels of perfectionistic self-promotion and lower levels of  PSP in Group Psychotherapy/page  12  nondisclosure of imperfections predict more positive overall dyadic adjustment and higher scores on the subscales of consensus and satisfaction. Similarly, lower levels of nondisplay of imperfection, and elevations on perfectionistic self-promotion predicted greater affectional expression. These self-presentational variables continued to be important in dyadic adjustment even after controlling for depression and trait perfectionism. Thus, the protective forms of perfectionistic self-presentation appear to be particularly damaging in intimate relationships. The fact that perfectionistic self-promotion demonstrated a positive impact on dyadic adjustment suggests that such behaviour may used less frequently in intimate relationships. Alternatively, it could be viewed within the dyad as a positive expression of strengths. A second line of research has invited couples to discuss problems in their marriage and coded negative behaviours such as criticism, negative solutions and disagreements, and positive behaviours like positive solutions, agreements and acceptance (Habke et al., 1997). Men's reports of perfectionistic expectations for their partners predicted a greater proportion of male negative behaviour during the experimental interaction. Similarly, males who believed that their partner had perfectionistic expectations for them had mates who employed more negative and less positive behaviour than other women. This finding may reflect the men's accurate perceptions of their spouse's demanding and argumentative style or the women may be responding to their mate's hostility over perceived unreasonable demands. Consistent with the work on dyadic adjustment, self-oriented perfectionism in men predicted a greater frequency of men's positive behaviours further supporting the proposition that this form of perfectionism may motivate them to work hard to achieve relationship success. In sum, the work on perfectionism in intimate relationships suggests that it does interfere with relationship adjustment and satisfaction perhaps by promoting the use of negative interpersonal behaviour and decreasing the frequency of positive behaviours like acceptance and agreement. Further, in proposing a mechanism by which perfectionism leads to poor dyadic adjustment, researchers have theorized that perfectionism may interfere with intimacy and communication. Once again, these observations are indicative of potential problems in a therapy relationship, which requires communication and intimacy. A large proportion of negative  PSP in Group Psychotherapy/page 13  interpersonal behaviours also would be detrimental, particularly in group therapy where other members may be less forgiving than the therapists of an abrasive interpersonal style. Perfectionism in the Therapy Context In considering perfectionism's impact on psychotherapy, it is essential to identify the various stages of therapeutic engagement (Hewitt & Flett, in press). First, an individual must be able to identify that they are experiencing difficulties and choose to contact a mental health professional for assistance. This stage can be referred to as help seeking. Once the individual has made this initial contact, they must be capable of participating in the process of therapy, which may include learning to trust the therapist, disclosing information about his or her difficulties, and establishing an alliance with the therapist. Finally, we would hope that the individual could become sufficiently engaged to experience a positive outcome as a result of their psychotherapy experience. Research on the impact of perfectionism at each of these stages will be addressed in turn. As perfectionistic self-presentation is the focus of the present study, where possible these results will be presented rather than those for trait perfectionism. Help Seeking. Ey, Henning, and Shaw (2000) investigated help-seeking behaviour in medical and dental students. They divided their sample into three categories: students who were not distressed, students who were in psychological treatment, and students who were distressed but were not in treatment. The distressed but not in treatment group was distinguished from the other samples by higher levels of socially-prescribed perfectionism, as well as more negative attitudes about mental health services. Although correlations between perfectionism and health seeking attitudes were not provided in the article, it would appear that socially-prescribed perfectionism is associated with negative attitudes about help-seeking, or at the very least, with a reluctance to present for treatment despite elevated levels of emotional distress. One team of researchers explored connections between help-seeking and perfectionistic selfpresentation (Nielsen, Hewitt, Han, Habke, Cockell, Stager & Flett, 1997). A sample of undergraduates was asked to complete measures of trait and self-presentational perfectionism as well as the Attitudes Toward Seeking Professional Psychological Help Scale (Fisher & Turner, 1970). At the bivariate level, all three types of perfectionistic self-presentation were negatively  PSP in Group Psychotherapy/page  1.4  correlated with all forms of help seeking. After controlling for gender, ethnicity, help-seeking history, current depression levels, and trait forms of perfectionism, nondisclosure of imperfection predicted low levels of interpersonal openness and confidence in the mental health profession. In addition, perfectionistic self-promotion was linked to problems with stigma tolerance and interpersonal openness. These findings have been replicated in other student and community samples, and subsequent work suggests that protective forms of perfectionistic self-presentation also predict greater fears of psychotherapy (Hewitt, Flett, Han, Tomlin & Nielsen, 2001). From these findings, it is clear that perfectionistic self-presentation is likely to impede a potential client's ability to seek help from a mental health professional. Such clients may wait until their symptoms are particularly troubling to present for help and may continue to experience some resistance to the therapy process. For example, both perfectionistic self-promoters and those with troubles disclosing their imperfections would have difficulty in being open with the therapist and may have trouble establishing a trusting therapeutic alliance. These findings are supported by work on self-concealment (Cepeda-Benito & Short, 1998). All three dimensions of perfectionistic self-presentation are positively correlated with self-concealment (Hewitt et al., 2000). Cepeda-Benito and Short's (1998) study of undergraduate students found that selfconcealment was associated with more frequent reports of having required psychological assistance but not seeking it. Thus, investigations of both attitudes and behaviour suggest that perfectionistic self-presentation is negatively associated with help-seeking. Process. Although studies like the ones described above (e.g., Nielsen et al., 1997) suggest some ways in which perfectionists may have trouble in therapy, only one study has attempted to investigate the role of perfectionistic self-presentation in therapy process. Habke and Hewitt (2000) assessed 90 clients presenting to a number of outpatient clinics. In addition to completing a number of self-report measures, participants engaged in a brief interview that included a discussion of past mistakes. Habke and Hewitt investigated participants' reactions to the interview using cognitive, affective and behavioural measures of distress. As a cognitive measure, they inquired about participants' appraisals of their own and the interviewer's expectations and satisfaction with their performance. To tap the affective dimension, participants  PSP in Group Psychotherapy/page 1 5  completed self-report measures of anxiety and distress at the start of the assessment and following the interview. In addition, physiological measures of heart rate and skin conductance levels provided information on physical arousal related to emotional distress. Finally, participants were given an opportunity to self-handicap by providing excuses in advance for why their interview performance might be impaired as well as an estimate of the magnitude of the impact of these handicaps. These last measures provided behavioural indicators of discomfort. At the bivariate level, there were strong indications that protective forms of perfectionistic self-presentation were linked to greater distress during the interview. These forms of perfectionism predicted appraisals of the interviewer as wanting more than the participant could provide and as being disappointed following the interview. In addition, these individuals reported more emotional distress and showed both higher heart rate and greater heart rate change from relaxation during the interview. Finally, perfectionistic self-presentation predicted greater use of self-handicapping excuses. Many of these relationships continued to be significant even after controlling for demographics, trait perfectionism, emotional distress (anxiety and depression), and impression management. In particular, nondisclosure of imperfection continued to predict appraisals of threat and of the interviewer as being disappointed with their performance, as well as greater heart rate elevation during the interview. This study suggests that protective forms of perfectionistic self-presentation may lead clients to experience greater emotional distress when disclosing about mistakes or shortcomings to their therapists. In addition, these clients may perceive the therapist as being more threatening or demanding and as being more disapproving of them. Clearly, these beliefs would make it difficult to establish a positive working alliance. Although this study presents many intriguing findings, it has several limitations. First, the study was an analogue because the interviewers did not carry out a typical assessment interview, nor were they intended to become the participant's continuing therapist. The limited role of the interviewer and the fact that participants did not have to meet with them again may, in fact, have decreased participants' concerns about the interviewers' perceptions and minimized some effects. Secondly, because this study focused on reactions during an initial assessment session, it did not test the theory that such responses should  PSP in Group Psychotherapy/page 16  intensify during the course of therapy, as intimacy increases in the therapy relationship. Perhaps most importantly, and tied to the last point, is the question of whether such reactions are clinically meaningful. That is, does discomfort during an initial session predict early termination or more negative therapy outcome? Certainly, there are some who claim that initial anxiety levels are a predictor of therapy success (e.g., due to better cohesion; Sexton, Hembre & Kvarme, 1996; or due to less attrition; Saltzman, Luetgert, Roth, Creaser & Howard, 1976). Hence, the final section relating perfectionism to therapy outcome is of considerable import in demonstrating the potential impact of perfectionists' initial distress in therapy. Outcome. While the other sections boasted at least one relevant study on perfectionistic selfpresentation, few researchers have focused on the role of perfectionism on therapy outcome. Thus, a brief review is included here of other personality variables and interpersonal styles associated with perfectionism which have been studied in therapy outcome studies. In general, however, links to outcome are infrequently replicated, and personality factors explain only five to ten percent of outcome variance (Garfield, 1994). A few findings will be described here only to provide further clues to the likelihood of perfectionism also being linked to poor outcome. Anxiety has been conceptualized in both state and trait forms (Endler & Magnusson, 1976; Spielberger, Gorsuch & Lushene, 1970). State anxiety or tension in initial sessions has been associated with better alliance and possibly more engagement (Sexton et al., 1996). In addition, greater anxiety reduction in early sessions also predicts better outcome (Garfield, 1994). Similarly, initially high scores on trait measures of anxiety predict more substantial improvement post-treatment (Conte et al., 1988). In contrast to the work on anxiety, high levels of general psychiatric symptomatology are associated with more negative outcome in individual psychotherapy (Garfield, 1994). Elevated neuroticism scores and introversion also predict less improved social adjustment a year following treatment (Zuckerman, Prusoff, Weissman & Padian, 1980). Yet another study reported that patients with high scores on histrionic, paranoid and obsessive-compulsive personality scales also showed less improvement than other patients at an outpatient clinic (Conte et al., 1988). Interestingly, only the perfectionism component of the obsessive compulsive personality scale was used in this study. Finally, a study of cognitive  PSP in Group Psychotherapy/page 17  behavioural therapy found that willingness to try new coping strategies was an important predictor of therapy success (Burns & Nolen-Hoeksema; 1991). This list of findings includes many different traits with differing impacts on therapy outcome. Again, the precaution must be made that few of these results have been replicated. However, perfectionism's connection to greater levels of psychopathology, difficulties trusting others, and perhaps, some rigidity (Hewitt & Flett, 1991b; Hewitt et al., 2000) suggests that it will not be a benefit in psychotherapy. A more recent study by Rector, Bagby, Segal, Joffe and Levitt (2000) explored the association between self-criticism and outcome following cognitive or pharmacological treatment for depression. Although self-criticism failed to have an impact on pharmacological treatment outcome, it did reduce the therapeutic benefit of the cognitive therapy program. Even more striking was the association between change in self-criticism over the course of therapy, and cognitive therapy outcome. Greater reductions in self-criticism from pre- to post-treatment were associated with larger reductions in depression symptomatology. These authors suggest that self-criticism is moderately associated with more general measures of perfectionism. Therefore, perfectionism may also contribute to a diminished treatment response to cognitive therapy. Other studies have looked specifically at patterns of interpersonal relating. For example, Conte et al. (1991) studied 96 patients who participated in at least four therapy sessions at their outpatient clinic. They found that patients who reported that they were highly rejecting of others experienced less improvement after therapy as assessed by symptom reduction and global functioning. In another interesting study, Maling, Gurtman and Howard (1995) factor analyzed patient and student responses to the Inventory of Interpersonal Problems (Horowitz et al., 1988) and identified three factors. The factors included control, detached, and self-effacing. The researchers then plotted dose-response curves for over 300 patients according to scores on these factors. Problems with control seemed to resolve most quickly. Improvement began after about 10 sessions. Detachment showed slower reactivity with improvement beginning at session 17. Finally, the self-effacing scale did not change over the course of therapy. These findings stand in contrast to those of Gurtman (1996) reported earlier in which the hostile dominant style was  PSP in Group Psychotherapy/page  18  related to less therapy progress and other therapeutic liabilities. It would seem that many different forms of interpersonal difficulty can be associated with impaired therapy process and outcome. In addition to work on related traits and interpersonal styles, less well-defined measures of perfectionism have been used in a few projects and point to a growing recognition that this personality variable may be difficult to treat and signals less positive therapy outcome. In reviewing the literature, one risks being drawn into the debate regarding what constitutes good outcome measures. However, the key appears to be using outcome measures rated by a variety of people. One study illustrated the impact of patient expectations on perceived outcome (Goldstein & Shipman, 1961). In their investigation of 30 outpatients, they concluded that patient expectations showed a curvilinear relationship with perceived symptom reduction after one therapy session. That is, those with moderate expectations reported the greatest symptom relief. The authors suggested that individuals with rigidly high or markedly low expectations (both groups could include perfectionists, given their perceptions of the mental health profession) might have been less responsive to evidence of change during the session. The low expector may be unable to believe that change is possible, while the high expector is likely to always be disappointed with reality. Thus, while patient satisfaction continues to be an important consideration, the use of alternate ratings of outcome may increase the accuracy of assessment of change. The first line of outcome research suggests that perfectionism is fairly persistent when not directly targeted in treatment. Several studies on eating disorders have reported that perfectionism does not change significantly over the course of treatment although weight restoration may occur (Bastiani, Rao, Weltzin & Kaye, 1995; Kaye, 1997; Szabo & Terre Blanche, 1997). The study by Bastiani et al. (1995) is most relevant because it uses both Hewitt and Flett's (1991b) and Frost et al.'s (1990) multidimensional measures of perfectionism. These findings provide support for the notion that perfectionism may be a longstanding vulnerability factor for various forms of psychopathology (Hewitt & Flett, in press). As such, the development of treatments that directly target perfectionism may be essential to prevent relapse.  PSP in Group Psychotherapy/page  19  The second line of research comes from the NIMH Treatment of Depression Collaborative Research Program (TDCRP; Elkin et al., 1989; Imber et al., 1990) in which patients were randomly assigned to four treatment conditions: interpersonal psychotherapy (IPT), cognitivebehavioural therapy (CBT), imipramine and clinical management, or pill placebo and clinical management. One research team became more interested in differences between patient outcomes than treatment outcomes. Blatt, Quinlan, Pilkonis and Shea (1995a) factor analyzed the Dysfunctional Attitude Scale (Weissman & Beck, 1978) and identified two factors, namely need for approval and perfectionism. They found that the perfectionism factor predicted negative treatment outcomes regardless of the type of treatment . Outcome was assessed using both self1  report and clinical ratings of depression, general clinical functioning and social adjustment. Higher perfectionism scores were associated with worse outcome on all of these measures. Once again, the authors suggested that perfectionism may need to be targeted directly using long-term, intensive psychoanalysis. However, they also proposed that perfectionism interferes with clients' ability to engage successfully in the process of therapy, particularly over a brief period of time. Three subsequent analyses have attempted to clarify perfectionism's impact. First, Blatt and his colleagues (1996) inspected the quality of the therapeutic relationship. Individuals who reported average levels of perfectionism experienced very different levels of outcome success depending on patients' ratings of the therapeutic relationship after the second treatment session. In this sub-sample, individuals who rated the therapeutic relationship positively also achieved greater symptom reduction at the end of therapy. In contrast, poor quality alliance predicted less successful outcome. This correspondence between therapeutic alliance and outcome is commonly recognized. The surprising phenomenon here was that individuals who scored more than one standard deviation below or above the mean on the perfectionism measure did not show this association. Those scoring low on perfectionism achieved a positive outcome regardless of therapeutic alliance ratings. In contrast, high scores on perfectionism predicted negative outcome even in the presence of a positive therapeutic alliance. These results suggest that highly ' See also Blatt et al. (1995b) for a correction to the original paper. This correction does not alter the original conclusions but presents revised statistical findings.  PSP in Group Psychotherapy/page 20  perfectionistic individuals may be unable to establish and benefit from an initially positive alliance with their therapist. In a subsequent publication (Blatt, Zuroff, Bondi, Sanislow & Pilkonis, 1998), these researchers confirmed that perfectionism also impeded successful outcomes as rated by independent clinical evaluators, the therapists and the patients even 18 months following treatment termination. In addition, they identified an intriguing time-line for these impairments. Until the eighth session, perfectionists did not lag behind other participants. It was only in the last third of the treatment that perfectionists did not continue to make gains. While Blatt et al. (1998) propose that perfectionists may have had difficulty accepting the imposed termination date, it is also possible that the process of therapy changed during the later period. For example, therapists may have pushed for greater openness at that point or attempted to focus more specifically on perfectionistic behaviours. In response, perfectionistic patients may have become anxious and resistant. Though interesting, these results do little to explain the mechanism by which perfectionism interferes with therapy process. Most recently, Zuroff et al. (2000) proposed that specific aspects of the therapeutic alliance might mediate the association between perfectionism and therapy outcome. Specifically, these authors tested associations between perfectionism and observer ratings of patient and therapist contributions to the therapeutic alliance. As in previous analyses, perfectionism was not linked to the magnitude of either alliance rating early in the course of therapy. However, perfectionism did predict less of an increase in patients' contributions to the alliance over the course of therapy. The patient's contribution to the alliance included their ability to engage in the work of therapy and to agree on goals and tasks of therapy with their therapist. Further analyses demonstrated that the association between perfectionism and negative therapy outcome was partially mediated by the lack of increase in patient alliance over time. Thus, one mechanism that may explain the link between perfectionism and poor therapy outcome is that perfectionism leads to difficulties developing a positive therapy alliance. However, the patient's impact on the alliance does not fully account for the perfectionism-outcome link suggesting that other mediators may be present.  PSP in Group Psychotherapy/page 21  In addition, Zuroff et al.'s study does not explore why perfectionistic individuals have difficulty engaging in therapy. Although this series of papers provides support for the notion that perfectionism may be associated with negative treatment outcome, other authors have questioned this association (Mussell, Mitchell, Crosby, Fulkerson, Hoberman & Romano, 2000). This research team investigated predictors of bulimia symptom remission following a 12-week program of cognitive-behavioural therapy. They noted that the perfectionism subscale of the Eating Disorders Inventory (EDI; Garner, Olmsted & Polivy, 1983) failed to predict the likelihood of post-treatment remission after controlling for pre-treatment depression and vomiting frequency. Unfortunately, the paper did not document whether perfectionism was associated with outcome at a bivariate level. A key difference between this study and that of Blatt et al. (1995) is the measure used to assess perfectionism levels. Both studies used general measures of perfectionism that may emphasize slightly different aspects of perfectionism. By exploring differences between specific perfectionism dimensions in the present study, this discrepancy may be better explained. In sum, it is apparent that perfectionism is related to problematic interpersonal behaviour. These individuals are often domineering and hostile and appear to have difficulty establishing intimate and satisfying partnerships. Given that psychotherapy is an interpersonal process requiring at least a modicum of self-disclosure and intimacy, it is not surprising that perfectionists often have less successful therapy outcomes. However, much work remains in order to clarify the impact of perfectionism on psychotherapy. To date, no study has considered multiple dimensions of perfectionism in the therapy context. As suggested by Habke and Hewitt's (2000) study, the protective forms of self-presentation may be particularly troublesome in psychotherapy. In addition, there is little information about how perfectionism interferes with therapy process and therapeutic alliance. The help-seeking literature implies that perfectionists may have difficulty identifying their need for outside assistance and being open enough to become involved in a treatment program. However, once in therapy, we have little  PSP in Group Psychotherapy/page 22  understanding of how perfectionists experience the therapy encounter or of how their behaviour may prevent optimal progress. This project seeks to fill these knowledge gaps. Why Group Psychotherapy?  Group psychotherapy was selected as the treatment modality for this study. Beyond the practical research advantages of many clients being seen for few therapist hours, group psychotherapy appears to be a promising and important context for perfectionistic clients. Three primary rationales are offered here. First, group psychotherapy, in general, has demonstrated considerable effectiveness and efficiency making it a treatment of choice in our increasingly cost-conscious society (Bednar & Kaul, 1994; Piper, 1991). Second, group psychotherapy may be an especially challenging environment for perfectionistic individuals. Finally, the group therapy context may be particularly therapeutic for perfectionistic individuals. Each of these points will be discussed in turn. There is increasing awareness that group psychotherapy offers comparable treatment success to individual therapy (e.g., Grunebaum, 1975). Dies (1993, p. 475) noted, "contemporary surveys.. .indicate widespread endorsement of group psychotherapy as a viable and vital treatment modality". In a review of 32 well-controlled studies (Toseland & Siporin, 1986), group psychotherapy proved to be more effective than individual therapy in 25% of the studies and comparable in the remaining 75%. Similarly, in a more recent analysis, Bednar and Kaul (1994) concluded that there is sufficient evidence to assert that group psychotherapy is more effective than no treatment, placebo, nonspecific effects and certain other recognized treatments. In addition to growing belief in its effectiveness, the efficiency and cost-effectiveness of group psychotherapy also have been touted (Piper, 1991). In Toseland and Siporin's (1986) review, they reported that of 12 studies to assess efficiency, 10 concluded that groups were superior to individual treatment. To put this enthusiasm for groups into perspective, it is also important to recognize that there are limitations in the group psychotherapy research reported to date. Dies (1993) and Bednar and Kaul (1994) present several suggestions for improving future studies. These suggestions include: better descriptions of group treatments and checks to ensure that treatments are delivered as intended, more representative samples and settings (rather than  PSP in Group Psychotherapy/page 23  students and trainees), more psychometrically sound measures, outcome measures that tap into change that may be specific to group psychotherapy, better definition and operationalization of variables, and follow-up after treatment. Although there is room for improvement in research design, the overall body of research does endorse the utility and efficiency of group psychotherapy. The second proposed benefit of studying perfectionism in group psychotherapy is that this setting may be challenging for perfectionists. Piper (1991) observed that many patients are reluctant to try group therapy due to anxiety about revealing private and painful information to a group of strangers. Given the emotional and physiological arousal noted by Habke and Hewitt (2000) when their subjects disclosed about past mistakes to a single interviewer, one can imagine that this effect would be intensified in a group format. Dies (1993) acknowledged that selection procedures for groups are not highly developed. However, he described a number of characteristics that might make group psychotherapy more challenging, including three factors relevant to perfectionists. First, he suggested that group members need to be motivated and have reasonable goals and expectations. While this is probably true for all forms of therapy, it may be particularly important in groups where therapists may be less able to correct any given member's erroneous beliefs or expectancies. Perfectionists' elevated expectations of self and others may lead to considerable disappointment and frustration in group therapy. Next, Dies described two factors relating to interpersonal relationships within the group, cooperation with others and alliance. In light of the research suggesting that perfectionists can be hostile and domineering as well as have likely difficulties with intimacy, it may be quite challenging for them to form cohesive and constructive relationships with fellow members and therapists. Indeed, Zuroff et al.'s (2000) study confirms that perfectionists have difficulty developing an increasingly positive alliance over the course of therapy. The combination of anxiety about disclosure to the group, frustration over elevated expectations for self and others in the group, and difficulties in creating cohesive relationships with a number of different people could combine to make group psychotherapy quite challenging for perfectionists. In the group format, one might anticipate  PSP in Group Psychotherapy/page 24  slightly earlier interference due to perfectionism in contrast to the ninth session reported by Blatt et al. (1998). Despite the general level of support for groups in the literature, the obvious challenges faced by perfectionists in this format may appear daunting. However, some of these challenges can be addressed by appropriate therapist interventions, and others provide ideal therapeutic opportunities. For example, Piper (1991) and MacKenzie (1990) suggest that patient preparation through pregroup training sessions may alleviate many of their concerns about potential threats in the group. Similarly, Dies (1993) advocates the use of pregroup sessions to deal with negative expectancies and create reasonable, positive expectations. In addition, explicit instruction can be given on helpful client behaviours in group therapy such as disclosure, here and now focus, interpersonal feedback and anxiety management. Stumbling blocks and potential hurdles can be identified in hopes that such predictions will enable clients to raise concerns and work through them rather than being overwhelmed. Thus, appropriate preparation and continued monitoring by therapists may ease some anxiety. At the same time, many would argue that facing fears and challenging situations could have considerable therapeutic benefit. Toseland and Siporin (1986) suggested that groups are ideal for individuals who are socially or psychologically isolated and who have relationship concerns and conflicts. Hewitt and his colleagues (Flett & Hewitt, in press; Flynn, Hewitt & Flett, 1996) have theorized that perfectionism often develops when a child is uncertain of their parents' love and attention. In order to prevent neglect and abuse or to ensure affection, children conclude that good things may occur if they are, or are perceived as being, perfect. This 'rule' remains with them into adulthood. The group situation provides an opportunity to test that rule with a variety of different people. While a therapist expressing acceptance despite imperfections can be helpful, having a room full of 'real' people do the same could have an even greater impact. The group context provides a safe environment to test assumptions, work through conflict to achieve better understanding, and practice new styles of relating. For a perfectionist, this may include admitting to faults, expressing 'imperfect' emotions like anger and depression, and learning to  PSP in Group Psychotherapy/page 25  ask more directly for feedback about others' expectations and perceptions. Therefore, groups may be a particularly therapeutic context for highly perfectionistic individuals. Therapeutic Factors in Group Psychotherapy  Having selected group psychotherapy as the intervention in this project, it is essential to identify the primary therapeutic factors in group psychotherapy. Then, specific hypotheses can be enumerated concerning perfectionism's interference in this form of psychotherapy. According to Bloch and Crouch (1985; p. 4), a therapeutic factor is "an element of group therapy that contributes to improvement in a patient's condition and is a function of the actions of the group therapist, the other group members, and the patient himself." They are careful to distinguish therapeutic factors from 'techniques', which are specific interventions designed to help the group members, and 'conditions for change', which they view as necessary conditions for therapeutic improvement to occur. While this definition is useful, not all theorists' proposed therapeutic factors fit Bloch and Crouch's criteria (e.g., Yalom, 1995), nor do all theorists agree with their definition. Yalom (1995) states that therapeutic factors may be conditions for change. Bloch and Crouch's definition is offered here merely as a guideline. Yalom's (1995) list of eleven therapeutic factors is used frequently as a starting point for research in this area. His factors are: 1. Installation of hope 2. Universality 3. Imparting of information (including didactic instruction and advice) 4. Altruism 5. The corrective recapitulation of the primary family group 6. Development of socializing techniques 7. Imitative behaviour (or vicarious learning) 8. Interpersonal learning 9. Group cohesiveness 10. Catharsis 11. Existential factors. Bloch and Crouch (1985) argue that family recapitulation and existential factors are specific to one theoretical model. Perhaps these factors would fit better in their classification of techniques. In addition, these reviewers favour distinguishing between catharsis and self-disclosure. Whereas Yalom (1995) believes that the two are intrinsically connected, Bloch and Crouch (1985) state that they have different effects on the group member. Catharsis often leads to  PSP in Group Psychotherapy/page 26  feelings of relief and self-disclosure is linked to feeling open and honest. Despite these disagreements, Yalom's list has provided an important research impetus for the field (Bloch & Crouch, 1985). In addition to the list, Yalom developed a Q-sort technique in which group members rank various factors on a normal distribution. This technique has allowed researchers to speculate about which factors group participants perceive as most important. As noted by Bednar and Kaul (1994), therapeutic (or curative) factors vary between groups, types of client, and the stage of group development. However, some factors do appear to be more highly valued. In general, reviews of this type of research come to similar conclusions. Group members report that the following factors, though not necessarily in this order, are key to their therapeutic progress: cohesion, interpersonal learning, catharsis and self-understanding (Bednar & Kaul, 1994; Yalom, 1995). These factors can be combined to describe a potential sequence of therapeutic benefit in group therapy (Yalom, 1995). First, a member displays their pathology. Other members provide feedback and the member in question participates in self-observation. This process permits recognition of the impact of his or her behaviour on the feelings of others, others' perceptions of him or her, and his or her own self-perception. With this new knowledge, the member may express some emotion and choose to risk new ways of being. By taking this risk, the member learns that a feared event does not occur. Yalom adds the caution that these ratings come mostly from long-term outpatient psychotherapy groups, which may have different goals from other group formats. Returning to the plight of highly perfectionistic individuals, one can imagine that this personality trait could interfere with each of these key therapeutic elements. For this study, perfectionists' personal distress during group psychotherapy was selected as the focus. Thus, the factors of interpersonal learning and catharsis were most relevant to our investigation. Each of these factors and their relationship to therapy outcome will be described more thoroughly in the following sections.  PSP in Group Psychotherapy/page 27  Interpersonal Learning/Self-Disclosure Interpersonal learning or interpersonal work generally includes the concepts of both selfdisclosure and feedback from other group members (Tschuschke, MacKenzie, Haaser & Janke, 1996). Because this paper focuses primarily on what group members bring to therapy, selfdisclosure will be explored more thoroughly here. Bloch and Crouch (1985) define selfdisclosure as "a patient's direct communication of personal material about himself to other group members" (p. 127). Other researchers have been more specific about the content of these disclosures. Tschuchke and Dies (1994) state that such expressions must "refer to the self in the group situation or what the person thinks or feels about him/herself and in relation to others in the group" (p. 192). Thus, the self-disclosure field provides further ammunition for Bednar and Kaul's (1994) complaints about an ill-defined group psychotherapy literature that lacks conceptual clarity. Despite variations in the definition of self-disclosure, many researchers have pointed to its importance in mental health and psychotherapy. One early review noted that self-disclosure can be a presenting problem, a means of increasing adjustment and a goal of psychotherapy (Allen, 1974). Reviews also noted that the association between self-disclosure and adjustment or mental health appeared to be curvilinear (Allen, 1974; Cozby, 1973). That is, individuals who disclosed in the moderate range achieved the most positive adjustment. Cozby (1973) suggested that excessive self-disclosure causes an individual to lose his or her sense of individuality and privacy. Low levels of self-disclosure prevent connections to other people. Allen (1974) commented more on the interpersonal consequences of these forms of self-disclosure. Excessive levels may lead to rejection by the listener because the disclosing individual is perceived as threatening. Generally, there is reciprocity in levels of self-disclosure and the recipient of high levels of disclosure may feel obligated to respond in kind. Low levels of self-disclosure may lead to alienation. In addition, he identified a reinforcing cycle that may be experienced by the low self-discloser. Such an individual may perceive their conflicts, fears and inadequacies as unique. They choose not to disclose and therefore fail to test this assumption, consequently  PSP in Group Psychotherapy/page 28  reinforcing feelings of low self-esteem and distinctness from others. Thus, links between selfdisclosure and mental health are fairly well established. Within the therapy context, numerous studies and reviews have concluded that self-disclosure is linked to better outcome. One study used audiotapes of the first three therapy sessions for a sample of college males (O'Malley, Suh & Strupp, 1983). These researchers concluded that patient involvement was the most powerful predictor of therapy outcome. This association appeared to develop over the first three sessions and was most consistent in the third session. In a general review of therapy process and outcome, patient verbal activity and particularly patient openness were quite consistently associated with better outcome (Orlinsky et al, 1994). Although the literature generally agrees on the importance of self-disclosure, some studies have failed to find the expected association between disclosure and therapy outcome (see Kelly, 2000a for a review). This inconsistency suggests that explorations of individual differences between clients (e.g., level of distress, perfectionistic self-presentation), and between therapists (e.g., capacity to tolerate negative disclosures without judging), as well as qualitative elements of selfdisclosure (e.g., level of intimacy) may be important to fully understand this phenomenon (Hill, Gelso, & Mohr, 2000; Kelly, 2000a, 2000b) Self-disclosure may have an even greater impact in the group setting than individual therapy because members disclose to a more diverse and representative group of people rather than a single therapist. This may improve generalization from the therapy context to the real world (Allen, 1974). Dies suggests that "interaction among group members is the most direct mechanism of change in the therapeutic process" (p. 504) and clearly, such interaction is initiated by member self-disclosure. Some evidence has accumulated to support these assertions. When Bloch and Crouch (1985) reviewed the literature several years ago, they lamented that only sporadic, specialized populations had been assessed for the impact of self-disclosure. They observed that self-disclosure seemed to improve schizophrenics' behavioural ratings and benefit other inpatients, yet appeared to be linked to greater difficulties for delinquents. They concluded that it is important to specify the type of patients, the type of therapy, and the type of selfdisclosure, as varied results may occur in different settings. In the period since this review, a few  PSP in Group Psychotherapy/page 29  interesting studies have been conducted on long-term, inpatient psychotherapy groups. Tschuschke and his colleagues (1994; 1996) conducted a series of intensive analyses on two long-term inpatient groups consisting of neurotic and personality disordered patients. Outcome was assessed in terms of symptom reduction and goal attainment at 12 and 18 months posttreatment. Results indicated that more successful patients self-disclosed more during therapy sessions (Tschuschke & Dies, 1994). In addition, early interpersonal work, or self-disclosure and feedback in the first six months of therapy, was more strongly correlated with improvement than later interpersonal work (Tschuschke et al., 1996). One obvious limitation in this research was the very small sample size of 16 patients. In addition, the focus on long-term, inpatient samples may not be directly generalizable to the short-term, outpatient groups studied in this project. It is interesting to note that the curvilinear relationship between self-disclosure and benefit suggested by Allen (1974) and Cozby (1973) has not been identified within the psychotherapy context. It may be that psychotherapy demands a high level of self-disclosure and is one context in which negative consequences of such openness do not occur (for a review of the impact of context on associations between self-disclosure and liking, see Collins & Miller, 1994). In part, this may be due to the lack of reciprocity between client and therapist. That is, regardless of the client's level of self-disclosure, the therapist will not feel obligated to match it, and therefore may not experience the same discomfort that a listener might in a social context. Of course, within a group context, other group members might feel pressured to respond to a heightened level of intimacy. Therefore, a curvilinear relationship, with moderate levels of self-disclosure being ideal, may exist within the group context, but has yet to be empirically demonstrated. A recent series of articles reviewed research on the impact of self-presentation on therapy outcome, focusing primarily on the implications of self-concealment or keeping secrets within the therapy context (Arkin & Hermann, 2000; Hill et al., 2000; Kelly, 2000a, 2000b). Kelly (2000a) proposed that it may be necessary for clients to hold back some negative information about themselves in order to preserve the opportunity for positive feedback from their therapists. She suggested that therapeutic change stems from improvements to a client's self-concept as a consequence of positive self-presentations and positive feedback from his or her therapist. If a  PSP in Group Psychotherapy/page 30  client reveals objectionable information about him or herself, the therapist may be perceived, realistically or unrealistically, as responding negatively, a reaction that may then be internalized by the client. She concluded that some self-concealment may be necessary for positive therapy outcome. In recommending more moderate levels of self-disclosure, this theory is consistent with Allen (1974) and Cozby's (1973) descriptions of a curvilinear relationship between the degree of self-disclosure and health. Although the studies on self-disclosure may be rather specific, there is some evidence that this variable is relevant to therapy outcome. Other research hints that perfectionism may interfere with self-disclosure and thereby therapy success. Two reviews noted studies on social desirability and need for approval and concluded that these traits are linked to lower levels of self-disclosure (Cozby, 1973; Bloch & Crouch, 1985). These variables are also highly correlated with perfectionistic self-presentation (Habke & Hewitt, 2000; Parkin, Hewitt, Flynn & Flett, 1999). Similarly, in a project on social anxiety, anxious subjects failed to match the disclosure levels of experimental partners, instead choosing to self-disclose at moderate levels (Meleshko & Alden, 1993). These anxious participants also reported that they had fears of disapproval and were attempting to protect themselves in the interactions. Again, protective forms of perfectionistic self-presentation are strongly linked to social anxiety (Hewitt et al., 2000) and one may expect that these forms of perfectionism also will be associated with lower levels of disclosure in the highly intimate context of psychotherapy. Dies (1993) called for more work on discrete aspects of self-disclosure including frequency, depth, timing, content and valence. One study took this extra step and looked at intimacy of self-disclosures (Altman & Haythorn, 1965). They found that dominant participants self-disclosed more nonintimate information and less intimate information than less dominant participants. Thus, given the ties between perfectionism and dominance, one might expect perfectionists to show a similar lack of intimate selfdisclosure. Finally, recent work on self-presentation emphasizes that self-presentations in therapy must be believable, both to the client and the therapist (Hill et al, 2000; Kelly, 2000b). Since perfectionistic self-presentation requires the maintenance of an impossibly perfect image, it is likely that neither client nor therapist will view this presentation as credible. As such, any  PSP in Group Psychotherapy/page 31  advantage that might be gained from a positive self-presentation is lost when the client and therapist know that this image is false (Kelly, 2000a). Therefore, evidence on need for approval, social anxiety, and dominance, as well as from the self-presentation literature suggest that perfectionistic self-presentation may have a negative impact on achieving necessary levels of disclosure during therapy. Catharsis and Emotional Experience As was noted earlier, there is some debate over whether catharsis is a separate therapeutic factor (Bloch & Crouch, 1985) or is intrinsically associated with self-disclosure (Yalom, 1995). Bloch and Crouch (1985) define it as "emotional release.. .which brings some measure of relief' (p. 162). Although many authors (e.g., Grunebaum, 1975) and even group participants (Bloch & Crouch, 1985) advocate the expression of positive and negative feelings in therapy, most outcome research does not show strong connections between catharsis and therapeutic outcome (Bloch & Crouch, 1985; Yalom, 1995). One exception is the review by Orlinsky et al. (1994), which reported that in 63% of studies, patient expressiveness was associated with more positive outcome. Some variability may be explained by the notion that emotional expressiveness is only advantageous when combined with personal insight (Yalom, 1995) or cognitive learning (Bloch & Crouch, 1985). Other research has focused more on the experience of various emotions during therapy rather than on their expression. For example, McCallum, Piper and Morin (1993) investigated postsession affect in 12 psychoanalytically oriented short-term therapy groups. Ratings of positive and negative affect were made by the patients themselves, the therapists and fellow group members. Positive post-session affect predicted more favourable therapy outcome as judged by goal attainment and therapy usefulness ratings completed by the patients and therapists. Negative affect when accompanied by self-focused attempts at understanding also predicted more positive outcome. This last finding provides some support for the earlier assertion that catharsis may be beneficial when connected to insight and learning. Further evidence comes from Orlinsky's (1994) review paper, which reported on associations between negative affect and therapy outcome. In 35% of the findings, it led to poor outcome, in 39%, there was no  PSP in Group Psychotherapy/page  32  significant effect and in 25%, negative affect lead to positive outcome. Consistent with other work, positive affective responses were consistently tied to positive outcome. Based on Habke and Hewitt's (2000) work, it is likely that perfectionistic self-presentation will be associated with negative post-session affect. Their study found that protective forms of perfectionistic self-presentation predicted elevated ratings of negative affect following an assessment interview. Similar reactions are likely during group therapy sessions. Evidently, in and of itself, negative affect may not be detrimental to therapy progress. However, it is important for the experience and expression of negative affect to be worked through in an effort to lead to greater insight and self-understanding. Unfortunately, perfectionists also are likely to be uncomfortable with such discussion and may avoid the important psychological work that needs to occur. If so, these perfectionists may experience a less successful therapy outcome. Attendance and Attrition In the previous sections, factors were identified which appear to improve therapeutic effectiveness. However, before these factors can play their part, members have to be physically present. Attendance patterns can be used to predict early termination or attrition (Stone, Blaze & Bozzuto, 1980). In addition to limiting a departing member's opportunities for therapeutic benefit, there is some evidence that attrition occurs in waves in psychotherapy groups (Stone et al., 1980). That is, one member's departure seems to precipitate other premature terminations. Sadly, rates of attrition from group psychotherapy are quite high. Some research teams have reported rates of 21 to 33% (Connelly & Piper, 1989; Falloon, 1981; McCallum, Piper & Joyce, 1992). In a more dramatic presentation, Klein and Carroll (1986) tracked all individuals who were referred to their group therapy clinic. Of the 719 referrals in a two-year period, 41% never attended a therapy group. Of those who did attend, one quarter dropped out within the first four sessions and only 28% of the total sample attended more than 12 sessions. These high rates of attrition are alarming since treatment duration is positively correlated with outcome (Conte, Plutchik, Picard & Karasu, 1991; Luborsky et al., 1971). Thus, considerable attention has been devoted to factors that predict attendance and attrition from group psychotherapy.  PSP in Group Psychotherapy/page 33  Several studies and reviewers report that group cohesion and attraction to the group predict lower levels of attrition and better attendance (Bloch & Crouch, 1985; Falloon, 1981; Yalom, 1995). Other studies have attempted to identify client variables associated with group attendance. Falloon (1981) reported that the following characteristics predicted dropout: being a single, unemployed male, high levels of social anxiety and general psychological symptoms, and the lack of a supportive significant other. In general, these characteristics suggest someone who is unable to connect with others in their lives. Similarly, Stone et al.'s (1980) clinical observations of dropouts focused on problems of intimacy and self-disclosure. Saltzman and his colleagues (1976) assessed a variety of relationship dimensions in individual psychotherapy and concluded that clients who experienced little anxiety in an initial session and later, characterized the therapy relationship as lacking respect, understanding, openness, security, uniqueness and hope were more likely to terminate therapy early. Thus, clients who dropped out of therapy clearly were unable to establish a positive working alliance with their individual therapists. Piper, Ogrodniczuk, Joyce, McCallum, Rosie, O'Kelly and Steinberg (1999) used video coding of individual therapy sessions, and also concluded that weaker therapeutic alliance and less engagement in therapeutic work by the client predicted dropping out. In another investigation of 16 groups, low psychological mindedness, elevated psychiatric symptoms and patient ratings of the severity of their target therapy objectives predicted attrition from short-term groups (McCallum et al., 1992). This description appears to emphasize the degree of psychopathology and the patient's ability to make use of the therapy process. Attendance and attrition obviously are important predictors of therapy outcome. Low levels of group cohesion, perhaps due to member psychopathology and difficulties with intimacy and self-disclosure, appear to be the primary predictor of early termination. Again, this raises concerns for perfectionistic self-presenters who also seem to have trouble with intimacy and establishing connections with other people. Thus, in therapy, these types of perfectionists may attend fewer sessions and dropout more frequently than other participants.  PSP in Group Psychotherapy/page 34  Overview of Primary Objectives  In reviewing the literature, it is clear that there are many ways in which perfectionism, particularly the protective forms of perfectionistic self-presentation, could interfere with group psychotherapy process and outcome. Perfectionism may impair self-disclosure, catharsis and therapy attendance and attrition. This study seeks to assess that impact on short-term group psychotherapy for perfectionism. By focusing the treatment groups on perfectionism, individuals with higher levels of each form of perfectionism can be recruited. As described earlier, there are several forms of perfectionism that are related but not redundant with one another (Hewitt et al., 2000); thus, large variances are still maintained for each subscale within the study sample. This study was also careful to use a variety of measures to prevent some of the confounds that have been described in the literature. One problem can be the source of information. For example, a self-oriented perfectionist may be highly critical of his own performance in therapy and consequently may rate his own therapy outcome more stringently. For this reason, both process and outcome measures will be collectedfromseveral sources including therapists, observers, and close friends or family members of group participants. In addition, group members will make self-evaluations because self-perception can be important and informative. To further tap each individual's experience, both physiological and behavioural measures will be employed. Frequently, physiological measures of stress are not highly correlated with participants' self-report of stress (e.g., Curtis, Buxton, Lippman, Nesse & Wright, 1976; Walsh, Wilding & Eysenck, 1994). Therefore, it is important to assess both subjective and physiological responses in order to tap the breadth of the human stress response. In addition, this study investigates short- and long-term indicators of stress by testing heart rate and skin conductance, which respond to the sympathetic nervous system adrenal-medullary system (AMS) and Cortisol from the pituitary-adrenal cortical system (PAC; Arnetz & Fjellner, 1986). Studies have concluded that an unhealthy stress response involves prolonged activation of the AMS and any activation of the PAC (Dienstbier, 1989). Thus, prolonged heart rate and skin conductance elevations and any elevation of Cortisol will be of interest in this study. These physiological measures tap into different aspects of stress experience in this project. Heart rate and skin  PSP in Group Psychotherapy/page 35  conductance elevations were used as outcome measures to evaluate whether the treatment program was effective in moderating participants' stress reactivity. Flynn et al. (2001) demonstrated that perfectionists experience greater physiological arousal during these tasks and show a slower recovery following them. Therefore, intense and prolonged stress responses are typical of perfectionists and provide a target for change in psychotherapy. Heart rate and skin conductance reactions to achievement tasks were measured before and after the group therapy experience. In contrast, Cortisol served as an indicator of members' distress in anticipation of and following group sessions. Behavioural coding of self-disclosure from videotapes has been advocated in the group therapy literature because self-report is rarely correlated with actual performance (Allen, 1974). Assessment with multiple methods provided a more complete and accurate picture of how perfectionists experience group psychotherapy and the impact that it has on psychotherapy outcome. In addition to demonstrating links between perfectionistic self-presentation and therapy process and outcome, this study also sought to determine whether perfectionism was a unique predictor of these phenomena. Work by Hewitt et al. (2000) illustrated that perfectionistic selfpresentation is strongly associated with symptoms of both depression and social anxiety. Therefore, any effects noted in the present study between perfectionism and therapy process could simply reflect indirect associations due to the impact of higher levels of pre-treatment depression or social anxiety. In other words, individuals scoring highly on perfectionistic selfpresentation may do poorly in therapy simply because they are more depressed or socially anxious. Indeed, several of the studies already mentioned described connections between depression and social anxiety and the process variables to be studied in this project. Several research teams have commented on the fact that higher levels of general psychiatric symptomatology (such as depression and anxiety), as well as introversion predict more negative psychotherapy outcome (Garfield, 1994; Shea, Elkin & Sotsky, 1999; Zuckerman et al., 1980). Meleshko and Alden (1993) highlighted the fact that socially anxious individuals did not match self-disclosure levels of experimental partners, choosing instead to self-disclose at more moderate levels. Thus, socially anxious group members might be expected to self-disclose at a  PSP in Group Psychotherapy/page 36  lower rate than other participants. In addition, data on attrition has identified both individuals with social anxiety (Falloon, 1981) and those with higher levels of psychiatric symptoms (McCallum et al., 1992) as being more likely to terminate therapy early. Because both social anxiety and depression are associated with perfectionism, and with therapy process and outcome, these variables will be used as controls in this study's statistical analyses. Doing so will identify whether perfectionistic self-presentation is a unique contributor to therapy process and outcome, over and above these other signs of psychopathology. Another marker of the intensity of psychiatric distress is the use of psychiatric medications. Medications also can alter physiological responses such as heart rate and Cortisol readings. Therefore, the use of medications will be entered as a third control variable. Hypotheses Two general hypotheses will be tested in this study. In each case, basic associations will be tested first, followed by a more controlled analysis to test for the unique predictive power of perfectionistic self-presentation. In this second stage, social anxiety, depression and use of medications will be controlled, as will baseline levels of relevant variables for outcome analyses. 1. Protective forms of perfectionistic self-presentation should be associated with greater distress d u r i n g therapy.  As noted earlier, a variety of methods will be used to investigate this hypothesis. On selfreports, individuals with greater difficulty disclosing and displaying imperfections are likely to report higher levels of anxiety prior to therapy sessions and greater increases in anxiety over the course of each session due to their concerns about intimate disclosures. Whereas most members should show a small decrease in anxiety across sessions, those scoring high on nondisplay and nondisclosure of imperfections should continue to experience high levels of post-session anxiety as a result of the increasing intimacy of the group and growing expectations of openness. Physiologically, protective forms of perfectionistic self-presentation (P-PSP) should be associated with higher levels of anticipatory Cortisol when controlling for baseline Cortisol levels. As with the self-report measures, post-session Cortisol levels should not show as much adaptation  PSP in Group Psychotherapy/page 37  across sessions for those scoring highly on nondisplay and nondisclosure of imperfection as compared to other group members. Several behavioural measures can act as indicators of a group member's distress. P-PSP should be positively correlated with the number of late arrivals, missed sessions, and dropouts. In addition, these subtypes of perfectionism should predict fewer self-disclosures, particularly disclosures of a negative valence, focusing on the group, and of a highly personal nature. Finally, although perfectionistic self-presentation is likely to be correlated with the experience of negative affect during therapy, it is unlikely to be expressed or worked through due to fears of being perceived as imperfect. Thus, P-PSP should predict lower coder ratings of negative affect during therapy sessions.  2. Protective forms of perfectionistic self-presentation should predict less positive outcome at the end of therapy. First, each member's self-report of symptom reduction can be used to test this hypothesis. Higher levels of P-PSP should be linked to less reduction in trait perfectionism scores and less reduction in depression, anxiety and interpersonal problems. In addition, individual goal achievement ratings at the midpoint and end of therapy should be negatively predicted by P-PSP. Finally, the member's own rating of global improvement should be reduced in individuals with higher levels of P-PSP Secondly, physiological measures can be indicative of change during the course of therapy. High levels of P-PSP should be associated with greater physiological arousal during and after an achievement challenge task. Following the treatment program, P-PSP should predict less decrease in physiological reactivity, in comparison to pre-treatment reactivity. As noted earlier, physiological measures will include heart rate and skin conductance levels. Finally, others' ratings of member change also should reflect more limited improvements for perfectionistic self-presenters. Improvement ratings by close others and by the therapists should be negatively correlated with both the nondisplay and nondisclosure of imperfection. Close others assessed change in perfectionism levels, as well as global improvement. Ratings by therapists focused on overall benefit and were made at the end of the therapy program.  PSP in Group Psychotherapy/page 38  Exploratory analyses. Although few gender differences have been noted in clinical samples in the perfectionism literature in general (Hewitt et al., 2000; Spangler & Burns, 1999), or as it relates to psychotherapy (e.g., Spangler & Burns, 1999; Zuroff et al., 2000), tests were performed to investigate differences between male and female group members. In addition, mediational models were tested to explore whether P-PSP had an indirect effect on therapy outcome by interfering with therapy process. Similar models have been proposed by Hewitt and Flett (in press). They described models in which perfectionism leads to problematic behaviours and poor coping strategies, which in turn lead to elevations on measures of psychopathology. In the present study, process and outcome variables that demonstrated an association to P-PSP were used to construct mediational models. These models then were statistically analyzed to determine their significance.  METHOD Participants  Participants were recruited by posters at college campuses and community centers, newspaper articles in leading Vancouver newspapers and radio talk show interviews (Appendix A). Each of these communications advertised a group treatment program for trait perfectionism. They included brief descriptions of the perfectionism construct and common outcomes, and advised interested participants to contact our lab for further information. Two hundred and sixty people (98 men, 162 women) made initial contact with our lab, leaving some form of identifying information. Of these, 55 were unable to be reached, 30 later indicated that they were not interested in therapy, and the remaining 175 took part in a phone screen. During this phone assessment, potential participants were screened to rule out severe psychopathology, prior hospitalizations due to psychotic symptoms, as well as extreme difficulty with the English language as the groups were conducted in English. Only seven individuals were ruled out for these reasons. In addition, participants were informed of the approximate time scheduling of  PSP in Group Psychotherapy/page 39  groups to ensure their availability, and their difficulties with perfectionism were discussed briefly to ensure that they understood the nature of perfectionism and had not confused it with obsessiveness or obsessive-compulsive disorder. Of the 168 people approved for the assessment phase of the program, 127 scheduled and attended their appointment. Assessment sessions permitted further screening of subjects to select participants with high levels of at least one form of trait or self-presentational perfectionism (at least half a standard deviation above the mean calculated from community samples in Hewitt and Flett's previous research studies; all but seven participants met this criterion). Potential group members who demonstrated severe psychopathology (e.g., currently suicidal, psychotic, or antisocial personality disorder) or who had never had a close relationship with another person were not invited to join the treatment groups (15 participants). The latter criterion is one that has determined the success of group treatment in previous studies (Yalom, 1995). Participants who did meet the above criteria were assigned to individual treatment groups with some consideration given to ensuring a good mix of gender, age groups and diagnoses. Of the 105 participants offered positions in the groups, 72 actually attended at least the first group session . 2  Based on work by Green (1991), 72 group members (25 men and 47 women) should ensure that this project's regression procedures using the two protective factors of perfectionistic selfpresentation would have power statistics of .80 for a medium effect size. Participants' ages ranged from 20 to 66 years (x=41.5, a=10.33). The sample was fairly evenly divided between single (37.5%) and married (40.3%) individuals. The vast majority of the group members were Caucasian (87.5%). Despite their difficulties, 50% continued to work full-time, with a further 22.3% engaged in part-time or casual labour. Over half of the group members reported taking some type of medication (59.7%). To provide some indication of the proportion of the participants with severe difficulties with protective perfectionistic self-presentation, individuals were identified who scored more than one  A M A N O V A was conducted to test differences in perfectionism and psychological symptoms (depression and anxiety) between those participants who entered into the groups, and the remainder of those who were assessed. Overall, no differences were detected (F , =1.05, p>.40). However, the group sample showed slightly higher levels of self-oriented perfectionism and nondisplay of imperfection. Those who entered into the groups did not differ from those who did not on other measures of perfectionism, nor on ratings of anxiety and depression.  2  (7  16)  PSP in Group Psychotherapy/page 40  standard deviation above the clinical normative sample on each subscale (Hewitt et al., 2000). Approximately 30% of the sample reported strong concerns about disclosing imperfections. Thus, in the average group of ten members, three individuals were likely to have significant struggles with nondisclosure of imperfections. In contrast, a greater number of group members (44%) expressed intense fears about displaying imperfections. Roughly 18% of the sample scored highly on both forms of protective perfectionistic self-presentation. Approximately 30% of the sample scored below the clinical mean on nondisclosure of imperfections, and 13% below the mean for nondisplay of imperfections. Thus, some individuals who scored low on these dimensions also participated in the groups. These estimates must be treated with considerable caution since self-presentational perfectionism has been studied only as a continuous variable and cut-off scores have not been established to reflect meaningful differences between high and low scorers. Measures  Predictor Variables Perfectionistic Self-Presentation Scale (PSPS; Hewitt et al., 2000). This 27-item measure assesses the three forms of perfectionistic self-presentation: perfectionistic self-promotion, nondisplay of imperfection and nondisclosure of imperfection. This study focused on the latter two subscales both of which involve concealment. Participants rate each statement on a sevenpoint scale to indicate their level of agreement. Such statements include, "It would be awful if I made a fool of myself in front of others" (nondisplay of imperfection) and "I should solve my own problems rather than admit them to others" (nondisclosure of imperfection). These subscales have good internal consistency ranging from .83 to .91 for nondisplay of imperfection andfrom.78 to .86 for nondisclosure of imperfection (Hewitt et al., 2000). Self-presentation subscales are correlated but distinct from measures of trait perfectionism. For example, selfpresentation scores explain additional variance in measures of depression (Hewitt et al., 2000). Further validity is demonstrated by the fact that ratings of subjects by peers and therapists are highly correlated with subjects' self-ratings (Hewitt et al., 2000). Copies of measures can be found in Appendix B, with the exception of those scales that are copyrighted.  PSP in Group Psychotherapy/page 41  Multidimensional Perfectionism Scale (MPS; Hewitt & Flett, 1991b). This scale assesses the three trait dimensions of perfectionism, namely self-oriented, other-oriented, and socially prescribed perfectionism. Subjects are asked to make seven-point ratings on how much they agree with each of 45 statements such as: "When I am working on something, I cannot relax until it is perfect" (self-oriented perfectionism), "I have high expectations for people who are important to me" (other-oriented perfectionism), and "I feel that people are too demanding of me" (socially-prescribed perfectionism). Higher scores indicate greater endorsement of perfectionistic items. This measure has good reliability and validity (Hewitt & Flett, 1991b). The subscales are not significantly correlated with social desirability, but do correlate with ratings made either by people close to the subjects, or by clinicians. Furthermore, test-retest reliability over a three-month period was .88 for self-oriented, .85 for other-oriented and .75 for socially prescribed perfectionism, demonstrating that this scale measures a stable personality trait. Dependent Variables Measures of Distress Self-report. Mood Analogue Rating Scale. Analogue scales are an easy, reliable and quick way to make repeated assessments of participants' current moods (Aitken, 1970; Luria, 1975; Ramp ling & Williams, 1977). Although mood ratings in this study tapped into three forms of negative emotion: sadness/depression, tension/anxiety and frustration/anger, only ratings of anxiety were considered in the analyses. Participants simply placed a mark on each 10 cm line to indicate the degree to which they were experiencing each mood. Lines were anchored by the phrases "not at all" and "extremely". Values were assigned to each rating by measuring the distance from the left edge of the line to the patient's mark in millimeters. Behavioural measures.  Disruptive Behaviour Scales. Created for this study, these scales are simple charts completed by the group therapists at the end of each meeting. They recorded behaviours such as absence with or without an excuse, late arrivals, and incomplete questionnaires. Scores for each  PSP in Group Psychotherapy/page 42  participant were constructed by tallying the frequency of each of these behaviours. Records of dropouts also were maintained. Psychotherapy Self-Disclosure Coding System (Hewitt, Flynn & Low, 1999). This coding system is based on a Self-Disclosure Coding System developed in the marital intimacy literature (Waring & Chelune, 1983; Waring, Schaefer & Fry, 1994). This system has been described as reliable, clear and comprehensive (Tardy, 1988). The original Self-Disclosure Coding System included many important measures that give a more complete view of an individual's selfdisclosure. These measures included amount or frequency of self-disclosure, duration, rate, depth, affective manner, and self-reference rate. In addition, separatefrequencycounts were suggested for positive, negative and neutral self-references. Reports of reliability for the various measures are consistently high rangingfrom.76 to .999, and generally are over .95 (Waring & Chelune, 1983; Waring et al., 1994). While a useful system, no published article clearly defined each of these measures. Consequently, we operationalized these terms ourselves and added a few measures that are particularly relevant in a group therapy context. First, self-disclosure was defined as a self-reference that tells the listener something about who the person is, how they relate to others, or their affect. A speech segment that focuses on a single theme of affect, core beliefs or self-concept descriptors is equivalent to one self-disclosure. Amount provides a sum of the number of self-disclosures during a taped session. Valence is a key factor in self-disclosures and this coding system distinguishes between negative, positive, and neutral content in each self-disclosure. The personal relevance of the disclosure is tapped by the measure Depth of Revealing. This scale rates disclosures as marginally (1), moderately (2) or highly (3) personal. Coders were told to consider how difficult it seemed to be for the individual to make the disclosure and whether they have told others this information in the past. This evaluation was made for an entire speech period rather than each self-disclosure. Similarly, a second rating was made for each speech period to assess a primary concern of interpersonal group therapy, the Here and Now Focus. Coders evaluated whether the speaker was discussing content from within the group, outside of the group, both or neither. The final rating that was madefromthese tapes was an overall rating of Affect. As coders are unable to tell what each  PSP in Group Psychotherapy/page 43  member was feeling during the session, this rating is representative of expressed or displayed affect. Ratings are made on three, ten-point mood scales for anxiety, depression, and happiness. Three tapes were coded from each therapy group. An early tape was selected from either the second or third 'actual' group session. A midpoint tape was chosen from the fifth or sixth 'actual' session and a late therapy tape from the eighth or ninth 'actual' session. Selection was based on group attendance. If perfect attendance was not achieved in one of the two codable sessions, both tape options were coded to provide data for all group members. Ratings were made by mixed-gender teams of two trained undergraduate coders. Each team practiced on at least one complete tape from a non-selected session. Teams were requested to use consensus coding in which they must reach agreement on each rating. Inter-rater reliability was evaluated by having another undergraduate student recode three of the therapy tapes (representing 12.5% of the total coding sample) . Comparisons between raters 3  can be made by simply calculating bivariate correlations to determine whether scores vary in a similar manner between coders, or by calculating intra-class correlation coefficients (ICC) which account for differences in the actual number of behaviours coded. Because this study's analyses were correlational in nature and because a single coding team provided ratings of every subject for each code, simple bivariate correlations provide sufficient evidence of reliability of the codes. However, future uses of the measure may require that the quantities noted be reliable, so ICC values also are reported here. Reliability for codes of the total frequency of self-disclosure and valence was very good. Bivariate correlations ranged from .71 for neutral disclosures to .91 for negative self-disclosures. Total disclosures also were highly correlated between raters (.88). In addition, the ICC (3,1) values were promising (total frequency=. 86; positive disclosures=. 79; negative disclosures=. 83; neutral disclosures=. 69). In contrast, ratings for other elements of self-disclosure were less consistent. So few ratings of moderate and highly personal selfdisclosures were made, that reliability could not be calculated for these measures. Therefore,  Three female undergraduate students conducted this reliability check. One was responsible for coding the frequency of self-disclosures and the valence. The other two students worked together to code depth of revealing, here and now focus, and affect. The three tapes were randomly selected from sequential blocks of tapes coded by the original coding team. This procedure is an attempt to reduce the impact of rater drift and to reflect the reliability of codes throughout the protocol. 3  PSP in Group Psychotherapy/page  44  these ratings were not included in subsequent analyses. In terms of here and now focus, only ingroup disclosure demonstrated a strong bivariate correlation between raters (.87). The ICC (3,1) value was less robust (.48) suggesting that although raters noted a similar pattern of in-group disclosures between group members, they did not record comparable frequencies of these disclosures. Theoretically, both in-group disclosures and combined in-group and out-group disclosures meet the therapy objectives of identifying typical patterns of relating to others within the group context. Therefore, the reliability of combined 'in-group' and 'both in-group and outgroup' scores was assessed. This combined code showed reasonable consistency in variability between raters (bivariate correlation=.78) but a poor ICC (3,1) (.37). However, because this study is correlational in nature, this combined scale was still acceptable for use in subsequent analyses. Finally, observations of affect also suffered from poor reliability. The sole exception, ratings of sadness, resulted in a moderate bivariate correlation (.79) and acceptable ICC (3,1) value (.76). Because of poor reliability, observer ratings of anxiety and happiness were not used in this study. Physiological measure.  Salivary Cortisol. Following each saliva collection, samples were frozen in order to improve the viscosity of the sample by denaturing the mucin and trapping debris in the proteins (Ferguson, 1984; Riad-Fahmy et al., 1982; Vining & McGinley, 1985). Salivary Cortisol concentrations were evaluated using Neogen Corporation's (Lexington, KY, USA) ELISA (Enzyme-Linked Immunosorbent Assay) test kit. This kit was designed originally for use with serum but can be adjusted for use with saliva by diluting the serum standards 1:10 with phosphate-buffered saline (pH 8.0). This adjustment is required because salivary Cortisol concentrations represent the free unbound Cortisol concentrations in serum, which are about 810% of the total serum Cortisol concentration (Brien, 1980; Umeda et al., 1981). This sampling technique works by competition between Cortisol in the sample and an enzyme conjugate for binding sites on an antibody coated plate. The enzyme conjugate can be detected by the intensity of colour development following the addition of a substrate. Therefore, greater colour development suggests that more enzyme conjugate has bound to the plate, and that less salivary  PSP in Group Psychotherapy/page 45  Cortisol  was contained in the sample. Approximately eighteen percent o f the assays were run in  duplicate to account for intra-assay variation. The average intra-assay coefficient o f variation was 8.3%. All samples for a given participant were analyzed in a single batch to reduce variability within a sample set. Outcome Measures Self-report  Beck Depression Inventory. (BDI; Beck & Steer, 1987). This 21-item inventory assesses depression severity by inquiring about both cognitive and somatic symptoms. Each symptom or attitude is represented by four statements with increasing severity. Participants rate which statements from each set are most true o f them in the past week. Severity ratings are summed using the highest ranked statement for each item. Beck and Steer (1987) suggest that scores from 10-18 indicate mild to moderate depression,from19-29, moderate to severe depression, and from 30-63 severe depression. Years o f study with this instrument have demonstrated its strong internal consistency, concurrent and discriminant validity (Beck, Steer & Garbin, 1988). Beck Anxiety Inventory. (BAI; Beck, Epstein, Brown & Steer, 1988). This 21-item inventory assesses a variety of anxiety symptoms including subjective, neurophysiological, autonomic and panic symptoms (Beck & Steer, 1991). Participants are asked to rate the degree to which they have experienced each symptom over the past week using a four-point scale. Scores are summed to produce a total anxiety score rangingfrom0 to 63. Studies have demonstrated the test-retest reliability (Beck et al., 1988), and discriminant validity (Beck & Steer, 1991) of this scale. In addition, items on the BAI are discerniblefromthose on the BDI indicating that each scale measures a distinct form of distress symptomatology (Hewitt & Norton, 1993) Inventory ofInterpersonal Problems. (IIP; Horowitz, Rosenberg, Baer, Urefio & Villasefior, 1988). This 127-item inventory reviews a number of difficulties which people report in relating to others. Part one asks participants to rate how hard various behaviours are for them on a five point scale from 0=not at all to 4=extremely. Some o f these items include, "It is hard for me to...trust other people, make reasonable demands of other people, believe that I am lovable to  PSP in Group Psychotherapy/page 46 other people". Part two asks participants to make similar ratings of "things that you do too much". Some of these statements include, "I act like a child too much", "I am too sensitive to criticism", and "I try to change other people too much". Factor analyses have identified six primary areas of difficulty: assertiveness, sociability, intimacy, submissiveness, over responsibility, and being too controlling (Horowitz et al., 1988). These scales have good internal reliability ranging from .82 to .94 and good test-retest reliability (.80 to .90 over ten weeks; Horowitz et al., 1988). Particularly important for a psychotherapy study such as this, the IIP appears to be sensitive to therapeutic change and may continue to change even after initial shifts in mood and other symptomatology (Horowitz et al., 1988). Patient Ratings. Group members established three goals for themselves for the treatment period. Following session 7 and at the end of treatment, members were asked to rate their goal attainment on a ten point scale from "no progress" to "goal achieved". In addition, they made global ratings of their progress at the end of the group in response to the question, "How much do you think you benefited from this program?". These ratings were made using a seven point scale rangingfrom"did not benefit" to "benefited significantly". Other ratings.  Therapist Ratings. Therapists were asked to make a global rating of how much each patient benefitedfromthe program at the end of the treatment. Once again, they used a seven-point scale rangingfrom"not at all" to "very much". Other Ratings. In addition to completing a second set of perfectionism other-rating forms, an individual close to each group member will be asked to evaluate how much he or she benefited from the program using the same question and scale as the therapists. Physiological Measures.  Autonomic Reactivity. The Davicon MED AC System/3 monitors participants' heart rates and skin conductance levels using a pulse plethysmograph and skin electrodes fastened to the fingers of their non-dominant hand. These electrodes produce a constant current density of 3LiA/cm and dry gold-plated sensors determine the level of skin conductance. Once adequate 2  physiological data are being captured, participants are asked to read instructions on the computer  PSP in Group Psychotherapy/page 47  screen. Following a brief relaxation period, the computer announces a short IQ test involving four questions. At the end of this test, participants are again instructed to relax. Comparison of physiological arousal during these three periods provides information about the individual's reactivity to the IQ test stressor, as well as their ability to cope with this stressor and return to baseline arousal levels. Control Variable Interaction Anxiousness Scale (IAS; Leary, 1983). This 15 item scale was developed to assess the affective component of social anxiety in a variety of situations. Participants rate statements on a five-point scale to indicate how true each one is for them. Items inquire about situations such as casual get-togethers and talking to members of the opposite gender or people in authority. Responses are totaled to create a final score in which higher numbers indicate greater social anxiety. Coefficient alphas reflect good internal consistency of about .88 (Leary & Kowalski, 1993) and test-retest reliability over an eight-week period of .80 (Leary, 1983). Construct, discriminant and criterion-related validities have been investigated and suggest that this measure does assess social anxiety specifically (Leary & Kowalski, 1993) Therapy Format  This new therapy program combined knowledge of critical components of interpersonal group psychotherapy (MacKenzie, 1990; Yalom, 1995) and impressions of key ingredients in the treatment of perfectionists in individual psychotherapy from the clinical experience of Paul Hewitt, Ph.D., R.Psych. Hewitt's treatment approach is described in Hewitt & Flett (2000). Use of this interpersonal model meant that therapy focused on the interpersonal precursors and consequences of perfectionism rather than focusing directly on reducing perfectionistic behaviours (e.g., negative evaluations, stringent expectations, etc...). This approach is consistent with traditional and contemporary models of interpersonal therapy (Sullivan, 1953; Teyber, 2000). Therapists worked to keep group discussion in the "here and now". That is, group members were encouraged to explore their relationships and experiences within the group because these can be more fully explored than situations outside of the group. Therapists also emphasized expression of affect and interpersonal feedback between members. Such discussion allowed many of Yalom's (1995) essential therapeutic factors to be available to group members.  PSP in Group Psychotherapy/page 48  Some of these factors included: installation of hope, universality, altruism (an opportunity to help fellow group members), development of socializing techniques, imitative behaviour or vicarious therapy, interpersonal learning and the development of group cohesiveness. Much of the group therapy literature now advocates the use of preparatory sessions prior to beginning 'actual' therapy sessions (e.g., MacKenzie, 1990). Although Bednar and Kaul (1994) call pregroup training one of the most potent factors leading to successful outcome, most reviews have only noted improvement in therapy attendance (Piper & Perrault, 1989; Zimpfer, 1991). Yet Piper and Perrault (1989) assert that such training has the long-term goal of assisting members to remain in therapy, in addition to improving therapy process and outcome. These authors remain optimistic that future studies using larger sample sizes and more specified forms of training may provide evidence that these goals are being met. The present therapy program incorporated many of the ideas suggested in the literature, with the provision of information about perfectionism in particular. Two one and one half hour "pre-group" sessions introduced group members to the interpersonal model, a therapy contract and how to get the most out of group therapy. Structured exercises provided time for initial introductions and practicing nonverbal listening skills. At the second meeting, leaders reviewed research on the development of perfectionism and its related problem areas. Members were asked to generate lists of the positives and negatives of being perfectionistic and then were introduced to the differences between striving for excellence and perfectionism. This exercise appeared to reduce some resistance to changing one's perfectionism. In addition, members privately completed sentence fragments to explore some of their motivations for being perfect and fears of imperfection. Groups ran for eleven weeks. Both "pregroup" sessions were held during the first week. "Actual" group sessions were held once each week for ten weeks. Sessions were an hour and a half in length, which is a common duration of group sessions (e.g., Piper, 1991). Each group was composed of seven to ten members and a male and a female co-therapist. Co-therapists were senior level graduate students in clinical psychology working towards their Ph.D. Two male and three female therapists took part in the study . While a slightly large number for short-term 4  The author was forced to act as a co-therapist for two of the eight therapy groups, after one of the female therapists withdrew from the program. In order to explore whether knowledge of the study caused perceptible changes in  4  PSP in Group Psychotherapy/page 49  therapy, ten members are considered to be an acceptable group size (Yalom, 1995). Groups were closed, meaning that no new members were assigned once the group had commenced. As is recommended by MacKenzie (1990), members who expressed a desire to leave the group were invited to an individual session with the group leaders to assess their distress level and explore their areas o f concern. Members were encouraged to attend at least one final group session in order to say goodbye to the other group members and to allow the group time to process their departure. Procedure  Assessment During initial phone contact, participants were asked to refrain from flossing their teeth for the twelve hours prior to their assessment session and subsequent therapy sessions because flossing can cause gums to bleed and thereby interfere with salivary Cortisol measurement. All other precautions needed to ensure accurate physiological measurements (e.g., no eating, drinking, teeth brushing, smoking or exercising for 1 hour before assessment; Hellhammer et al., 1987; Landon, 1982; Vining & McGinley, 1985) were monitored during the lengthy assessment session. Upon arriving at the lab for the initial assessment, participants were asked to complete a consent form for both assessment and treatment. They then were given the MPS and PSPS to complete. These measures were given first so that they could be quickly scored to ascertain whether the participant was sufficiently perfectionistic to join the. group. Those participants who did not meet this criterion were offered an abbreviated assessment such that referral suggestions  therapist behaviour toward group members scoring highly on different aspects of perfectionism, group members from four groups (two in which the author was co-therapist, and two comparison groups) were asked to make ratings of therapist behaviour and attitude. Since some differences would be expected between co-therapists (e.g., one may be more encouraging of quiet group members), co-therapy teams were analyzed to determine whether the author was judged as being more different from her co-therapist than another co-therapy pair. Correlations were calculated between each type of perfectionism and each category of therapist rating (e.g., invokes anxiety, encourages input, addresses conflict) for each therapist. These correlations were compared within co-therapist teams. The author and her co-therapist did not differ significantly on any of these correlations. However, the comparison co-therapy team did differ on ten of the forty correlations. Thus, the author invoked reactions in various types of perfectionists that were more similar to her co-therapist than did therapists within the comparison team. As a result, data from the author-led groups were retained for use in the study. However, for those groups, only the male co-therapist's ratings of group members were used in therapist evaluations of benefit from the therapy program.  PSP in Group Psychotherapy/page 50  could be made. While the perfectionism measures were being scored, participants completed the BDI, BAI, and IAS. After completing these measures, participants engaged in an unstructured clinical interview with the author to assess their suitability for group treatment. As described earlier, participants were screened out if they demonstrated severe psychopathology, an inability to establish relationships with other people, or insubstantial problems with perfectionism. Those participants who continued to be appropriate for treatment were presented next with a more comprehensive clinical battery of questionnaires including the Personality Assessment Inventory (Morey, 1991) to assess a broad range of psychopathological symptoms, and the Inventory of Interpersonal Problems (Horowitz et al., 1988) which focuses on interpersonal difficulties. Participants also were monitored physiologically during an achievement challenge task. Finally, participants were guided through a fifteen-minute version of Jacobson's Progressive Muscle Relaxation (Barlow & Craske, 1994). This relaxation session was used to compensate for anxiety caused by the assessment procedure, in order to obtain a salivary Cortisol sample that would reflect baseline levels of the hormone. In addition, it ensured that all participants would experience the same non-stressful stimuli prior to the Cortisol sampling. Following a brief introduction to the technique, participants listened to a fifteen-minute tape which instructed them to tense, then relax various muscle groups. Other researchers have found this technique to be effective in decreasing autonomic tonus and inducing feelings of relaxation (Arnetz, Fjellner, Kallner & Eneroth, 1985). Next, the "assessment session" salivary Cortisol sample was taken. This sampling occured at the end of the assessment session in order to be more consistent with circadian rhythms at subsequent testings during group treatment, given that groups ran during the evening . 5  Participants were told to suck their cheeks and lips to increase saliva production (Landon, 1982). If this was insufficient, citric acid was available for subjects to touch against their tongues.  Cortisol concentration levels can vary plus or minus 100% from the mean throughout the day (Moore-Ede, Sulzman & Fuller, 1982). Levels are higher and less reactive in the morning (Weitzman, Fukushima, Nogeire, Roffwarg, Gallagher & Hellman, 1971). 5  PSP  in Group  Psychotherapy/page  51  Following the assessment, participants were given feedback about primary areas of difficulty in their lives and their suitability for group treatment. Subjects either were assigned to a treatment group or given referrals for more appropriate programs for their concerns. Subjects who planed to enter the treatment group were given packets of questionnaires to give to a close friend or relative. These "other ratings" provided additional support for the participants' selfreport ratings of perfectionism. We anticipated that these ratings would closely match those of the group participants (Hewitt et al., 2000). Therapy At the beginning of each therapy session, group members were asked to complete the mood visual analogue measure. Following each session, members once again filled in mood visual analogue scales. At the beginning of session 1 and end of sessions 2,3 and 11, members provided saliva samples for Cortisol analysis. Video-recordings of three selected sessions were coded by two trained observers for frequency and quality of self-disclosures. In session 2, members decided on three primary goals for therapy. At the end of session 7, which falls at the midpoint of the actual group therapy sessions, members were asked to rate their degree of goal attainment for each of their goals on a ten-point scale. A similar assessment was made at the end of session 12.' At the end of the last session, group members were given packets to pass on to the close friend or family member who completed initial perfectionism ratings. This second packet included the same perfectionism measures and instructions to return the questionnaires to the group member to bring in to their post-therapy assessment. Therapists were asked to complete another set of questionnaires following each session. This set included ratings of disruptive behaviours (absence, lateness, incomplete questionnaires). At the end of the therapy program, therapists rated each group members' overall benefit from the program. Post-therapy assessment The focus of this assessment was to evaluate progress made during the course of therapy. Group members completed perfectionism measures, the MPS and PSPS, and measures of psychopathology symptoms including the BDI, BAI, and the IIP, as well as a clinical interview.  PSP in Group Psychotherapy/page  52  Participants also repeated the achievement challenge task with a second set of questions used for the IQ test. Finally, members evaluated their progress during the course of therapy. Statistical Analyses  For the first hypothesis, statistical analyses occurred in three stages. First, bivariate correlations were computed between each form of protective perfectionistic self-presentation (PPSP) and the process variables of interest. For those correlations that were significant, an initial regression analysis was conducted to determine which subscales of P-PSP were most predictive of each variable. If the regression results continued to indicate a significant relationship between perfectionistic self-presentation and a process variable, a more stringent hierarchical regression analysis was conducted to determine the unique contribution of this style when controlling for the related constructs of depression, social anxiety, and use of medications, as well as trait perfectionism. "Related" control variables were entered as a first block, followed by trait perfectionism on the second block, and perfectionistic self-presentation on the third block. In evaluating the results of the second hypothesis concerning outcome, all statistical tests had to control for pre-therapy levels of the outcome variable in question. First, partial correlations were computed between perfectionistic self-presentation and each outcome variable while controlling for pre-therapy scores on that variable. Once again, regression analyses were used to explore the predictive power of perfectionism for each of these outcome measures. Pre-therapy scores on the outcome variable in question were entered in the first block, followed by protective self-presentational perfectionism. As was done in the analyses for hypothesis one, if significant, these analyses were repeated while controlling for related variables that were not already in the equation and trait perfectionism . 6  Recently, group based research has been criticized for not taking into statistical consideration differences between groups in a sample. Investigations of individual difference variables can be influenced by group level effects inherent in a nested design. Without controlling for group differences, conclusions regarding individual differences may be erroneous. Ideally, we would need to establish that the effects of perfectionism on the process and outcome variables of interest are consistent across groups. To do so, hierarchical regression analyses would need to be performed for each dependent variable, controlling for the main effects of perfectionism and group on the first step, and entering the interactions between group and perfectionism on the second step. With eight groups, seven dummy coded variables would be need to be formed to test the main effect, and fourteen different interaction terms to assess effects on nondisplay and nondisclosure of imperfection. Thus, a minimum of 23 predictor variables would be needed to evaluate this assumption of equality between groups. Clearly, this study did not have a large enough sample size to accurately evaluate such a large number of predictors. Therefore, alternative methods for addressing this question had to be found.  6  PSP in Group Psychotherapy/page 53  Finally, it also was important to illustrate the proposed mediational model. That is, given negative associations between P-PSP and outcome, tests must determine whether one mechanism for this may be specific problems with therapy process. This model was examined for each process variable that was robustly linked to P-PSP, and for each outcome variable negatively associated with P-PSP. Using Baron and Kenny's (1986) approach to mediational analysis, hierarchical regression analyses were employed to identify mediators. This technique will be explained in more detail while presenting the results. Due to the paucity of previous research on perfectionism in the psychotherapy context, it was difficult to narrow the focus of this study to a few specific variables of interest. In allowing for the exploratory nature of this work, several different measures of distress and outcome were included. While such breadth is one strength of this study, the risk of Type I error increases with the number of tests conducted. Each sub-hypothesis within the project (e.g., P-PSP is expected to be linked to self-reported anxiety) can be considered as a separate study, within which the number of tests should be weighed in the interpretation of bivariate results (Cohen & Cohen, 1983). Because each hypothesis made specific predictions about the direction of association, one-tailed significance levels can be used (p<.10) and divided by the number of analyses within the section. Correlations that meet this conservative significance level are indicated in b o l d in the tables. Because of the complex nature of this study, including the number of measures and time points involved, it must be acknowledged that very few of the tests reach this ideal statistical significance standard. To some extent, this must be expected given the small effect F i r s t ,aM A N O V A was conducted to assess whether any differences existed in self-reported perfectionism b e t w e e n t h e e i g h t g r o u p s . N o s t a t i s t i c a l l y s i g n i f i c a n td i f f e r e n c e w a s f o u n d ( F =1.02,p>.40). Because groups did not differsystematicallyon perfectionism,differencesbetween the groups on process and outcome variables should have added error to the data making itmore difficultto reject null hypotheses. Therefore, the findings that are most l i k e l y t o b e e r r o n e o u s d u e t o g r o u p d i f f e r e n c e s a r e t h o s e h y p o t h e s e s w h i c h w e r e n o t c o n f i r m e d . Af a c t o r a n a l y s i s was conducted for most of these variablesand one primary factorwas identified.Ratings by friendsand family could not be included in this analysis due to alow return rate of these questionnaire packets by members of group 8. T h e c o m m o n f a c t o r r o u g h l y r e f l e c t e d n e g a t i v e o u t c o m e a n d e l e v a t e d s a l i v a r y Cortisol f o l l o w i n g g r o u p s e s s i o n s . A n ANOVA was conducted to see whether the groups differedon thisfactor(F = 3 . 5 3 ,p < . 0 1 ) . P o s t - h o c a n a l y s e s indicated that group one differed from groups six and eight on thisoutcome factor. However, thisdoes not n e c e s s a r i l y m e a n t h a t t h e s e g r o u p s s h o w d i f f e r e n ta s s o c i a t i o n sb e t w e e n p e r f e c t i o n i s m a n d t h e o u t c o m e v a r i a b l e s . Hierarchical regression analyses were repeated for measures that loaded over .40 on this factor,but without including data from the extreme groups (1,6 and 8). Associations between the outcome variablesand P-PSP remained non-significant,with the exception of the partialcorrelationbetween nondisplay of imperfection and posttreatment interpersonalproblems, controlling for pre-treatment interpersonalproblems. Without the extreme groups i n c l u d e d , n o n d i s p l a y o f i m p e r f e c t i o n w a s a s s o c i a t e d w i t h ag r e a t e r r e d u c t i o n i n i n t e r p e r s o n a l p r o b l e m s o v e r t h e course of treatment (r =-.34,p<.05). ( 5 3 5 )  ( 7 3 7 )  ( 3 4 )  PSP in Group Psychotherapy/page 54  sizes typically noted in research on individual differences in psychotherapy process and outcome (Garfield, 1994). Recent literature has criticized the requirement of strict probability levels, and cautions of the comparable risk of making a Type II error, in failing to reject a false null hypothesis (e.g., Hammond, 1996; Hunter, 1997). Thus, a loosening of these strict rules was permitted in conducting multivariate analyses provided a minimal level of significance (p<.10) was attained at the bivariate level. Analyses conducted on trend level effects (.05<p<.10) are clearly flagged as exploratory, and as requiring cautious interpretation. Interpretation of trend level effects is not unusual in other published work in this area (e.g., Blatt et al., 1995). Effect sizes (R ) are noted for all multivariate analyses, and these provide an alternate indication of the 2  magnitude and importance of the effect being studied. RESULTS Data were examined using procedures based on Tabachnick and Fidell (1989). First, missing data were identified and in most cases replaced by subscale means (unless more than 10% was missing from that set in which case the cell was considered missing). Next, distributions of variables with considerable skew and kurtosis were transformed in an attempt to normalize their distributions and eliminate outliers. The following transformations successfully normalized the shape of the distributions and removed outliers: Using square root: all disruptive therapy behaviours; and using LN(x+l): Cortisol at assessment, prior to first pregroup session, and following the second pregroup session and the ninth 'actual' session. Extreme outliers on the physiological measures were eliminated from the analyses due to the likelihood of measurement error. One participant's post-treatment skin conductance readings were deleted, in addition to three Cortisol data points from two other participants. Although slight skew and outliers were noted in other variables, these were not transformed in order to facilitate interpretation of the results. Multivariate outliers were tested by calculating Mahalanobis distance for each case. One case was deleted from evaluations of post-treatment interpersonal problems. The means, standard deviations, and coefficients alpha for each self-report measure can be found in Table 1. All scales demonstrated adequate internal consistency. Both trait and perfectionistic self-presentation subscales were tested for gender differences, but none were  PSP in Group Psychotherapy/page 55 noted (F =0.91, p=.481). All self-report measure means for this sample were compared to (565)  published normative clinical samples. As this was the first study to recruit individuals with high levels of perfectionism, average perfectionism scores for this sample were significantly higher than those reported in other clinical settings (self-oriented: t =8.17, p<.001; other-oriented: (332)  t =9.53, p<.001; socially-prescribed: t =5.17, p<.001; nondisplay: t (332)  (332)  perfectionistic nondisclosure: t  =6.56, p<.001;  (1110)  =4.07, p<.001; Hewitt & Flett, 1991; Hewitt et al., 2000).  (1110)  Although this sample scored more highly on measures of interpersonal problems (t =2.31, (172)  p<.05) and social anxiety (t =6.91, p<.001) than normative groups (Horowitz et al., 1988; 7  (g54)  Leary & Kowalski, 1993), average scores on depression and anxiety indices were lower than normative clinical samples (t =-4.56, p<.001 and t =-4.11, p<.001 respectively; Beck & (317)  (227)  Steer, 1987; Beck et al., 1988). In part, this lower level of subjective distress may be explained by the fact that these participants chose to seek assistance from a university-based research program rather than a hospital clinic. Correlations between P-PSP and trait perfectionism, depression, social anxiety, and the use of medications are presented in Table 2. As expected, strong positive associations were noted between the P-PSP variables and all three forms of trait perfectionism, as well as the other control variables . As expected, nondisplay and nondisclosure 8  of imperfections were highly, but not perfectly correlated with one another (r =.51, p<.001). (71)  Hypothesis 1: Distress during therapy. Several methods were used to investigate this hypothesis including self-report of anxiety, salivary Cortisol levels and behavioural measures. Analyses were conducted separately for each of these dependent variables. Self-report of anxiety. First, correlations were computed between protective forms of perfectionistic self-presentation and anxiety ratings before and after sessions at various points in As no clinical norms were available for the IAS, comparisons were made against an undergraduate university sample. Previously, scores on the impression management scale of the Balanced Inventory of Desirable Responding (Paulhus, 1991, 1994) also were included as control variables because impression management seemed to represent a broader class of presentational style. However, impression management was not positively correlated with either form of protective perfectionistic self-presentation. In fact, it was negatively associated with nondisplay of imperfection (r =-.26, p=.029; nondisclosure of imperfections: r =-.06, p=.615. Because impression management did not relate to perfectionistic self-presentation in the expected manner, it was excluded from the paper. A l l analyses were conducted with impression management as an additional control measure, however this did not alter any of the findings reported in this paper. 7  8  (7l)  (71)  PSP in Group Psychotherapy/page 56  treatment (Table 3). Nondisplay o f imperfections was associated with higher ratings of anxiety both before and after sessions, particularly early in therapy. In contrast, nondisclosure of imperfections was not linked to pre-session anxiety, but was correlated with post-session anxiety late during the course of therapy. To determine whether these P-PSP variables were linked to change in anxiety over the course of therapy sessions, hierarchical regression analyses were conducted for each time period in therapy to predict post-session anxiety (Table 4). On the first step, pre-session anxiety was entered to account for general levels of anxiety. In each case, this accounted for a significant portion of the variance in post-session anxiety. This was followed by the second step, addition of the two P-PSP variables. Across all sessions, there was a trend for nondisclosure o f imperfections to predict higher levels of post-session anxiety. When different time periods in therapy were distinguished, it became evident that this relationship was significant for sessions late in the course of therapy. A second set of regression analyses was conducted for these two time periods in order to control for related variables such as social anxiety, depression, the use of medications and trait perfectionism. As before, pre-session anxiety was entered into the equation on the first step, followed by related variables on the second step, trait perfectionism on the third step, and finally P-PSP. Across all sessions, depression was a significant predictor o f post-session anxiety (Table 5). Nondisclosure of imperfections no longer was a unique predictor o f post-session anxiety across all sessions. However, when only sessions late in the course of therapy were considered (Table 6), neither related variables nor trait perfectionism were significant predictors of postsession anxiety. Nondisclosure o f imperfection continued to be a strong predictor of post-session anxiety. Salivary Cortisol. A similar sequence of analyses was conducted to investigate the association between P-PSP and salivary Cortisol. First, correlations were computed between these variables (Table 7). The only significant correlation was a positive association between nondisclosure of imperfections and high assessment levels of salivary Cortisol. A hierarchical regression analysis clarified whether perfectionistic nondisclosure was a unique predictor of  PSP in Group Psychotherapy/page 57  assessment Cortisol levels, even after controlling for the time of day that the sample was taken. Time of testing was entered into the regression equation on the first step and showed a trend to predicting C o r t i s o l levels (R =.050, F =3.39, p=.070, (3= -.22). P-PSP variables were entered 2  (165)  on a second step (R Change=.088, F =3.19, p=.048), and nondisclosure of imperfection 2  (263)  continued to predict higher C o r t i s o l levels during the assessment session (P=.35, t=2.53, p=.014). A subsequent regression analysis controlled for related variables on the second step, and trait perfectionism on the third step (Table 8). Despite a strong positive association between depression and  levels, perfectionistic nondisclosure continued to be a unique predictor of  Cortisol  elevated assessment salivary C o r t i s o l . This suggests that individuals who scored highly on nondisclosure of imperfections were more likely to experience physiological stress during the assessment session. A number of specific hypotheses were made concerning salivary C o r t i s o l , and these were tested with hierarchical regression analyses. First, tests were conducted to evaluate whether PPSP was associated with greater anticipatory stress prior to the first group session. Anticipatory anxiety prior to the first pregroup session was used as a dependent variable, controlling for individual differences in Cortisol levels by inserting Cortisol levels from the assessment and time of assessment sample on the first step. Assessment C o r t i s o l was a strong predictor of anticipatory Cortisol  (First block: R =.224, F =8.39, p<.00T; assessment C o r t i s o l p=-48, t=4.06, p<.00T). 2  (258)  Next, P-PSP variables were entered but these did not contribute significantly to the variance in anticipatory Cortisol levels (R Change=.018, F =0.68, p=.510). Thus, P-PSP did not appear to 2  (2S6)  predict greater arousal in anticipation of the first group therapy session. The next two regression analyses tested whether P-PSP had a stronger effect on C o r t i s o l at various times during group therapy. First, Cortisol following the first 'actual' group session was the dependent variable. Cortisol samples following the second "pregroup" session were used as a baseline to control for individual differences in Cortisol levels. This sample was an appropriate control since it was taken at the same time of day as the test sample, but followed a structured, and less stressful pre-group training session. On the first step, Cortisol following the second pregroup session was entered, and this variable explained significant variance i n  Cortisol  levels after  PSP in Group Psychotherapy/page 58  the first 'actual' group session (R =.299, F =26.00, p<.001, p=.55). Next, P-PSP variables 2  (161)  were entered into the equation, but these did not contribute further (R Change=.020, F =0.87, 2  ( 259)  p=.424). Next, to test whether an association between P-PSP and Cortisol might appear later in the course of therapy, C o r t i s o l following the ninth 'actual' session was used as the dependent variable, again controlling for levels of  Cortisol  after the structured pregroup session on the first  step of the analysis. Once again, the pregroup sample was a strong predictor of  Cortisol  levels  following a session late in the therapy program (R =.428, F =35.94, p<.001, p=.65). P-PSP 2  (148)  variables were entered on a second step, but were not significant predictors of post-session Cortisol  (R Change=.010, F =0.39, p=.678). Therefore, P-PSP was not a predictor of 2  (246)  Cortisol  stress response at any of stage in the therapy program. Disruptive therapy behaviours. Table 9 reports the correlations between P-PSP variables and five forms of disruptive therapy behaviour: not showing up to sessions with an excuse, not showing up to sessions without an excuse, late arrivals to sessions, and incomplete questionnaires, as well as attendance overall. Too few participants dropped out of the program (N=5) to use this variable in statistical analyses. A significant positive correlation was found between nondisclosure of imperfections and not showing up to sessions without providing an excuse. A regression analysis established that when both P-PSP variables were used as predictors, nondisclosure of imperfections continued to be a unique predictor of not showing up without providing an excuse (As a block: R =.067, F =2.44, p=.095; perfectionistic 2  (268)  nondisclosure P=.28, t=2.07, p=.043). A final hierarchical regression analysis controlled for related variables on the first step, and trait perfectionism on the second step (Table 10). Nondisclosure of imperfection continued to demonstrate a trend to contributing unique variance in the frequency of not showing up to group sessions without an excuse. Psychotherapy self-disclosure coding. In order to investigate whether P-PSP decreased the frequency and intensity of self-disclosures in therapy, correlations were calculated between PPSP and elements of self-disclosure at different time points in the therapy program. Table 11 presents correlations between the P-PSP variables and various aspects o f self-disclosurefroman early session (second or third 'actual' session). Nondisclosure o f imperfections was significantly  PSP in Group Psychotherapy/page 59  correlated with disclosures of a negative valence, and observer ratings of sadness. In addition, nondisclosure of imperfections showed a trend to predicting more frequent in-group or combined in-group/out-group self-disclosures. Nondisplay of imperfections was not significantly associated with any of the coded variables, although it did show a trend to being linked to displays of sadness. A similar set of correlations was computed for a session late in the course of N  therapy (Table 12; eighth or ninth 'actual' session). Nondisclosure of imperfections was not correlated with any of the coded variables. Similarly, nondisplay of imperfections showed only a trend to being associated with more negative self-disclosures. In an effort to simplify interpretation of this data, the total frequency of self-disclosure was calculated for each coded video. Nondisclosure of imperfections was positively correlated with more frequent self-disclosures, but only in sessions early in the course of therapy (Table 13). Table 13 also shows correlations with self-disclosures of different valence, and those that focused on in-group or combined in-group/out-group events and reactions. None of these correlations was statistically significant. Observer ratings of sadness also were averaged across all coded sessions. No associations were noted between perfectionistic self-presentation and observed mood. Next, a series of regression analyses were conducted to elucidate the unique contributions of the P-PSP variables. First, P-PSP was used to predict the total number of disclosures across all sessions. However, these variables did not show an association with total disclosures (R =.002, 2  F  =0.05, p=.950). Next, changes in the frequency of disclosures were evaluated between late  ( 2 63)  and early sessions in the course of therapy. Disclosure frequency late in therapy was the dependent variable, and early disclosure was entered on the first step, followed by the P-PSP variables on a second step. As would be predicted, early disclosure frequency strongly predicted late disclosure frequency (R =. 126, F =9.27, p=.003, (3=36), but the P-PSP variables showed a 2  (164)  trend to contributing additional unique variance to this equation (R Change=.074, F =2.89, 2  (262)  p=.063). Although caution is required in interpreting such trends, it is interesting to note that nondisplay and nondisclosure of imperfections demonstrated opposite effects. Nondisplay predicted higher frequency of late disclosures, when controlling for early levels (P=.27, t=2.04,  PSP in Group Psychotherapy/page 60  p=.045), whereas perfectionistic nondisclosure predicted a lower disclosure frequency late in therapy ((3=-.29, t=-2.14, p=.036). The perfectionistic nondisclosure effect remained significant even after controlling for related variables and trait perfectionism on the second and third steps (Table 14). A similar series of statistical analyses was repeated for each of the self-disclosure descriptives (valence and focus of self-disclosures). First, tests were conducted to evaluate whether P-PSP predicted fewer negative self-disclosures. As a unit, the P-PSP variables were not strong predictors of the total frequency of negative self-disclosure (R =.015, F =0.48, 2  (263)  p=.619). Next, analyses were conducted to test the hypothesis that P-PSP would predict fewer negative self-disclosures late in therapy, controlling for the frequency of negative self-disclosures early in therapy. Early frequency of negative self-disclosures was a strong predictor of late disclosures of this valence (R =.184, F =15.68, P=.44, p<.001). Consistent with the 2  (li64)  hypothesis, P-PSP predicted additional unique variance (R Change=.l 18, F =5.34, p=.007). It 2  (262)  is interesting to note that nondisplay and nondisclose have opposite effects on this variable ((3=35, t=2.89, p=.005 and p=-.36, t=-2.78, p=.007 respectively). In other words, nondisplay of imperfections predicted a relative increase in negative self-disclosures from early to late sessions in therapy, whereas nondisclosure of imperfections predicted a relative decrease in disclosures of this valence. Given the similar pattern of findings for nondisclosure of imperfections and total disclosures, it was essential to establish whether this effect was specific to negative selfdisclosures, or whether it reflected a more general decrease in any form of disclosure for group members with concerns about disclosing imperfections. One method of addressing this question would be to add total disclosures into the regression formula as a control variable. However, the high correlations between frequency of negative disclosures and total disclosures (>.90) would introduce multicollinearity into the equation. Instead, regression analyses were conducted for the frequency of positive and neutral disclosures late in therapy, controlling for early levels of disclosure of each valence, in an effort to provide evidence that this effect was unique to negative disclosures. Early positive self-disclosures showed a trend to predicting the frequency of  PSP in Group Psychotherapy/page 61  positive disclosures late in therapy (R =.047, F 2  (Ii64)  =3.16, p=.22, p=.080). However, P-PSP did  not contribute additional predictive power to this equation (R Change=.002, F 2  =0.06, p=.946).  ( 2 62)  Early neutral disclosures failed to predict the frequency of neutral disclosures late in therapy (R =.027, F =1.79, p= 185). Similarly, P-PSP did not predict the frequency of late neutral 2  (164)  disclosures (R Change=.023, F 2  =0.76, p=.473). Therefore, the impact of the P-PSP variables  ( 2 62)  appears to be unique to negative self-disclosures. They do not impact changes in the frequency of other forms of disclosure during the course of therapy. A final set of analyses was conducted to determine whether the P-PSP variables were unique predictors of negative self-disclosures late in therapy, after controlling for trait perfectionism and other related variables (Table 15). Social anxiety predicted a slight increase in negative selfdisclosure over the course of therapy. Although nondisplay of imperfections no longer contributed significant unique variance, nondisclosure of imperfections continued to be a strong predictor of a relative decrease in the frequency of negative self-disclosures from early to late sessions in therapy. The last self-disclosure descriptor to be analyzed was the focus of disclosure. The sum of in-group and combined in-group/out-group disclosures was used as the dependent variable. Together, the P-PSP variables did not contribute significant variance to the prediction of this form of disclosure (R =.024, F 2  ( 2 63)  =l .80, p=.173). However, at the univariate level,  nondisclosure of imperfections showed a slight trend to predicting morefrequentin-group disclosures (P=.27, t=l .89, p=.064). Although caution is required in conducting further analyses of these results due to the non-significant scores for the P-PSP block, controls were added to determine whether the impact of nondisclosure of imperfections was unique (Table 16). When trait perfectionism was evaluated, self-oriented perfectionism was a significant predictor of greater frequency of in-group disclosures. After controlling for these variables, nondisclosure of imperfections no longer contributed additional unique variance. This suggests that shared components of self-oriented perfectionism and nondisclosure of imperfections may be important in determining the frequency of in-group disclosures. Next, in-group disclosures late in therapy were considered, controlling for this type of disclosure early in therapy (R =.l 16, F 2  (164)  =8.43,  PSP in Group Psychotherapy/page  62  p=.005). Perfectionistic self-presentation did not significantly predict change in in-group disclosures over the course of therapy (R Change =.018, F =0.65, p=.528). 2  (262)  The last set of variables from the videotape coding system was observer ratings of mood. Hypotheses suggested that P-PSP would be associated with less expression of negative affect, although such individuals would actually be experiencing more sadness than other group members. Observer ratings of sadness were not significantly predicted by P-PSP (R =.044, 2  F  =1.46, p=.239). As with the self-disclosure variables, differences in sadness ratings were  ( 2 63)  investigated between sessions early and late in therapy. Ratings of sadness in late sessions were used as the dependent variable, while early ratings were controlled for by entering them on the first step of the analysis, followed by P-PSP on the second step. Early ratings of sadness did not significantly predict ratings late in therapy (R =.011, F =0.71, p=.402). In addition, P-PSP did 2  (Ij64)  not predict observer ratings of sadness late in therapy (R Change=.009, F 2  =0.29, p=.751).  ( 2 62)  This suggests that individuals with concerns about perfectionistic self-presentation were not perceived as being sadder than other group members at any point during treatment. Hypothesis 2: Therapy outcome. Once again, multiple dependent variables were employed to evaluate this hypothesis including: group member, close other and therapist ratings of benefit, post-treatment levels of perfectionism and other psychological symptoms, close others' ratings of perfectionism posttreatment, and changes in physiological reactivity. Each of these will be considered in turn. Ratings of satisfaction and benefit. Group members judged their own treatment success using several ratings at different points during therapy. Table 17 shows correlations between PPSP and members' post-treatment ratings of satisfaction with the group, how much they benefited from the treatment, and how successful they were in achieving their personal goals both at the midpoint of therapy and at termination. The majority of these correlations were not significant. However, there was a trend for nondisplay of imperfection to predict higher levels of satisfaction with the treatment program. A regression analysis demonstrated that nondisplay was not a unique predictor of treatment satisfaction when entered along with nondisclosure of imperfections (R =.055, F =1.74, p=. 185). 2  (260)  PSP in Group Psychotherapy/page 63  People close to the group members also rated their overall benefit from the treatment program. These ratings did not correlate strongly with P-PSP (Table 17). However, statistically significant correlations did exist between both P-PSP variables and therapists' ratings of global benefit. Counter to predictions, higher scores on both nondisplay and nondisclosure of imperfections were associated with higher ratings of benefit by the therapists (Table 17). This link was further investigated with a regression analysis to see which P-PSP variable was most relevant. As a unit, the P-PSP variables were significant predictors of therapist ratings of benefit (R =.124, F =4.39, p=.016). However, only nondisclosure of imperfections continued to show 2  (262)  a trend to contributing unique variance to this benefit rating (P=.25, t=1.83, p=.071). This suggests that therapists viewed individuals with concerns about disclosing their imperfections as having benefited morefromthe therapy program than other participants. Additional controls were added in a second regression equation (Table 18). Related variables were entered on the first step, followed by trait perfectionism on the second step and P-PSP on the third. A trend existed for higher levels of patient depression to predict greater therapist perceptions of benefit from the treatment. Having controlled for these variables, neither P-PSP variable contributed additional unique variance. Thus, therapists viewed those patients who were more depressed at intake as having gained the mostfromthe group therapy experience. Self-report ofperfectionism and psychological symptoms. First, correlations were calculated between P-PSP and each of the outcome measures post-treatment (Table 19). Nondisplay of imperfection was associated with higher levels of other-oriented and socially prescribed trait perfectionism following treatment. It also demonstrated trends to being linked to higher levels of all psychological symptoms including anxiety, depression and interpersonal problems. Nondisclosure of imperfections was significantly correlated with higher self-oriented and socially prescribed perfectionism post-treatment, as well as higher levels of depression. Thus, both nondisplay and nondisclosure of imperfections were associated with more elevated psychological symptomatology following treatment. These correlations do not answer the question of whether P-PSP predicted less benefit from the treatment. If P-PSP is linked to greater psychopathology prior to starting treatment,  PSP in Group Psychotherapy/page 64  individuals scoring high on these dimensions still may have made comparable gains to other group members, yet continue to show more elevated scores on measures of distress. Therefore, partial correlations were calculated for each of the outcome variables, controlling for pretreatment levels of each measure (Table 20). Only nondisclosure of imperfection continued to show a trend to being associated with higher post-treatment levels of depression. All other correlations were not statistically significant. This suggests that although individuals scoring highly on P-PSP dimensions continue to show greater psychopathology post-treatment, in general, they do make comparable gains during treatment when compared to other group members. The only exception to this was the link between nondisclosure of imperfections and depression. An initial regression analysis confirmed that nondisclosure of imperfections continued to show a trend in predicting variance in post-treatment depression, even after controlling for pretreatment depression levels (R =.192, F =13.56, p<.001) and nondisplay of imperfections 2  ( ] 57)  (Block 2: R Change=.047, F =1.70, p=.192; perfectionistic nondisclosure: p=.25, t=1.81, 2  (255)  p=.075). A second hierarchical regression analysis controlled for pre-treatment depression on the first step, related variables (other than depression) on the second step, and trait perfectionism on the third (Table 21). Although as a block, P-PSP was not statistically significant, at a univariate level, nondisclosure of imperfection continued to be a unique predictor of post-treatment depression. Once again, the lack of significance of the P-PSP block requires cautious interpretation of the univariate results. Higher scores on nondisclosure of imperfection appear to predict higher levels of depression even after a course of group psychotherapy. Close others' ratings ofperfectionism. Correlations between group members' self-ratings of P-PSP and close others' ratings of pre-treatment perfectionism are presented in Table 22. Ratings of P-PSP by members and their friends or family, were not highly correlated. Members' own ratings of nondisplay of imperfections showed a trend to being associated with ratings of this variable by close others, but was more strongly linked to others' ratings of nondisclosure of imperfection. Patients' ratings of nondisclosure of imperfections were not correlated with any of the perfectionism ratings completed by individuals close to them. This suggests that members in  PSP in Group Psychotherapy/page 65  these groups may have been somewhat successful at hiding the extent of their perfectionism from others around them. Table 22 also lists correlations between P-PSP and close others' post-treatment ratings of perfectionism. Close others' ratings of nondisclosure of imperfection following treatment were significantly associated with both nondisplay and nondisclosure of imperfection as rated by the patient prior to starting therapy. This suggests that members who scored highly on P-PSP at pretreatment, continued to struggle with disclosing mistakes and problems at the end of therapy, more so than other group members. Partial correlations were calculated to explore posttreatment perfectionism ratings, controlling for pre-treatment ratings by close others. However, none of these were statistically significant. Therefore, patient ratings of P-PSP did not predict the degree of change in other-rated perfectionism from pre- to post-treatment. Autonomic reactivity. The last measure of outcome was change in autonomic reactivity in response to a stressful task from pre- to post-treatment. Table 23 presents the correlations between P-PSP and heart rate and SCL reactions pre- and post-therapy. Both nondisplay and nondisclosure of imperfection were associated at a trend level with a larger magnitude of heart rate increase during the pre-treatment stress test . This association was no longer apparent 9  following the group therapy treatment. This suggests that perhaps P-PSP is linked to elevated stress and task involvement when confronted with an achievement task. As with the self-report data, partial correlations were calculated between P-PSP and post-treatment indicators of arousal, while controlling for pre-treatment arousal (Table 24). The only significant correlation was between nondisclosure of imperfections and SCL elevation following treatment. Higher levels of perfectionistic nondisclosure predicted greater SCL reactivity, even after controlling for pretreatment arousal. This implies that individuals scoring highly on nondisclosure of imperfections did not show a comparable decrease in stress reactivity, compared to other group members. Hierarchical regression analyses then were conducted to explore the unique contribution of perfectionistic nondisclosure to the prediction of post-treatment SCL reactivity. SCL elevation assessed prior to treatment showed a trend in predicting post-treatment SCL elevation (R =.063, 2  9  Neither P-PSP variable was correlated with accuracy of performance on the " IQ" computer task.  PSP in Group Psychotherapy/page 66  F  ( 1 5 2  =3.52, p=.066). P-PSP was entered as a block on the second step and also showed a trend to  adding predictive value (R Change=.083, F 2  (25 0  =2.43, p=.098). In particular, nondisclosure of  imperfection was linked to greater SCL elevations post-treatment (P=.34, t=2.13, p=.038). That is, higher scores on perfectionistic nondisclosure predicted greater stress reactivity as measured by SCL, following group therapy treatment. Nondisclosure of imperfections continued to be a unique predictor of SCL stress elevations even after controlling for related variables and trait perfectionism (Table 25).  Gender Differences. Although not a major focus of this paper, data were reanalyzed to test for gender differences in the impact of P-PSP. Gender was entered along with the P-PSP variables as a block to test for main effects. On a subsequent step, interaction terms between gender and each P-PSP variable were entered. Very few of these analyses resulted in significant findings, either of gender main effects or interactions. Although interesting, main effects of gender are not directly relevant to the hypotheses in this paper. Those that were identified are listed in Table 26, along with separate means and standard deviations for men and women. To summarize, women were more likely to miss sessions without providing an excuse. Main effects of gender were consistently identified on ratings of satisfaction with the treatment program, perceived benefit, and goal achievement ratings at the midpoint of therapy (at trend levels) and post-treatment. In every case, female group members rated the program and their therapy gains more positively than male members. Finally, there was a trend for men to show greater post-treatment SCL elevations in response to the stress test, and poorer return to baseline relative to pre-treatment, compared to female group members. In other words, the SCL stress pattern for women improved over the course of therapy, whereas for male group members, the post-treatment SCL pattern appeared to be less healthy, marked by stronger elevations and slower return to baseline levels following the test. Of greater importance to this study were interactions noted between perfectionism and gender in predicting therapy process and outcome variables of interest. Those interactions that were statistically significant are described in more detail below.  PSP in Group Psychotherapy/page 67  Self-disclosure. The interaction between gender and nondisclosure of imperfections demonstrated a trend to predicting thefrequencyof disclosures late in the course of therapy, while controlling for disclosures early in therapy (Table 27). Among women, nondisclosure of imperfection was a statistically significant negative predictor of disclosures late in therapy (First Block - early disclosures: R =.060, F =2.57, p=.l 17; Second block - P-PSP: R Change 2  2  (140)  =.163, F  ( 2 3 8  =3.98, p=.027; perfectionistic nondisclosure: (3=-.48, t=-2.80, p=.008; Figure 1).  That is, higher scores on nondisclosure of imperfections led to fewer disclosures late in the course of therapy. Male group members did not show this effect (First Block - early disclosures: R =.270, F =8.13, p=.009; Second block - P-PSP: R Change =.067, F =1.02, p=.379). 2  2  (122)  (220)  A similar pattern was noted for the frequency of negative self-disclosures late in therapy, controlling for early negative disclosures (Table 28). Among women, nondisclosure of imperfections significantly predicted relatively fewer negative self-disclosures late in therapy (First Block - early negative disclosures: R =.049, F =2.08, p=.157; Second block - P-PSP: 2  (140)  R Change =191, F 2  ( 2 3 8  =4.79, p=.014; perfectionistic nondisclosure: P=-.51, t=-2.97, p=.005;  Figure 2). Men did not show this effect (First Block - early negative disclosures: R =.431, 2  F =16.64, p<.001; Second block - P-PSP: R Change =.092, F =1.93, p=.171). 2  (122)  (220)  Tests of late levels of positive self-disclosure showed a somewhat different pattern (Table 29). After controlling for early levels of positive self-disclosure (R =.090, F =2.19, p=.153), 2  (122)  P-PSP predicted a relative increase in the use of positive self-disclosures late in therapy, by male group members (R Change =.317, F =5.34, p=.014; perfectionistic nondisclosure: P=.57, 2  (220)  t=2.46, p=.023; Figure 3). That is, concerns about perfectionistic nondisclosure predicted an increase in men's production of positive self-disclosures late in the course of therapy. Female group members did not show a significant association between perfectionistic nondisclosure and positive self-disclosures late in therapy (First Block - early positive disclosures: R =.033, 2  F =1.35, p=.253; Second block - P-PSP: R Change =.075, F =1.59, p=.218). Thus, 2  (li40)  (238)  although both men and women show an effect of nondisclosure of imperfections on selfdisclosure frequency late in therapy, men react by using more positive self-disclosures, whereas women use fewer negative self-disclosures over time.  PSP in Group Psychotherapy/page 68  One more interaction between P-PSP and gender was noted for the self-disclosure data. A significant interaction was noted between nondisclosure of imperfections and in-group disclosures late in therapy, controlling for early in-group disclosures (Table 30). As with the positive disclosure data, men appeared to show a stronger effect for in-group disclosures. After controlling for early in-group disclosures (R =.192, F =5.23, p=.032), P-PSP showed a trend 2  (122)  to predicting additional unique variance in late in-group disclosure frequency, among male group members (R Change =.180, F =2.88, p=.080; perfectionistic nondisclosure: P=.35, t=1.47, 2  (220)  p=.159; Figure 4). In contrast, women did not show a significant impact of P-PSP on late ingroup disclosure frequency (First Block - early in-group disclosures: R =.068, F =2.91, 2  (140)  p=.096; Second block - P-PSP: R Change =.008, F =0.16, p=.852). 2  (238)  Close others' ratings of benefit. As a main effect, gender did not predict how individuals close to the group members would rate their benefit from therapy. However, the interaction between gender and nondisplay of imperfections was a significant predictor of these ratings (Table 31). Specifically, male group members who initially scored highly on nondisplay of imperfection were rated by individuals close to them as having benefited from the treatment program more than other participants (R =.516, F =5.34, p=.026; nondisplay: P=.98, t=2.95, 2  (210)  p=.015; Figure 5). This effect was not found for female group members (R =.018, F =0.26, 2  (229)  p=.771). It should be noted that these analyses were based on a very small subset of male group members who returned ratings by close others. Autonomic reactivity. No gender effects existed for measures of heart rate during or after the stress test. However, interactions were noted for skin conductance levels. In addition to the main effects described earlier, gender/P-PSP interactions also were significant predictors of posttreatment SCL return to baseline (Table 32). Among female group members, neither nondisplay nor nondisclosure of imperfections had a significant impact on the ability to recover from stress as indicated by SCL (First Block - pre-treatment SCL return to baseline: R =.035, F =1.20, 2  (133)  p=.282; Second block - P-PSP: R Change =.095, F 2  (231  =1.70, p=.200). However, for male  participants nondisplay of imperfections was associated with a better return to baseline posttreatment (First Block - pre-treatment SCL return to baseline: R =.047, F =0.78, p=.390; 2  (116)  PSP in Group Psychotherapy/page 69  Second block - P-PSP: R Change =.192, F =1.76, p=.208; nondisplay: p=-.59, t=-1.70, 2  (214)  p=. 112; Figure 6), whereas nondisclosure of imperfections predicted less return to baseline following the stressor delivered at post-treatment (P=.60, t=l .77, p=.099; Figure 7). Although the effect with nondisplay was not statistically significant when regressions were conducted independently for men and women, the interaction term of gender and nondisplay in the initial hierarchical regression analysis was significant. This discrepancy may be explained by the small number of male group members and poor power for analyses involving only this sub-sample. These interactions suggest that in terms of ability to recover from stressors physiologically, nondisplay of imperfections was associated with a better treatment outcome in male group members, whereas nondisclosure of imperfections was associated with a poorer outcome for males. Ratings ofperfectionism by close others. An interaction between gender and self-rated nondisplay of imperfections predicted others' post-treatment ratings of nondisclosure of imperfection, after controlling for others' ratings of this variable at pre-treatment (Table 33). Among female group members, no association was found between self-rated nondisplay of imperfection and others' ratings of nondisclosure of imperfections (First Block - pre-treatment other ratings of perfectionistic nondisclosure: R =.154, F 2  (131  =5.64, p=.024; Second block - P-  PSP: R Change =.082, F(2,29) 1.55, p—.229). However, among male group members, there was a 2  —  significant link between these variables (First Block - pre-treatment other ratings of perfectionistic nondisclosure: R =.686, F , =28.35, p<.001; Second block - P-PSP: R Change 2  2  (  13)  =.111, F , =3.01, p=.090). Specifically, initial self-report of elevated nondisplay of (2 1}  imperfection predicted lower ratings of nondisclosure of imperfection by someone close to the group member at post-treatment (p=-.50, t=-2.45, p=.032; Figure 8). Once again, it appears that higher levels of nondisplay of imperfection in males are associated with better treatment outcome, this time in the form of others' ratings of men's avoidance of disclosing imperfections. Testing the Mediational Model. In order to test the hypothesis that nondisclosure of imperfection has an adverse impact on therapy outcome due to its impact on psychotherapy process, potential mediational mechanisms  PSP in Group Psychotherapy/page 70  were identified. The relationship between perfectionism and poor outcome was most clearly demonstrated by group members' self-report of depression, and the level of SCL reactivity preand post-treatment. Therefore, these variables were used to operationalize therapy outcome. Two variables showed robust associations between nondisclosure of imperfection and therapy process: elevated post-session anxiety following sessions late in the course of therapy, and the frequency of not showing up to sessions without providing an excuse. In addition, among female group members, nondisclosure of imperfection predicted a relative decrease in negative disclosures, and total disclosures from early to late therapy sessions. Each process variable was tested as a potential mediator for each of the two outcome variables. For example, one series of analyses tested whether the association between nondisclosure of imperfection and posttreatment depression could be explained by perfectionistic nondisclosure's impact on postsession anxiety. Analyses were conducted according to guidelines suggested by Baron and Kenny (1986). First, a significant association is confirmed between nondisclosure of imperfection and the outcome variable, using a hierarchical regression analysis to control for pre-treatment levels of the outcome variable on the first step. Next, the association between the process variable and outcome variable is tested. For the mediational model to hold, this test should also be statistically significant. Finally, a hierarchical regression analysis is conducted with pretreatment levels of the outcome variable on the first step, the process variable on the second step, and nondisclosure of imperfection on the third step. If nondisclosure of imperfection no longer explains significant unique variance in the outcome variable, it can be assumed that the process variable mediates its effect. Baron and Kenny (1986) suggest a second method for confirming the significance of the indirect, or mediational, model, using Sobel's (1982) formula (Appendix C). This test will be conducted for any mediational models identified by the first methodology. In the analyses for post-treatment self-report of depression, after controlling for pretreatment depression (R =.192, F =13.56, p<.001), nondisclosure of imperfection showed a 2  (157)  trend to predicting higher levels of depression (R Change =.044, F 2  ( 1 5 6  =3.22, [3=22, p=.078).  Tests for each process variable then were conducted. Only one succeeded in significantly  PSP in Group Psychotherapy/page 71  diminishing the contribution of nondisclosure of imperfection. Table 34 shows the analysis for the model of nondisclosure of imperfection leading to greater post-session anxiety following sessions late in the course of therapy, which, in turn, leads to less decrease in depression posttreatment. After controlling for post-session anxiety, nondisclosure of imperfection no longer explained unique variance in depression outcome. The proportion of the variance in depression scores uniquely predicted by nondisclosure of imperfections decreased from 4.4% in the unmediated model to 2.3% in the mediated model. This mediational model is illustrated in Figure 9a. Although this suggests a mediational model, post-session anxiety was not a strong predictor of depression outcome. To fully qualify as a mediational relationship, this processoutcome link ought to have been stronger. The significance of the indirect model was tested using Sobel's (1982) equation. It was not statistically significant (t = 1.20, p>.10). Thus, the experience of post-session anxiety may partially explain the elevated post-treatment depression scores associated with nondisclosure of imperfections, but anxiety is not a strong mediator of this relationship. The analyses involving SCL elevation during the stress test showed clearer examples of mediation. Once again, the association between nondisclosure of imperfection and posttreatment SCL elevation was established, after controlling for pre-treatment SCL reactivity (First Block - pre-treatment SCL reactivity: R =.063, F  ( 1 5 2  =3.52, p=.066; Second block -  perfectionistic nondisclosure: R Change =.074, F  (15]  =4.41, P=.27, p=.041). Next, each process  2  2  variable was considered as a possible mediator, and two showed signs of playing this role. Postsession anxiety following sessions late in the course of therapy was significantly linked to less reduction in SCL reactivity post-treatment (Table 34). After controlling for post-session anxiety, nondisclosure of imperfection no longer was a unique predictor of SCL reactivity. The proportion of variance in post-treatment SCL elevation uniquely predicted by nondisclosure of imperfection diminished from 7.4% in the unmediated model to 3.4% when using post-session anxiety as a mediator. Figure 9b illustrates this mediational model. Once again, Sobel's (1982) equation was used to verify the significance of the indirect model. In this case, the mediational model was statistically significant using a one-tailed test because the direction of the association  PSP in Group Psychotherapy/page 72  was predicted a priori (t = 1.72, p<.05). Thus, post-session anxiety appears to be one mediator of perfectionism's impact on stress reactivity outcome. Among female group members, a second process variable, self-disclosure late in therapy controlling for disclosure in early sessions, also predicted post-treatment SCL reactivity. For women, the association between perfectionistic nondisclosure and post-treatment SCL reactivity was somewhat stronger than for the sample as a whole (First Block - pre-treatment SCL reactivity: R =.160, F =6.49, p=.016; Second block - perfectionistic nondisclosure: R 2  2  (134)  Change =131, F  =6.11, (3=36, p=.019). As with post-session anxiety, once the variance  ( ] 33)  attributed to early self-disclosure was partialled out on a previous step, nondisclosure of imperfection was no longer a significant predictor of post-treatment SCL reactivity (Table 34). The proportion of variance in SCL elevation uniquely predicted by nondisclosure of imperfections shrank from 12.2% in the unmediated model to 5.8% when using frequency of self-disclosures as a mediator. In terms of the significance of the indirect model (Figure 9c), Sobel's (1982) test indicated that frequency of self-disclosure met trend levels of significance as a mediator (t = 1.63, p<.10 for a one-tailed test). Although caution must be used in interpreting such trend level effects, the negative impact of nondisclosure of imperfection on stress reactivity seems to be mediated by two process variables: increased post-session anxiety following sessions late in the course of therapy, and a relative decline in self-disclosure from early to late sessions in therapy for female group members. DISCUSSION This study sought to explore the association between protective forms of perfectionistic selfpresentation and group psychotherapy process and outcome. Two primary hypotheses were explored. First, P-PSP was expected to be linked to greater stress during the course of therapy as reflected by self-report of anxiety, physiological signs of stress, behavioural resistance to therapy, and fewer self-disclosures, particularly of a negative valence. Secondly, P-PSP was anticipated to lead to less positive therapy outcome, measured by self-report, reports of family, friends, and therapists, as well as indices of physiological reactivity to stress. Finally, these elements were expected to combine to create a mediational model in which P-PSP led to  PSP in Group Psychotherapy/page 73  problems with group process, which, in turn, led to less positive therapy outcome. Support was found for each of these hypotheses. Evidence existed that P-PSP was linked to greater stress during therapy as indicated by several diverse measures. P-PSP also was associated with less of a decrease in depression and SCL stress reactivity following treatment. Most relevant were the results indicating that both self-report of post-session anxiety and decreases in the frequency of self-disclosures over the course of treatment appear to act as mediators in the association between P-PSP and therapy outcome. Impact on Therapy Process  In considering therapy process, some of the most robust results came from analyses of PPSP and self-report of anxiety. As expected, concerns about disclosing imperfections predicted greater increases in anxiety over the course of therapy. This effect remained significant even after controlling for trait dimensions of perfectionism and related variables such as depression, social anxiety, and impression management. Thus, nondisclosure of imperfection was a unique predictor of post-session anxiety late in therapy. These results build on those of Habke and Hewitt (2000) who demonstrated that P-PSP was associated with elevated negative mood before and following an assessment interview. Similar to the present study, Habke and Hewitt noted that nondisclosure of imperfection predicted post-assessment negative mood, after controlling for initial mood ratings. Even more striking are the parallels between the time-line noted in this study and that described by Blatt et al. (1998). In analyses of the NIMH TDCRP study, they noted that perfectionism showed a negative impact on therapy outcome only later in the course of therapy, after approximately the eighth session. Results concerning anxiety ratings in this study provide further evidence that perfectionism's impact may occur late in the course of therapy. They also suggest that the experience of anxiety during therapy may be one mechanism by which perfectionism leads to more negative therapy outcome, although this issue is more directly addressed by the mediational analyses. Differences in the impact of the two P-PSP variables on anxiety also were noteworthy. Whereas nondisplay of imperfection was strongly linked with pre- and post-session anxiety, nondisclosure of imperfection was linked to post-session anxiety only. It seems that individuals  PSP in Group Psychotherapy/page 74  who are concerned about being seen making mistakes or looking foolish experience considerable anticipatory anxiety prior to group therapy sessions. In contrast, individuals with concerns about disclosing shortcomings experience greater anxiety after such sessions, when they may feel overwhelmed by the group's expectations of disclosure, or by the fact that they did disclose which is a new and frightening experience for them. This difference also was observed in Habke and Hewitt's (2000) study. They reported that nondisplay of imperfection was a unique predictor of negative mood prior to an assessment interview, although nondisclosure of imperfection was a better predictor of post-interview negative mood ratings. The P-PSP variables also were distinguished by the timing of greatest anxiety. Nondisplay showed stronger associations with anxiety during sessions early in the course of therapy, whereas perfectionistic nondisclosure predicted increased post-session anxiety late in the course of therapy. This contrast supports the theory that concerns about displaying imperfections may be most relevant in contexts with strangers or people with whom the individual is less familiar (Hewitt et al., 2000). Nondisclosure of imperfection tends to be most relevant in closer or more intimate relationships, in which demands exist for a greater level of openness and sharing. At the start of group psychotherapy, the group is a collection of strangers who gradually reveal more intimate details of their lives and learn to care for one another. It makes sense that initially, those with elevated nondisplay of imperfection would fear being seen as foolish by this unfamiliar group. However, as they get to know other members, these fears diminish and these individuals become more comfortable. Quite the opposite, individuals with disclosure concerns begin the group with less anxiety. In the beginning phases, discussion remains at a more superficial level, and members often discuss minor problems or issues they have already resolved to some extent. It is only as cohesion grows in the group that members start to reveal personal flaws and more grave concerns. As people start to open up in this way, an expectation may develop that other members will respond with a similar level of openness and self-revelation (Allen, 1974). At this point, nondisclosure of imperfection shows increasing links to anxiety, likely reflecting discomfort with this new phase in therapy.  PSP in Group Psychotherapy/page 75  To test the theory that changes in anxiety paralleled increases in engagement and decreases in avoidance within the groups, measures of perceived group engagement and group avoidance from the Group Climate Questionnaire (Mackenzie, 1983) were tested for differences in early and late sessions. As expected, group members did report increasing engagement and decreasing avoidance over time (MANOVA: F(2,60)=14.26, p<.001; univariate tests for engagement F(2,60)=26.61, p<.001 and avoidance F(2,60)=3.62, p=.021). Therefore, these changes in group climate may be partially responsible for the increasing anxiety of individuals with disclosure fears. Salivary Cortisol measurements were hoped to mirror the results of self-reported anxiety; however, few associations between perfectionism and Cortisol were noted. Several explanations exist for these null findings. First, Cortisol can be influenced by many factors such as exercise, eating, drinking, smoking, and brushing one's teeth (Hellhammer et al., 1987; Landon, 1982; Vining & McGinley, 1985). For most of the samples, participants were in the experimental or therapeutic environment for sufficient time to prevent them from engaging in any of these behaviours prior to sampling. However, for the anticipatory sample (prior to thefirstsession), participants could only be instructed not to do any of these activities for at least an hour prior to the session. Although not assessed formally, research assistants reported that while many participants took these instructions quite seriously, others arrived for the session rushing and out of breath or having eaten shortly beforehand. These activities may have contributed significant error to the anticipatory sample . For samples taken following therapy sessions, another 10  possible explanation is that participants may have felt anxious but not helpless. Cortisol tends to be released in settings in which the individual is not only stressed, but also feels out of control and unable to cope (Dienstbier, 1989 Frankenhaeuser, 1979). Perhaps the use of pre-group sessions helped to normalize the anxiety experienced by perfectionists and provided possible coping strategies for these stressful periods (e.g., listen for other group members to describe similar feelings, let the group know about your feelings). Armed with such information, group  Due to poor participant compliance, a baseline sample collected at home was discarded from the analyses.  PSP in Group Psychotherapy/page  76  members may have avoided more distressing feelings of helplessness and therefore, limited the release of Cortisol.  One association was found between P-PSP and salivary Cortisol. Nondisclosure o f imperfection predicted elevated Cortisol during the assessment session, even after controlling for trait perfectionism and related variables. In other words, higher levels of perfectionistic nondisclosure were associated with greater experience of physiological arousal during the assessment session. This is particularly striking since participants engaged in a fifteen-minute relaxation procedure prior to producing the assessment sample. The relaxation session may have been too brief to counteract the anxiety of the assessment. Other researchers have suggested that salivary Cortisol peaks 20-30 minutes following a mild laboratory stress (Kirschbaum & Hellhammer, 1989). Linden, Dadgar, and Earle (1994) noted that stressors involving the presence of other people tend to show shorter latency periods in Cortisol activation. Therefore, the fifteen-minute relaxation task plus time to teach the exercise and explain the Cortisol sampling method was thought to be adequate. Alternatively, nondisclosure of imperfection may have caused participants to benefit less from the relaxation procedure. However, this rationale is not easy to explain since no disclosure was required during „the relaxation protocol. One might have expected individuals scoring high on nondisplay of imperfection to have greater difficulty with the muscle relaxation exercises due to fears of looking foolish, yet results did not show this. In any case, perfectionistic nondisclosure did predict elevated Cortisol during the assessment. Given the previousfindingswith self-reported anxiety, it may seem strange that perfectionistic nondisclosure had an effect so early in the course of the therapeutic relationship. However, an assessment session includes clear expectations of disclosure, particularly of problem areas in the individual's life. Unlike early group sessions in which members are permitted to slowly build cohesion and trust, assessments require fairly immediate discussion of intimate details and difficulties. Thus, it is perhaps not surprising that nondisclosure o f imperfections is associated with strong physiological and emotional arousal during psychological assessment. An obvious and essential ingredient for positive psychotherapy outcome is attendance at therapy sessions (Conte, Plutchik, Picard & Karasu, 1991; Luborsky et al., 1971). Several  PSP in Group Psychotherapy/page 11  methods are available for avoiding the process of psychotherapy, from planning other events during the weekly session time and using these as an excuse, to arriving late for sessions to limit the time spent in the therapy environment, to simply not showing up. In this study, nondisclosure of imperfections predicted increased frequency of not showing up to group sessions, without providing an excuse. Perfectionistic nondisclosure continued to be a unique predictor of this behaviour, at trend levels of significance, even after controlling for trait perfectionism and other related variables. Thus, individuals scoring high on this perfectionism dimension appear to use the ultimate in avoidance techniques for missing sessions. By not providing an excuse, they need not contact the therapists, nor risk being persuaded to attend the therapy session. It is interesting that perfectionistic nondisclosure did not predict attendance overall. Therefore, this aspect of perfectionism specifies how group members may miss sessions, rather than whether they will have poor attendance. This strategy for missing sessions has several negative consequences. Fellow group members and group therapists may become frustrated by the lack of communication about absences. Groups often worry about absent members, and may become resentful about such individuals' disregard for the rest of the group (Yalom, 1995). By not discussing absences with the therapists, group members also miss the opportunity to explore any resistance issues or emotional reactions that may be occurring at that time. Group therapists working with perfectionistic clients should monitor these absences and be prepared to address them in the context of concerns about disclosure and increasing group intimacy. Despite periodic absences from the therapy groups, continued participation in the program was excellent. Only five of the 72 group members dropped out of the therapy program. This represents a dropout rate of seven percent, substantially lower than the twenty-one to thirty-three percent reported by some other research teams (Connelly & Piper, 1989; Falloon, 1981; McCallum, Piper & Joyce, 1992). The use of pre-group training sessions may have helped to reduce dropout (Piper and Perrault, 1989). Alternatively, individuals who volunteered to participate in a non-clinic based research program may have felt a greater obligation to remain in the treatment groups because of the research element. That is, although they may have felt  PSP in Group Psychotherapy/page 78  uncomfortable or dissatisfied with the groups, they may have remained in the program to provide feedback to the research team. Finally, the brief nature of the therapy offered in this program may have helped members to avoid premature termination. Another behavioural indicator of discomfort in group therapy is the frequency of selfdisclosure. Several aspects of self-disclosure were investigated in this study; however, the most striking findings were with the total frequency of self-disclosures. Somewhat surprisingly, nondisclosure of imperfections predicted higher levels of self-disclosure during a session early in the course of therapy. It would appear that individuals, for whom disclosure is an issue, began therapy with a resolve to work on that problem area. Assessment feedback and pre-group training emphasized the importance of disclosure and participation in the groups, and these interventions may have had some effect. These results are consistent with Habke and Hewitt's (2000) study in which perfectionistic nondisclosure did not predict differences in mistakes admitted to during an assessment interview. They found that participants who scored highly on nondisclosure of imperfection talked as openly as other participants about their problems, although they experienced considerable anxiety and worry over doing so. In addition to more frequent disclosures early in therapy, nondisclosure of imperfections also predicted a greater number of self-disclosures involving in-group references. This type of self-disclosure is believed to be more therapeutic, allows greater interpersonal work to be done, and was actively promoted during pre-group training sessions. Thus, this finding provides further evidence that individuals with disclosure concerns invested considerable effort in being "good group members". Although these initially high rates of self-disclosure of personal information may reflect a certain degree of openness and willingness to address important therapeutic issues, several authors have cautioned that moderate levels of disclosure produce better interpersonal consequences (Allen, 1974; Cozby, 1973; Kelly, 2000a, 2000b). Excessively high levels of selfdisclosure can strip the speaker of their sense of individuality and privacy (Cozby, 1973), as well as lead to discomfort among the listeners due to pressure to reciprocate in kind (Allen, 1974). In addition, disclosures of objectionable information can cause therapists, and presumably other group members, to react negatively to the client (Kelly, 2000a). In turn, clients may internalize  PSP in Group Psychotherapy/page 79  these reactions, leading to more negative self-esteem (Kelly, 2000a). Thus, it is unclear whether the association between nondisclosure of imperfections and elevated frequency of early selfdisclosure is a positive indicator of their engagement in therapy or a threat to their ability to connect with the group. It may be that individuals who score highly on this perfectionism dimension have little practice or skill in determining an appropriate level of self-disclosure. When they decide to try revealing information about themselves, they may do so too rapidly or too early on in a relationship. These errors may lead to negative interpersonal consequences like rejection or lack of support (Allen, 1974), reactions which would confirm their beliefs about the risk of disclosure. This theory is consistent with Kelly's (2000a) suggestion that in order to be successful in therapy, clients must be capable of making good decisions about whether, when, and how to disclose negative personal information. Despite this energetic start to therapy, for female group members, perfectionistic nondisclosure also was a unique predictor of lower levels of disclosure late in therapy, after controlling for early levels of disclosure. In particular, perfectionistic nondisclosure was associated with fewer negative self-disclosures over the course of therapy. Thus, women with concerns about disclosing their imperfections showed a general decline in self-disclosure between early and late therapy sessions, and this decline was especially pronounced for disclosures with negative content. This effect is opposite to what one would normally expect in group therapy. As the group grows in trust and openness, self-disclosure ought to increase (MacKenzie, 1990). Information that relates to distress or negative self-perceptions should to be easier to share once the group has developed some sense of cohesion. This lack of increase in self-disclosure over the course of therapy provides further support for the notion that nondisclosure of imperfections is associated with increasing discomfort with therapy process over time. Perhaps, as these women come to know and care about fellow group members, they become more concerned about how the group will react to perceived flaws or weaknesses. This rationale fits with the theory that nondisclosure of imperfections is most relevant in close, intimate relationships (Hewitt et al., 2000).  PSP in Group Psychotherapy/page 80  In part, this effect was due to the low levels of self-disclosure in early sessions by women who scored low on nondisclosure of imperfection. These women cautiously entered into the group, disclosing less than other members initially, but allowing their self-disclosures to increase over time. At first, this pattern seems to contradict what is known about the nondisclosure of imperfections variable. These women stated that they do not have difficulties disclosing, yet they did not engage in self-disclosing behaviour at the start of therapy. In contrast, men and those women who scored high on nondisclosure of imperfections engaged in fairly constant levels of disclosure throughout the course of therapy. One explanation may be that women who do not have concerns about disclosure are more skilled at selecting an appropriate time and place to reveal their personal stories and feelings. Because they do not fear this type of behaviour, they may have more experience with self-disclosure in other contexts. In contrast, women who do fear disclosure of their imperfections, and men more generally (Shay, 1996), may have had little practice at determining an appropriate point in relationships for self-disclosure. Their high levels of self-disclosure at the start of therapy may reflect an effort to engage fully in the therapy process. However, if cohesion and trust in the group have not yet been established, such disclosures may fail to elicit needed support and feedback (Yalom, 1995). Among perfectionists, these fumbling attempts to meet perceived demands for openness in therapy are likely to lead to feelings of self-criticism, regret, and embarrassment when too much personal information is divulged too soon. In contrast, women with no concerns about disclosure may be able to wait until the therapy setting feels secure, prior to revealing and discussing their concerns. This pattern of engagement may be healthier and more productive. The association between nondisclosure of imperfections and decreases in self-disclosure over time is particularly important given the anxiety and distress experienced by individuals scoring high on this dimension of perfectionism. Several researchers have suggested that negative affect can be therapeutic, but only if accompanied by insight (Yalom, 1995), and cognitive learning (Bloch & Crouch, 1985), which typically are arrived at through exploration (McCallum et al., 1993) and disclosure. Without such exploration, the experience of negative affect can be detrimental to therapy outcome. Additionally, several studies have emphasized the  PSP in Group Psychotherapy/page 81  importance of self-disclosure, regardless of associated affect, for positive therapy outcome (O'Malley et al., 1983; Orlinsky et al., 1994). It is interesting to note that the more traditional pattern of increasing self-disclosure over time was predicted by nondisplay of imperfection. Nondisplay of imperfections also predicted a relative increase in negative self-disclosures from early to late therapy sessions. Again, this is consistent with findings on self-reported anxiety in which nondisplay was linked with early levels of anxiety. Individuals scoring highly on this dimension of P-PSP show greatest discomfort at the beginning of group therapy, when they are unfamiliar with the other group members. As the group develops and gets to know one another, individuals with concerns about being seen making a mistake appear to relax and grow more comfortable with the group context. This comfort is demonstrated by declining anxiety and increasing self-disclosure. Another possibility is that these individuals learn how to "behave" during group sessions, due to an understanding of group norms, and grow more confident about their ability not to be perceived as foolish. Male group members with concerns about disclosing imperfections showed a different pattern of self-disclosure over the course of therapy. Rather than a decrease in negative selfdisclosures, these men showed an increase in positive self-disclosures from early to late therapy sessions. This behaviour might also reflect discomfort with the therapy process. As cohesion developed in the group, these men chose to disclose their strengths and positive feelings. In addition, they showed a slight increase in the use of in-group disclosures. Thus, they may be expressing things like positive feelings toward others in the group, or satisfaction with how the group is functioning. These verbalizations may be attempts to be accepted by the group and to avoid discussion of weaknesses and shortcomings. It is interesting that men with protective selfpresentational concerns would employ an apparently acquisitive style of self-presentation. Analyses were conducted to test whether perfectionistic self-promotion, the acquisitive form of perfectionistic self-presentation, might better explain these findings. Although perfectionistic self-promotion was associated with more frequent positive disclosures by male group members throughout the course of therapy (R =.386, F =4.20, p=.019; self-promotion (3=.70, p=.006), it 2  (320)  PSP in Group Psychotherapy/page 82  did not explain shifts in positive self-disclosure over time. Thus, for men, nondisclosure of imperfections was a unique predictor of increasing positive self-disclosures over the course of therapy. Once again, the small number of male group members means that these findings must be interpreted with caution. However, these results suggest that women and men may act on their disclosure concerns by using different patterns of self-disclosure in psychotherapy. It will be important to study this in greater depth in future studies. The final distress-related hypothesis in this study was that despite the fact that P-PSP is linked to greater experience of negative affect, these emotions would not be visible to an observer. Both P-PSP variables are strongly correlated with depression ratings (nondisplay: r =.41, p<.001, nondisclose: r =.30, p=.010), yet they did not predict higher ratings of (71)  (71)  observed sadness during the therapy sessions. Although this does not provide support for the null hypothesis (the traditional significance test assumes that the null hypothesis is true and only tests whether the data deviate significantly from that assumption; Hammond, 1996), it is striking that such strong associations with self-report of negative mood cannot be identified by others. It also is noteworthy that self-reported depression did predict observer ratings of sadness (r  (66  =.32,  p=.010). One possible explanation for these findings was that nondisclosure of imperfections might moderate the impact of negative mood on observer ratings of sadness. However, a posthoc test of this interaction was not significant (First Block - main effects: R =.l 11, F 2  (263)  =3.94,  p=.024; Second block - interaction: R Change =.005, F _ =0.32, p= 575). Although 2  (1  62  perfectionistic nondisclosure does not act as a statistical moderator, it appears to be associated with some ability to mask negative affect. Unfortunately, this ability may diminish the likelihood that others will offer support and assistance (Clark et al., 1996). If other group members and therapists cannot discern these individuals' distress, little can be done to help them work through these feelings. In addition, studies have shown that inhibiting the expression of negative affect does not diminish the subjective experience of negative feelings (Gross & Levenson, 1997). Rather, such inhibition is likely to result in physiological stress in the form of sympathetic arousal. In turn, this physiological stress may impede cognitive processing and performance (Gross & Levenson, 1997). Clearly, disruptions to one's ability to concentrate and  PSP in Group Psychotherapy/page  83  integrate information explored during psychotherapy would be likely to have negative consequences for therapy outcome. In sum, several lines of evidence converge to suggest that P-PSP predicts increased stress and unease during group psychotherapy. Although nondisplay of imperfection is associated with anxiety in anticipation of therapy sessions and during the early phases of therapy, nondisclosure of imperfections is more strongly linked with post-session anxiety and for women, decreasing negative and overall self-disclosure as therapy progresses. Men with disclosure concerns show a different pattern of increasing positive and in-group disclosures from early to late sessions. This also may reflect growing discomfort and attempts to gain acceptance from the group. In addition, perfectionistic nondisclosure predicts an increased frequency of not showing up to sessions without providing an excuse, a behaviour that may be a sign of avoidance. Despite early efforts to fully participate in the process of therapy through disclosures with an in-group focus, nondisclosure of imperfections is linked to emotional discomfort and disengaging behaviour later in the course of therapy. Finally, this discomfort and distress does not appear to be easily noted by an observer, making it difficult for therapists and fellow group members to offer assistance. Although these findings suggest that perfectionists may find group psychotherapy to be challenging, their full importance can only be recognized in light of their impact on therapy outcome. Impact on Therapy Outcome  Before exploring the impact of perfectionistic self-presentation on therapy outcome, it is important to note that overall, the group therapy program was successful. The focus of this paper is not on the utility of the treatment. However, without the knowledge that the treatment program was helpful for most participants, it would be meaningless to comment on individual differences in treatment outcome. A MANOVA was conducted to test change on trait perfectionism, depression, anxiety, interpersonal problems, and physiological reactivity to stress. Overall, these dimensions showed significant improvement over the course of the treatment program (F  0040)  =6.08, p<.001). Several participants acted as wait-list controls prior to being  assigned to one of the treatment groups. After being assessed initially, they were required to wait  PSP in Group Psychotherapy/page  84  for a minimum of one treatment group duration prior to starting therapy. Before starting their group (and after the minimum three month wait), these participants completed the assessment measures a second time to evaluate change over the wait period. In comparison to participants who directly entered groups, wait-list controls showed a trend to having significantly less improvement on measures of trait perfectionism, depression, anxiety and interpersonal functioning (F =2.14, p<.08). Univariate tests indicated that the treatment group showed (554)  greater improvement than controls on measures of self-oriented perfectionism and depression, in particular. Thus, the treatment program appeared to be successful in reducing psychological symptoms of distress as well as perfectionistic motivations. Having established this fact, our attention now can turn to individual differences in treatment outcome. Several variables were used to explore the association between P-PSP and group psychotherapy outcome. These outcome variables spanned several modalities including selfreport, others' report, and physiological data. In general, P-PSP was correlated with higher posttreatment levels of psychological symptoms and trait perfectionism. Specifically, greater concern about disclosing imperfections was linked to greater struggles with depression, selforiented and socially prescribed perfectionism following group therapy. Nondisplay of imperfections was associated with continuing struggles with depression, anxiety, interpersonal problems, and interpersonal dimensions of trait perfectionism. Although these findings suggest individuals with P-PSP concerns are likely to show greater psychopathology than other group members after treatment, they do not specify whether such individuals benefit from treatment. Given the strong pre-treatment correlations between P-PSP and psychopathology, comparable degrees of change during therapy may have occurred. After controlling for pre-treatment levels of the outcome variables, only one self-report measure showed the expected negative association with nondisclosure of imperfections. Higher scores on nondisclosure of imperfections at pre-treatment predicted higher post-treatment depression scores, even after controlling for pre-treatment depression ratings. Thus, perfectionistic nondisclosure predicted less of a reduction in depression over the course of treatment. Even after controlling for trait perfectionism and related variables such as social  PSP in Group Psychotherapy/page 85  anxiety, this finding remained significant, indicating that perfectionistic nondisclosure is a unique predictor of post-treatment depression. This association matches that reported by Blatt et al. (1995) in their work on the treatment of depression. It is curious that the other self-report outcome measures did not demonstrate a similar negative impact of perfectionistic self-presentation. One possible explanation is that the treatment was too brief to create sizable changes in core beliefs such as trait perfectionism, or in interpersonal functioning. The latter has been shown to continue to shift following initial improvements in psychological symptoms such as anxiety and depression (Horowitz et al., 1988). With smaller changes among the sample as whole on these measures, it could have been difficult to discern individual differences in improvement. A post-hoc MANOVA was conducted to evaluate change in each of the self-report outcome variables. All showed marked decreases in the sample over the course of therapy, with levels of significance of p<.001. Therefore, sufficient change appears to have occurred, even on core problem areas. A second explanation for the limited outcome findings may be that by focusing the treatment program on trait perfectionism the range of possible scores on measures of perfectionistic self-presentation were restricted. If all group members scored highly on nondisplay and nondisclosure of imperfections, it would be impossible to accurately discern differences in outcome according to these variables. While it is true that this sample scored more highly on measures of protective perfectionistic self-presentation than other clinical samples (e.g., Hewitt et al., 2000), considerable variability remained. For example, the full range of possible scores on the subscale of nondisclosure of imperfections is 7 to 49. In this study, participants' scores ranged from 8 to 49, and the standard deviation for this measure (8.95) was comparable to past research. This range in scores was preserved because group members were selected on the basis of elevated trait perfectionism. Because trait and self-presentational forms of perfectionism are only moderately correlated, a full range of scores was still possible on the self-presentational measures. Therefore, restricted range of perfectionism scores is unlikely to fully explain the limited associations between protective perfectionistic self-presentation and treatment outcome.  PSP in Group Psychotherapy/page 86  In comparing data from this study to that conducted by Blatt et al. (1995), it is noteworthy that their research focused on the treatment of depression. Although they used a composite measure of outcome to evaluate the impact of perfectionism, this composite may have been somewhat weighted by multiple measures of depression. Indeed, they report using the Beck Depression Inventory and the Hamilton Rating Scale for Depression, in conjunction with three more general ratings of outcome. Thus, their finding of a negative association between perfectionism and therapy outcome, may also have been driven by a link between perfectionism and depression. This theory of a specific association between perfectionism and depression outcome might also explain that lack of perfectionism findings reported by Mussell et al. (2000) in predicting change in bulimia symptoms following treatment. Why, then, would perfectionism make it particularly difficult to lift feelings of depression during psychotherapy? Perhaps, this effect is caused by the very nature of perfectionism. Through the course of therapy, interpersonal skills and behaviours may be acquired that minimize problematic interpersonal issues and help individuals to better manage stress and some of their maladaptive beliefs. However, because of perfectionistic expectations, these individuals may demand a complete "cure" and a problemfree life. When this goal is not attained, they may continue to feel dissatisfied with themselves (and the program!), and helpless to change their lives as completely as they would like. Thus, abatement of depression may not accompany changes in other problem domains. This explanation is consistent with Goldstein and Shipman's (1961) work on the impact of expectations on therapy outcome. They noted that extremely high expectations reduce individuals' perceptions of symptom reduction, perhaps due to their disappointment with the reality of more modest levels of change. Hewitt and his colleagues (1998) commented that trait perfectionism is a vulnerability factor for chronic symptoms of depression. They cautioned that if perfectionism is not addressed and resolved through treatment, it will place individuals at risk for nonremission or recurrence of depression. However, they also acknowledged that the treatment of perfectionism is likely to require intensive, long-term therapy. Given the mixed results with self-reported outcome measures, it was striking that one physiological outcome measure showed a similar pattern of results to those for depression.  PSP in Group Psychotherapy/page 87  Initially, both P-PSP variables showed a connection to more elevated heart rate during the pretherapy stress test. This suggests that individuals scoring highly on these perfectionism dimensions showed greater physiological stress as a result of the mock "IQ test". This is consistent with findings by Flynn et al. (2001) in which self-oriented perfectionism was associated with elevated heart rate arousal during another type of academic exercise. More generally, it also supports Habke and Hewitt's (2000) observations that nondisclosure of imperfection predicted elevated heart rate during another type of stressor, discussion of past errors during an assessment interview. Thus, several studies have noted that perfectionism is associated with increased physiological arousal, in the form of elevated heart rate, during stressful events (Flynn et al., 2001; Habke & Hewitt, 2000). To evaluate the impact of therapy on diminishing the degree of arousal, the stress test was repeated again following the end of the treatment program. Hierarchical regression analyses, controlling for pre-treatment arousal, showed that nondisclosure of imperfections predicted greater skin conductance arousal at post-treatment. Thus, whereas other group members were less reactive to stress following treatment, individuals with strong disclosure concerns did not show a comparable moderation of SCL arousal. Even after controlling for trait perfectionism and related variables, perfectionistic nondisclosure continued to be a unique predictor of posttreatment SCL arousal. Two important questions arisefromthis finding. First, it is curious that perfectionistic nondisclosure was the P-PSP variable to show an effect on the stress test when self-disclosure was not required during the test. One might have expected nondisplay of imperfection to be more relevant given the potential for looking foolish or being observed making errors on the test questions. However, it is important to remember that this finding looks at change in arousal rather than arousal at a given point in time. In fact, at pre-treatment, nondisplay did show a slightly stronger association with heart rate elevation during the test, as one might predict. When change in arousal at post-treatment is explored, one must consider not only the stress situation itself, but also what has been gained over the course of therapy. That is, although an IQ test may not be the most stressful situation for individuals scoring highly on nondisclosure of  PSP in Group Psychotherapy/page  88  imperfection, the fact that their perfectionism concerns prevented them from self-disclosing during the course of therapy may have prevented themfromgaining as many stress reduction skills as other group members. Thus, the association between nondisclosure of imperfections and maintained SCL arousal likely reflects problems benefiting from the group therapy program. The second question that emerges is the contrast in findings between heart rate and skin conductance. Both of these indices reflect changes in autonomic nervous system arousal, although they are activated by different systems, noradrenergic and cholinergic respectively (Venables & Christie, 1980 ). Generally, these physiological markers are thought to act together, although occasionally differences have been found. Lang, Levin, Miller, and Kozak (1983) suggested that heart rate responded more to conditions requiring imminent action, whereas skin conductance arousal, on its own, may reflect response to emotional threat in the form of orienting and heightened attention. Thus, it may be that feelings of emotional threat during a stressful task were more difficult for individuals with disclosure concerns to overcome. Gross and Levenson (1997) also noted that inhibition of sad affect activated skin conductance arousal, but not heart rate. Therefore, individuals who scored highly on nondisclosure of imperfections may have attempted to hide their distress during the achievement taskfromthe observation of the experimenter. This suppression of affect also may explain the association between perfectionistic nondisclosure and elevated skin conductance reactivity. P-PSP did not have a significant impact on recovery from stress arousal. Dienstbier (1989) proposed that stress reactions are only pathological if they persist beyond the initial stressor. In this case, perfectionistic self-presentation was not associated with an enduring stress response. This is contrary to findings presented by Flynn et al. (2001) in response to an achievement stressor, but consistent with that presented by Habke and Hewitt (2000) in their study of stress during an assessment interview. Flynn et al.'s (2001) study focused on the impact of selforiented perfectionism, whereas the present study and that of Habke and Hewitt (2000) targeted perfectionistic self-presentation. Therefore, different aspects of perfectionism may show different stress profiles. Another possibility is that both this study and that of Habke and Hewitt (2000) used clinical populations, whereas Flynn et al. (2001) explored stress reactions in  PSP in Group Psychotherapy/page 89  university students. For clinical participants, other real life stressors (such as their clinical symptoms, participation in a psychological assessment or in a therapy program) may be more salient than a stress manipulation, causing them to obsess less about it following the stress task. It will be worthwhile to clarify this distinction in future studies, in order to better understand whether stress reactions among individuals with perfectionistic self-presentation can also have detrimental health consequences. Another source of information about therapy outcome came from ratings of perfectionism by individuals who were emotionally close to each group member (e.g., a family member or a close friend). Results from this source did not support the initial hypothesis. First, ratings by friends and family of perfectionistic self-presentation in the group members were not strongly correlated with members' own ratings, particularly for nondisclosure of imperfection. This suggests that group members may have been somewhat successful at masking their nondisclosure of imperfections. In contrast, nondisplay of imperfections was more likely to be recognized as a problem by close others, although they had some difficulty in identifying the motivations for the members' behaviour (i.e., distinguishing between nondisplay and nondisclosure of imperfections). Still, a regression analysis predicting others' ratings of nondisplay of imperfection showed stronger links to clients' own ratings of nondisplay than nondisclosure of imperfection (F =3.08, p=.068, nondisplay P=.36, p=.022; perfectionistic nondisclosure P=(2 58)  .23, p=.145). In contrast, others' ratings of nondisclosure of imperfection were not associated with either P-PSP self-rating (F  (258  =2.28, p=.120). Initially, this result may call into question the  validity of the nondisclosure of imperfections subscale. However, this finding stands in stark contrast to past studies, which have shown strong associations between self- and other-ratings of nondisclosure of imperfection (Hewitt et al., 2000, Study 3). Both college students' peers and therapists during an assessment interview were accurately able to discriminate relevant aspects of perfectionistic self-presentation. This study is different from those reported by Hewitt et al. (2000) in two key ways. First, this study involved a clinical sample being evaluated by nonprofessionals. In Hewitt et al.'s (2000) work, therapists evaluated perfectionistic selfpresentation in the clinical sample. It may be that in conjunction with clinical problems, it is  PSP in Group Psychotherapy/page 90  more difficult to discern concerns about disclosing imperfections without training in psychology, or knowledge of the level of disclosure expected and obtained in a therapy context. Another explanation is that participants in this study were selected for having high levels of perfectionism. Since all participants in the study suffered from at least one subtype of perfectionism, it may have been more difficult for other people to identify which aspects were most relevant for each individual. Finally, individuals with extremely high levels of nondisclosure of imperfections may be somewhat successful at preventing others from knowing the extent to which they keep information to themselves. It is fascinating to note that postsession ratings of perfectionism by group members and their friends or family are much more highly correlated (for nondisplay: r =.31, p=.030; perfectionistic nondisclosure: r (48  (4g  =.55,  p<.001). Perhaps, over the course of the therapy program, group members revealed more of their struggles to friends and family, which allowed them to gain a better appreciation for group members' levels of perfectionistic self-presentation. Although friends and family continued to view individuals with P-PSP as less disclosing of imperfections at the end of therapy, in general, neither P-PSP variable predicted change in others' ratings of perfectionism over the course of treatment. This is consistent with the results noted for group members' self-report in which P-PSP did not adversely impact improvements in perfectionism scores. When gender differences were explored, one association was noted for male participants. Male group members who initially scored highly on nondisplay of imperfections were rated by those close to them as more improved on the dimension of nondisclosure of imperfections at the end of therapy. While thisfindingmay initially appear to be an anomaly, it is supported by several other pieces of data collected on ratings of global benefit. The final source of treatment outcome information was global ratings of improvement by group members, therapists, and friends or family members. These results proved to be quite mixed. Nondisplay of imperfection was associated with moderately higher ratings of treatment satisfaction by group members. For male group members, ratings by close others' also support thesefindings.Friends or family members rated men scoring more highly on nondisplay of  PSP in Group Psychotherapy/page 91  imperfection as having benefited more from the treatment program. As noted earlier, nondisplay of imperfection also was associated with greater improvements in men's willingness to disclosure imperfections, as rated by close others. In addition, for male group members, nondisplay predicted a better return, to baseline for skin conductance reactivity following a stress test administered post-treatment. Diestbier (1989) reminds us that this ability to recover from stress is the most important factor for physical health. These positive implications of nondisplay of imperfection were not anticipated and will need to be confirmed by future studies. When stages of group development are considered, one possible explanation for this finding can be construed. Initially, a group of strangers has to tentatively form a group through finding commonalities and shared experiences (Yalom, 1995). Later, group members vie for power and come to know one another as distinct individuals (MacKenzie, 1990). Only as these processes occur do group cohesion and trust emerge allowing in-depth therapeutic work. If we consider the essential therapeutic work of individuals concerned with displaying imperfections, just attending the group is a huge step towards allowing others to see themselves as less than perfect. The first stage of finding a common ground, speaking and interacting in the group environment would be quite threatening for them. This assertion is supported by the high levels of anxiety reported by individuals scoring highly on nondisplay of imperfections prior to and following early therapy sessions. Sexton et al. (1996) and Saltzman et al. (1976) suggest that early anxiety levels can actually be suggestive of more successful therapy due to better cohesion and reduced attrition. Thus, nondisplay concerns are addressed from the very start of the therapy program. In contrast, concerns about disclosure only become apparent as the group reaches later stages in development when trust has begun to form. Thus, a group member who wishes to work on problems with nondisclosure of imperfections may only have started their therapeutic work quite late into a twelve-session program. For this reason, individuals who scored highly on nondisplay of imperfections may have appeared to progress more quickly and further than other group members. It would be useful to evaluate the impact of different numbers of therapy sessions on these contrasting aspects of perfectionistic self-presentation.  PSP in Group Psychotherapy/page 92  It is intriguing that the association between nondisplay of imperfection and greater therapy benefit is more pronounced for male group members. Few gender differences have been previously noted in the perfectionism literature (Hewitt et al., 2000). This finding also contrasts with consistent evidence that female group members reported greater satisfaction and benefit from the therapy program. These findings suggest that males with concerns about displaying their imperfections do particularly well in this type of group therapy program. Following from the premise proposed above, for males with this form of perfectionistic self-presentation, entering into therapy may be particularly threatening, and therefore therapeutic. Women tend to be more willing to seek psychological treatment (Good, Dell & Mintz, 1989), perhaps because therapy involves behaviours and attitudes that are counter to the traditional male role, including expression of emotion and feelings of intimacy (Good et al., 1989; Shay, 1996). Thus, men who joined groups in this program immediately displayed 'imperfection' by acting counter to society's expectations of the male role. They also inserted themselves into an environment that tended to be female dominated. Perhaps because this experience was so challenging, these men were able to quickly address key issues about displaying flaws and shortcomings, and make important therapeutic gains. Another explanation is that in female dominated groups, male participants may have felt safer because of a perception that women are less powerful or threatening. In fact, when asked about their reactions to a group composition of two men and the rest women, both men in one group replied that they had hoped to be the only male group member because that would have been more comfortable for them. Indirect empirical support for this explanation comes from work on therapist gender in individual psychotherapy, which has indicated that male clients who express a preference usually prefer female therapists (Pikus & Heavey, 1996). A sense of safety in female dominated groups may have allowed men with concerns about displaying imperfections to engage in more therapeutic work on these issues. Caution is needed in interpreting these findings given the small number of male group members in the study. Because of the unusual nature of this finding, future studies will be needed to confirm and explore its meaning.  PSP in Group Psychotherapy/page 93  Therapists' ratings of overall benefit reflected a somewhat different perspective. Therapists viewed both forms of P-PSP as being predictive of greater benefitfromthe group therapy program. After controlling for trait perfectionism and related variables, it became apparent that associations between P-PSP and depression seemed to explain the effect. That is, therapists rated group members who were most depressed at pre-treatment, as having progressed the most during therapy. In fact, additional analyses indicated that depression was not a strong predictor of selfreported changes in anxiety, interpersonal functioning, nor trait perfectionism. Perhaps these group members showed more easily identifiable improvements because of the striking nature of their initial depression. To test whether depression was observable early in therapy, correlations were calculated between depression scores and coders ratings of sadness during an early session. These two variables were positively correlated (r =.38, p<.001). In contrast, initial ratings of (68  depression were not associated with observations of sadness in later sessions (r =.16, p=.199). (66)  Thus, even our coders noted a shift in the affect of depressed group members. Therapists may have based their evaluations of benefit on such easily observable changes. To summarize findings on therapy outcome, only ratings of depression and skin conductance arousal in response to a stress test confirmed the hypothesis that nondisclosure of imperfections would be detrimental to therapy outcome. Other measures tended to show no impact of perfectionistic nondisclosure on outcome. Perfectionistic nondisclosure may have a particularly negative impact on people's ability to recover from feelings of depression because of persisting expectations of perfection which lead to disappointment, or because these symptoms require more intimate disclosure of particular problems, issues, and beliefs. Similarly, while skills may be learned to manage stress and function better interpersonally, remaining concerns may be evidenced in these individuals' reflexive physiological arousal in response to a stressful test. In contrast, among male group members, nondisplay of imperfection actually appeared to predict greater benefit from the program as judged by the members themselves, theirfriendsand family, as well as being reflected in better physiological recovery from stress. Although this finding will require confirmation by future research, it may reflect the challenging but therapeutic nature of group therapy for these individuals.  PSP in Group Psychotherapy/page 94  The Mediational Model  Having established that nondisclosure of imperfections was associated with greater discomfort with therapy process, and that it showed some negative impact on therapy outcome, a mediational model was proposed to link these components. Using Baron and Kenny's (1986) approach to identifying possible mediators, three possible pathways were identified. First, the association between nondisclosure of imperfections and elevated post-treatment depression was mediated by post-session anxiety late in the course of therapy. Similarly, the association between perfectionistic nondisclosure and post-treatment skin conductance reactivity also appeared to be mediated by post-session anxiety late in the course of therapy. This was the only mediated model that was statistically significant using Sobel's (1982) formula. For female group members, the perfectionistic nondisclosure - skin conductance reactivity link also was mediated by a relative decline in self-disclosure over the course of therapy. The strength of this mediational model met trend levels of significance. It was interesting that overall levels of selfdisclosure mediated this relationship whereas the frequency of negative self-disclosures did not. The combination of declining negative and positive self-disclosures must have a greater impact on treatment outcome. Together, these behaviours would minimize opportunities to discuss problems and difficulties, as well as limiting chances for boosting self-esteem through the discussion of positive affect and successes. In each of these three models, the mediator explained approximatelyfiftypercent of the unique contribution of nondisclosure of imperfections in predicting negative therapy outcome. Therefore, they were not ideal mediators, and they only partially explain why perfectionism may lead to more negative treatment outcome. This is similar to the mediational model proposed by Zuroff et al. (2000). In their study, change in patients' contribution to the therapeutic alliance acted as a mediator of the association between perfectionism and a general outcome factor, decreasing the unique prediction of perfectionism from 9% to 4%. Thus, there is evidence indicating that three process variables may be mechanisms by which perfectionism interferes with therapy outcome: anxiety, and lack of increase in self-disclosure and therapy alliance over the course of therapy. All of these factors seem to be most relevant late in the course of therapy,  PSP in Group Psychotherapy/page 95  corresponding to Blatt et al.'s (1998) observation that perfectionism interferes with outcome after approximately the eighth session. The factors identified in this study, anxiety and selfdisclosure, also may mediate the impact of perfectionism on patient alliance. That is, a client who is experiencing considerable anxiety over growing intimacy in the therapeutic relationship and does not show a typical increase in self-disclosure over time, may fail to establish as positive an alliance with their therapist, or group, as other clients. It would be worthwhile to explore this model in future studies. In addition, the residual unique predictive abilities of perfectionism, and specifically nondisclosure of imperfections, may reflect a direct association with negative outcome or the presence of other mediator variables. Other mediators might include patient expectations, beliefs about the perceptions of the therapist or group, reactions of therapists or other group members, and ability to feel connected to a therapy group. These factors will need to be explored in future studies. In interpreting mediational data, it is important to recall that mediation does not demonstrate causation. In other words, one cannot draw the conclusion that nondisclosure of imperfections causes post-session anxiety, which in turn causes negative outcome. By controlling for related variables such as depression and social anxiety in earlier analyses, some alternate causal models can be disregarded, but an infinite number of possibilities remain. Because personality research is destined to be correlational in nature (research participants cannot be randomly assigned to having a specific personality trait), an estimate of temporal sequencing has to suffice. Implications for the Perfectionism Construct  Results from this study promote and expand current conceptualizations of perfectionism as a multidimensional construct with both direct and indirect effects on psychopathology. First, this study advances work by Blatt et al. (1995, 1996. 1998) by introducing a conceptually distinct measure of perfectionism. Whereas Blatt and his colleagues used a unidimensional measure of general perfectionism attitudes, the use of the Perfectionistic Self-Presentation Scale (Hewitt et al., 2000) in this study provided insight into two distinct patterns of interpersonal behaviour related to perfectionistic concerns. This study demonstrated that perfectionistic behaviour styles can have important implications for therapy process and outcome, over and above perfectionism  PSP in Group Psychotherapy/page 96  traits or motivations. In other words, perfectionistic self-presentation contributes unique knowledge to our understanding of the therapeutic experience of perfectionistic clients. Thus, this study reinforces the importance of considering both perfectionistic motivations and the interpersonal expression of perfectionism. Hewitt and Flett (in press) discussed two primary models by which perfectionism could have an impact on psychopathology. First, perfectionism could lead directly to psychological difficulties through the creation of stress and enhancement of the stress experience. Several investigations have provided evidence in support of this direct model (e.g., Flynn et al., 2001; Hewitt & Flett, 1993; Hewitt, Flett & Ediger, 1996). In their second model, Hewitt and Flett (in press) proposed that perfectionism could act indirectly by preventing help-seeking or by having a detrimental impact on therapy outcome. Through these pathways, psychological symptoms might be exacerbated and prolonged. Research work related to the indirect model began only recently (e.g., Blatt et al., 1995; Habke & Hewitt, 2000). The present study was the first to illustrate that specific dimensions of perfectionism may be most intrusive in the therapy context. Along with Zuroff et al.'s (2000) paper, this study also was unique in explicitly testing mediational links between perfectionism, therapy process and therapy outcome. Therefore, this study provides essential support for Hewitt and Flett's (in press) indirect model, and expands on the model by exploring the mechanisms by which perfectionism might influence therapy outcome. Limitations & Future Directions  As with any applied clinical study, several limitations were inherent in the design. First, the program was based in a university research laboratory. While this allowed groups to be designed and conducted in an optimal manner for the study, results may not be generalizable to more typical outpatient settings. Group members often referred to the research nature of the therapy and the university setting. For example, some members spoke of group sessions as "class" and wondered if events during sessions were experimental manipulations. However, these individuals were seeking treatment for very real psychological problems that are typical of other outpatient clinics, including depression, past suicide attempts, relationship problems, anxiety,  PSP in Group Psychotherapy/page 97  stress-related physiological problems, and achievement difficulties. Thus, it is likely that many of the findings here would generalize to other settings. In addition to being based in a university setting, therapists for the groups were senior level graduate students in clinical psychology. While some research implies that therapist experience does not impact on therapy success (Durlak, 1979; Herman, 1993), more recent studies and reviews have suggested that greater experience and training is associated with improved therapy outcome (Burlingame, Fuhriman, Paul & Ogles, 1989; Stein & Lambert, 1995 ). Thus, level of 11  therapist experience is another factor that may influence the generalizability of these results to other settings. It is possible that more experienced therapists could have managed the groups in such a way as to minimize the negative impact of nondisclosure of imperfection. Still, the same might be said of any challenging issue in therapy, and this does not diminish the importance of identifying those clients who will require greater therapist attention and care. Undergraduate students assumed the role of coders in this study. While these students were carefully trained to understand the concepts of self-disclosure and different elements of the coding system, group therapists and supervisors sometimes were surprised by the low ratings of sadness within the groups, and by the inconsistencies between raters in codes for anxiety and happiness. In part, this may reflect the challenge of discerning affect from videotape. However, individuals more experienced with the therapy setting may have been better able to distinguish affective cues. If students and therapists do judge affective displays differently, it suggests that three questions can be considered. By using student coders, the present study may better address how non-clinicians, and perhaps other group members, would view a client's level of affect. Separate issues would be the level of affect noted by trained therapists, and the client's own assessment of their affective experience. Future studies interested in exploring the experience of different emotions during therapy sessions may wish to evaluate the phenomenon from each of these perspectives. Because of the time and investment required to provide quality assessment and treatment to clients, this study included only the minimum number of participants needed to attain reasonable  11  Also note the correction to this paper (Crits-Christoph, 1995).  PSP in Group Psychotherapy/page 98  power for the basic statistical analyses (Green, 1991). Other analyses, such as gender differences, were assessed on an exploratory basis, but without sufficient power to draw confident conclusions. A larger sample size, particularly with a larger pool of male group members, would offer greater power for discerning differences in these supplementary analyses. Two other statistical issues are inherent in most applied clinical studies. First, because of the large number of data collection points and the extended timeline of this study (over a period of several months), missing data were inevitable. Attempts were made to ensure that data were as complete as possible (e.g., participants were asked to review questions that were left blank on questionnaires, alternate video tapes were coded for participants who were absent from a given session). Several strategies have been suggested for dealing with missing data (Acock, 1997). Subscale substitution was used for small amounts of missing questionnaire data, an approach which is considered reasonable for estimating homogeneous data. However, data substitution for information such as missing post-treatment questionnaires, or ratings by friends and family members did not seem appropriate, as a comparable data source was not available. Pairwise deletion of cases with missing data was used to provide a best estimate of correlations and greater power. Acock cautions that although this technique appears to be more appropriate than listwise deletion, on regression analyses, degrees of freedom are automatically set to the smallest sample involved which leads to conservative effect estimates. Ideally, it would be nice to eliminate missing data completely, but this does not seem to be a realistic goal for a clinical study. Another statistical issue is the growing trend to use random regression techniques such as hierarchical linear modeling (HLM; Bryk & Raudenbush, 1987; Arnold, 1992) in analyses of data drawn from groups of participants. This approach provides two primary benefits for analyses of group treatment data. First, it allows group effects to be considered and controlled. Second, this technique permits change to be viewed as a continuous process by mapping individual growth curves (Francis, Fletcher, Stuebing, Davidson & Thompson, 1991). Issues such as individual differences in rates of change can then be explored. The decision of whether to use this new technique or to rely on more standard linear regression models was based on  PSP in Group Psychotherapy/page 99  several considerations. Because they are so new to the field, important statistical issues such as the impact of violations of assumptions for random regression models are not clear. Several authors have suggested that these need to be better understood (Bryk & Raudenbush, 1987; Nich & Carroll, 1997). Of particular import for this study, these models are based on assumptions of large sample sizes, a requirement that is difficult to achieve in clinical research (Gibbons et al., 1993; Nich & Carroll, 1997). Another factor is that many programs exist for constructing random regression models, each using slightly different approaches. Francis et al. (1991) note that it is difficult to evaluate the practical implications of these differences. Similarly, de Leeuw and Kreft (1995) commented on the many decisions "users" of such software must make in conducting these analyses. Without a good foundation of background research, the impact of these decisions is unclear. While these authors view HLM as "an elegant conceptualization", they also conclude that it is "not always necessary". HLM can only create growth curves in contexts where three or more time points of data are available (Bryk & Raudenbush, 1987; Francis et al., 1991). While this condition was certainly met for certain process measures in this study, outcome only was assessed pre- and post-treatment. As with any study, participants can only be expected to complete a reasonable number of measures in a valid and thoughtful manner. Because this line of research is so new, a large number of diverse effects needed to be considered. Therefore, we chose to limit testing of outcome measures to two time points, using a variety of measures. A final consideration in opting not to use random regression models was a desire to compare results from this study more directly to those that are being published in the perfectionism and psychotherapy field (e.g., Zuroff et al., 2000). Although a few studies of psychotherapy process and outcome have been published using HLM, these tend to be reanalyses of previously published data to explore the impact of this new statistical technique. For example, Gibbons et al. (1993) reanalyzed data from the NIMH Treatment of Depression Collaborative Research Program, which had been previously analyzed using more standard techniques and published by Elkin et al. (1989). It is reassuring to note that findings were quite similar in these two papers. For all of the reasons listed here, this study was analyzed using linear and hierarchical regression and mediational analyses, rather than HLM. Given the need  PSP in Group Psychotherapy/page  100  for more studies to explore the impact of HLM, it may be worthwhile to re-analyze some of the data from this study using HLM in the future. However, it would be most appropriate for use with those variables with multiple data points, such as self-reported anxiety, Cortisol, and goal achievement ratings. In light of the work by Gibbons and his colleagues (1993), similar results to those presented here would be anticipated. A final limitation is that fact that perfectionistic self-presentation is a newly developed construct. Although these dimensions of perfectionism have demonstrated good reliability and validity (Hewitt et al., 2000), further research will enable better understanding o f these variables and how they function in the real world. It would also be helpful to evaluate perfectionistic selfpresentation using multiple measures. At the present time, an interview is being developed to allow an assessing therapist or experimenter to rate participants on both trait and selfpresentational dimensions of perfectionism. Although several questions were answered by this study, many more can be raised on this issue. First, this study explored the impact of perfectionistic self-presentation on group treatment for perfectionism. The treatment focus on perfectionism likely had advantages and disadvantages for the study of group process. Although it enabled the recruitment of individuals with extremely high levels of perfectionism, the fact that perfectionism was being treated in the groups may have actually minimized some o f the expected effects of PSP on therapy process and outcome. Therapists were knowledgeable about perfectionism, monitored the level o f disclosure of various participants, and explored issues that are relevant to perfectionists such as social anxiety, concerns about revealing oneself, and acceptance. In so doing, anxiety and resistance to therapy by individuals with disclosure concerns may have been softened. Additionally, group dynamics may be different for a homogeneous group of perfectionists compared to a heterogeneous outpatient group. Cohesion tends to be greater in homogeneous groups (Yalom, 1995). Also, in a group of individuals who all have difficulty disclosing mistakes and flaws, there may be more space for frightened individuals to talk. In contrast, in a heterogeneous group, perfectionists may be better able to stay on the "sidelines" while other group members more willingly discuss their struggles. It will be important to replicate findings from this study in a  PSP in Group Psychotherapy/page  101  heterogeneous outpatient group in a typical clinic setting. In addition, it would be interesting to explore these process phenomena in individual therapy as well. In creating this group treatment program, pre-therapy training was included for the first two sessions to provide information on how to get the most out of the therapy experience, to anticipate pitfalls and challenges in therapy, and to provide a common language for talking about perfectionism. Other research teams have advocated the use of such preparation to prevent drop out and to improve therapy outcome (Bednar & Kaul, 1994; MacKenzie, 1990; Piper & Perrault, 1989; Zimpfer, 1991). In viewing the results of this study, questions arise as to whether these sessions helped individuals who scored highly on perfectionistic self-presentation to remain in the program and attempt self-disclosure despite their anxiety. Determining whether preparatory sessions are particularly useful for perfectionistic clients would be an interesting focus for future research. Several other questions are not addressed by the present study. For example, does perfectionistic self-presentation also influence interpersonal functioning in group psychotherapy? Zuroff et al. (2000) found that perfectionism limited increases in patients' contribution to the therapeutic alliance over the course of therapy. In groups, perfectionists may be less engaged and less well accepted by other group members. Further exploration of the timing of perfectionism's impact also is warranted. Throughout this paper, the suggestion has been made that growing intimacy and openness in the group, or in individual therapy, may be threatening to individuals with concerns about disclosing imperfections. However, this theory was not directly addressed by this study, and requires empirical testing. In their investigations of self-criticism, Rector and his colleagues (2000) emphasized the importance of change in self-criticism over the course of therapy as a predictor of treatment outcome. Future studies may wish to test the impact of changes in trait or self-presentational perfectionism on therapy outcome. It is possible that investigations of change in perfectionism may yield stronger associations with a variety of outcome measures. Blatt, Zuroff and their colleagues (1995, 1998) noted that perfectionism's impact remains apparent eighteen months following the termination of therapy. It will be worthwhile to follow-up with participants in the present study to see whether perfectionistic self-  PSP in Group Psychotherapy/page  102  presentation demonstrates such lasting effects. Finally, the positive association between nondisplay of imperfections and positive therapy outcome was somewhat surprising. Future studies are needed to clarify whether this finding is consistent and how this association may come about. Implications  The effects noted in this study were, admittedly, fairly small in size. However, when addressing an issue as complex as how clients experience psychotherapy and whether they are successful in dealing with a myriad of issues through therapy, it is not surprising that many different factors will contribute. Most studies that investigate personality variables' impact on psychotherapy outcome show comparable effect sizes (Garfield, 1994). In choosing to study this area, one must accept that effects may be small, but that together these pieces of information will allow us to understand important clinical issues. Indeed, this study has important implications for the study of group psychotherapy, the field of perfectionism research, and clinical practice. Despite the limitations noted, this study included methodological innovations that should advance research in group psychotherapy. First, it fits into a growing area of interest concerning individual differences in therapy process and outcome (Blatt, 1999; Garfield, 1994; Janowsky, 1999; Toseland & Siporin, 1986). This study proposes a new construct, perfectionistic selfpresentation, which appears to be a relevant factor in this field. Second, this study made an effort to explore therapy process and outcome. Many studies of therapy process have failed to make this link to outcome explicitly, rather than simply implying that a connection is present (Dies, 1993) . Perhaps most importantly, this study has combined both self-report and physiological measures of stress and outcome. While self-report measures are essential to provide information about an individual's perceptions of their personal experience, discrepancies have been noted between self-report and physiological data in stress research (e.g., Curtis et al., 1976; Walsh et al., 1994). In addition, self-report data may be more strongly influenced by the demand characteristics of the situation, particularly in the case of treatment outcome data (Garfield, 1994) . This study used measures from four different domains: self-report, others' report, behavioural observation and physiological data. By combining these sources of information, one  PSP in Group Psychotherapy/page  103  can be far more confident in the validity of any research findings. The use of physiological data in a treatment study is particularly unusual. Yet, the methodology of this paper demonstrates that physiological measurements can be used in a practical and useful manner in clinical settings. Within the perfectionismfield,this study supplements previous research suggesting that perfectionistic self-presentation can be detrimental to getting clients into therapy (Nielsen et al., 1997), and to their experiencing the therapy environment in a positive manner (Habke & Hewitt, 2000). This is thefirststudy to demonstrate that nondisclosure of imperfections has a negative impact on therapy process and outcome. It provided further evidence of differences between the two protective forms of perfectionistic self-presentation. In fact, some results even suggested that nondisplay of imperfection may predict more positive outcome following short-term group psychotherapy. Although nondisplay and nondisclosure of imperfections both reflect protective and concealing forms of self-presentation, they appear to result in very different emotional and behavioural profiles. Thus, the conceptualization of perfectionistic self-presentation as multidimensional appears to be valid. In addition, nondisclosure of imperfections predicted problems with therapy process and outcome, over and above the impact of trait perfectionism. This supports the notion that perfectionistic self-presentation is distinct from trait perfectionism, and that it's influence may be particularly relevant in interpersonal contexts, such as psychotherapy. Differences in benefit associated with nondisplay and nondisclosure of imperfection, also raise the question of the ideal length of therapy. It may be that longer-term treatment is required to adequately address concerns about disclosure of imperfections. Blatt (1999) also came to the conclusion that perfectionism may require intensive, long-term therapy. In a reanalysis of data from the Menninger Psychotherapy Study, Blatt found that perfectionistic clients made significant gains in long-term psychoanalysis. Thus, future studies may need to consider the utility of matching various dimensions of perfectionism and treatment duration and modality. Findings in this study suggest that clinicians do need to pay extra attention to the functioning of clients who have fears of disclosing their imperfections. Therapists should be aware of the anxiety that these individuals experience and their inclination to avoid the challenge of therapy by missing sessions without notification. The timing of these difficulties is of  PSP in Group Psychotherapy/page  104  particular importance. When clients appear to be performing well in the early phases of treatment, therapists may forget to check in with them or to ensure that they remain comfortable later in the treatment process. Given that nondisclosure of imperfections is associated with greater discomfort late in the course of therapy, therapists need to be particularly vigilant as the group grows in trust and cohesion, to ensure that perfectionists are not left behind. Past studies have shown how difficult it is for these individuals to even seek psychotherapy assistance (Nielsen et al., 1997). Clinicians need to be aware of this struggle, of these clients' initial efforts to meet the demands of therapy, but also of their fear and temptation to avoid the challenges of self-revelation and intimacy. In exploring the impact of perfectionistic self-presentation on group therapy process and outcome, evidence was provided to support the hypothesis that nondisclosure of imperfections can be detrimental to clients' comfort and participation in therapy, as well as their ability to benefit from group therapy. Consistent with previous work (Blatt et al., 1998), the impact of perfectionistic self-presentation was noted in the latter third of the therapy sessions. This suggests that growing openness and intimacy in the group may be particularly troublesome for individuals who fear revealing their imperfections. Although effect sizes were small, they were significant and contribute to our understanding of the complex phenomenon of individual differences in therapy process and outcome. Clinically, this study highlights the importance of monitoring fears of disclosing imperfections to help predict difficulties during periods of increasing intimacy. With this awareness and through the persistence and courage of these clients, we may be able to develop strategies to minimize its impact.  PSP in Group Psychotherapy/page 105  REFERENCES Acock, A.C. (1997). Working with missing data. Family Science Review, 10(1), 76-102. Aitken, R.C. (1970). Communication of symptoms. Psychotherapy and Psychosomatics, 18 (1-6), 74-79. Alden, L.E., Wiggins, J.S., & Pincus, A. L. (1990). Construction of the circumplex scales for the Inventory of Interpersonal Problems. Journal of Personality Assessment, 55(3-4), 521-  536. Allen, J.G. (1974). Implications of research in self-disclosure for group psychotherapy. International Journal of Group Psychotherapy, 24, 306-321.  Altman, I., & Haythorn, W.W. (1965). Interpersonal exchange in isolation. Sociometry, 28, 411-426. Arkin, R.M., & Hermann, A.D. (2000). Constructing desirable identities - Self-presentation in psychotherapy and daily life: Comment on Kelly (2000). Psychological Bulletin, 126(4), 501-504. Arnetz, B.B., & Fjellner, B. (1986). Psychological predictors of neuroendocrine responses to mental stress. Journal of Psychosomatic Research, 30, 297-305. Arnetz, B.B., Fjellner, B., Kallner, A., & Eneroth, P. (1985). Stress and psoriasis: Psychoendocrine and metabolic reactions in psoriatic patients during standardized stressor exposure. Psychosomatic Medicine, 47, 528-541. Arnold, C.L. (1992). An introduction to hierarchical linear models. Measurement and Evaluation in Counseling and Development, 25, 58-90.  Assor, A. (1996). Two types of motivation for recognition: Secure and insecure. Psychological Reports, 79, 913-914. Barlow, D.H., & Craske, M.G. (1994). Mastery ofyour Anxiety and Panic II. Albany, New York: Graywind Publications Inc. Baron, R.M., & Kenny, D.A. (1986). The moderator-mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology, 51(6), 1173-1182.  Bastiani, A.M., Rao, R., Weltin, T., & Kaye, W.H. (1995). Perfectionism in anorexia nervosa. International Journal of Eating Disorders, 17, 147-152.  Baumeister, R.F., Tice, D.M., & Hutton, D.G. (1989). Self-presentational motivations and personality differences in self-esteem. Journal of Personality, 57(3), 547-579. Beck, A.T., Epstein, N., Brown, G., & Steer, R.A. (1988). An inventory for measuring clinical anxiety: Psychometric properties. Journal of Consulting and Clinical Psychology, 56, 893-  897.  PSP in Group Psychotherapy/page 106  Beck, A.T., & Steer, R.A. (1987). Beck Depression Inventory - Manual. Toronto: The Psychological Corporation, Harcourt Brace Jovanovich, Inc. Beck, A.T., & Steer, R.A. (1991). Relationships between the Beck Anxiety Inventory and the Hamilton Anxiety Rating Scale with anxious outpatients. Journal ofAnxiety Disorders, 5, 213-223. Beck, A.T., Steer, R.A., & Garbin, M.G. (1988). Psychometric properties of the Beck Depression Inventory: Twenty-five years of evaluation. Clinical Psychology Review, 8, 77100. Bednar, R.L., & Kaul, T J . (1994). Experimental group research: Can the canon fire? InA.E. Bergin, & S.L. Garfield (Eds.), Handbook of Psychotherapy and Behavior Change. New  York: John Wiley & Sons, Inc. Blankstein, K.R., Flett, G.L., Hewitt, P.L., & Eng, A. (1993). Dimensions of perfectionism and irrational fears: An examination with the Fear Survey Schedule. Personality and Individual Differences, 15(3), 323-328.  Blatt, S.J. (1999). Personality factors in brief treatment of depression: Further analyses of the NIMH-sponsored Treatment for Depression Collaborative Research Program. In Janowsky, D.S. (Ed.), Psychotherapy Indications and Outcomes. Washington: American Psychiatric Press, Inc. Blatt, S.J., Quinlan, D.M., Pilkonis, P.A., & Shea, M.T. (1995). Impact of perfectionism and need for approval on the brief treatment of depression: The National Institute of Mental Health treatment of depression collaborative research program revisited. Journal of Consulting and Clinical Psychology, 63(1), 125-132.  Blatt, S.J., Quinlan, D.M., Pilkonis, P.A., & Shea, M.T. (1995b). "Impact of perfectionism and need for approval on the brief treatment of depression: The National Institute of Mental Health treatment of depression collaborative research program revisited": Correction. Journal of Consulting and Clinical Psychology, 63(3), 494.  Blatt, S.J., Zuroff, D.C., Quinlan, D.M., & Pilkonis, P.A. (1996). Interpersonal factors in brief treatment of depression: Further analyses of the National Institute of Mental Health Treatment of depression collaborative research program. Journal of Consulting and Clinical Psychology, 64(1), 162-171.  Blatt, S.J., Zuroff, D.C., Bondi, C M . , Sanislow, C.A., & Pilkonis, P.A. (1998). When and how perfectionism impedes the brief treatment of depression: Further analyses of the National Institute of Mental Health Treatment of Depression Collaborative Research Program. Journal of Consulting and Clinical Psychology, 66(2), 423-428.  Bloch, S., & Crouch, E. (1985). Therapeutic Factors in Group Psychotherapy. Oxford: Oxford University Press. Brien, T.G. (1980). Free Cortisol in human plasma. Hormone Metabolism Research, 12,643650.  PSP in Group Psychotherapy/page 107  Bryk, A.S., & Raudenbush, S.W. (1987). Application of hierarchical linear models to assessing change. Psychological Bulletin, 101(1), 147-158.  Burlingame, G.M., Fuhriman, A., Paul, S., & Ogles, B.M. (1989). Implementing a time-limited therapy program: Differential effects of training and experience. Psychotherapy, 26(3), 303313. Burns, D.D. (1980). Feeling Good: The New Mood Therapy. New York: New American Library. Burns, D.D., & Nolen-Hoeksema, S. (1991). Coping styles, homework compliance, and the effectiveness of cognitive-behavioral therapy. Journal of Consulting and Clinical Psychology, 59(2), 305-311.  Bornstein, R.F., Riggs, J.M., Hill, E.L., & Calabrese, C. (1996). Activity, passivity, selfdenigration, and self-promotion: Toward an interactionist model of interpersonal dependency. Journal of Personality, 64(3), 637-673.  Buss, A.H, & Finn, S.E. (1987). Classification of personality traits. Journal of Personality and Social Psychology, 52(2), 432-444.  Cepeda-Benito, A., & Short, P. (1998). Self-concealment, avoidance of psychological services, and perceived likelihood of seeking professional help. Journal of Counseling Psychology, 45(1), 58-64. Clark, M.S., Pataki, S.P., & Carver, V.H. (1996). Some thoughts and findings on selfpresentation of emotions in relationships. In G. J.O. Fletcher, & J. Fitness, (Eds.), Knowledge structures in close relationships: A social psychological approach, (pp. 247-274). Mahwah,  NJ: Lawrence Erlbaum Associates. Collins, N.L., & Miller, L.C. (1994). Self-disclosure and liking: A meta-analytic review. Psychological Bulletin, 116(3), 457-475.  Cozby, P.L. (1973). Self-disclosure: A literature review. Psychological Bulletin, 79(2), 73-91. Connelly, J.L., & Piper, W.E. (1989). An analysis of pretraining work behavior as a composition variable in group psychotherapy. International Journal of Group Psychotherapy, 39(2), 173-189.  Conte, H.R., Plutchik, R., Picard, S., & Karasu, T.B. (1991). Can personality traits predict psychotherapy outcome? Comprehensive Psychiatry, 32(1), 66-72. Conte, H.R., Plutchik, R., Picard, S., Karasu, T.B., & Vaccaro, D. (1988). Self-report measures as predictors of psychotherapy outcome. Comprehensive Psychiatry, 29(4), 355-360. Crits-Christoph, P. (1995). "Graduate training in psychotherapy: Are the outcomes enhanced?": Comment on Stein and Lambert (1995). Journal of Consulting and Clinical Psychology, 63(5), 717.  PSP in Group Psychotherapy/page 108  Curtis, G., Buxton, M., Lippman, D., Nesse, R., & Wright, J. (1976). "Flooding in vivo" during the circadian phase of minimal Cortisol secretion: Anxiety and therapeutic success without adrenal Cortisol activation. Biological Psychiatry, 11, 101-107. de Leeuw, J., & Kreft, I.G.G. (1995). Questioning multilevel models. Journal of Educational and Behavioral Statistics, 20(2), 171-189.  Dienstbier, R.A. (1989). Arousal and physiological toughness: Implications for mental and physical health. Psychological Review, 96, 84-100. Dies, R.R. (1993). Research on group psychotherapy: Overview and clinical applications. In A. Alonso, & H.I. Swiller (Eds.), Group Therapy in Clinical Practice (pp. 473-518). Washington: American Psychiatric Press. Durlak, J.A. (1979). Comparative effectiveness of paraprofessional and professional helpers. Psychological Bulletin, 181(1), 80-92.  Elkin, I., Shea, M.T., Watkins, J.T., Imber, S.D., Sotsky, S.M., Collins, J.F., Glass, D.R., Pilkonis, P.A., Leber, W.R., Docherty, J.P., Fiester, S.J., Parloff, M.B. (1989. NIMH Treatment of Depression Collaborative Research Program: General effectiveness of treatments. Archives of General Psychiatry, 46, 971-983.  Endler, N.S., & Magnusson, D. (1976). Toward an interactional psychology of personality. Psychological Bulletin, 83, 956-97'4.  Ey, S., Henning, K.R., & Shaw, D.L. (2000). Attitudes and factors related to seeking mental health treatment among medical and dental students. Journal of College Student Psychotherapy, 14(3), 23-39.  Falloon, I.R.H. (1981). Interpersonal variables in behavioural group therapy. British Journal of Medical Psychology, 54, 133-141.  Ferguson, D.B. (1984). Physiological considerations in the use of salivary steroid estimation for clinical investigations. In Y. Kawamura & D.B. Ferguson (Eds.), Frontiers of Oral Physiology, 5, 1-20.  Fisher, E.H., & Turner, J.L. (1970). Orientations to seeking professional help: Development and research utility of an attitude scale. Journal of Consulting and Clinical Psychology, 35, 79-90. Flett, G.L., & Hewitt, P.L. (in press). Perfectionism in the family context. In Flett, G.L., & Hewitt, P.L., (Eds.), Perfectionism: Theory, Research & Practice. New York: American Psychological Association. Flett, G.L., Hewitt, P.L., Blankstein, K.R., & Dynin, C.B. (1994). Dimensions of perfectionism and type A behaviour. Personality and Individual Differences, 7(5,477-485. Flett, G.L., Hewitt, P.L., Blankstein, K.R., & O'Brien, S. (1991b). Perfectionism and learned resourcefulness in depression and self-esteem. Personality and Individual Differences, 12, 61-68.  PSP in Group Psychotherapy/page 109  Flett, G.L., Hewitt, P.L., Endler, N.S., & Tassone, C. (1995). Perfectionism and components of state and trait anxiety. Current Psychology, 13(4), 326-350. Flett, G.L., Hewitt, P.L., Garshowitz, M., & Martin, T.R. (1997). Personality, negative social interactions, and depressive symptoms. Canadian Journal of Behavioural Science, 29(1), 28-  37. Flynn, C.A., Hewitt, P. L., & Flett, G. L. (1996). Perfectionism and family relationships from a cross-cultural perspective. -Poster presented at the XXVI International Congress of Psychology, Montreal, PQ, Canada. Flynn, C.A., Hewitt, P.L., Flett, G.L., & Weinberg, J. (2001). Perfectionism and physiological response to an achievement-related stressor. Manuscript in preparation.  Flynn, C.A., Hewitt, P.L., Broughton, R., & Flett, G.L. (1998). Mapping Perfectionism Dimensions onto the Interpersonal Circumplex. Paper presented at the annual convention of the American Psychological Association, San Francisco, CA. Francis, D.J., Fletcher, J.M., Stuebing, K.K., Davidson, K . C , & Thompson, N.M. (1991). Analysis of change: Modeling individual growth. Journal of Consulting and Clinical Psychology, 59(1), 27-37.  Frankenhaeuser, M. (1979). Psychoneuroendocrine approaches to the study of emotion as related to stress and coping. In Howe, H., & Dienstbier, R. (eds.), Nebraska Symposium on Motivation. Lincoln: University of Nebraska Press. Frost, R.O., Marten, P., Lahart, C , & Rosenblate, R. (1990). The dimensions of perfectionism. Cognitive Therapy and Research, 14, 449-468.  Garfield, S.L. (1994). Research on client variables in psychotherapy. In A.L. Bergin, & S.L. Garfield (Eds.), Handbook of Psychotherapy and Behavior Change, 4th Ed. New York:  John Wiley & Sons, Inc. Garner, D.M., Olmsted, M.P., & Polivy, J. (1983). The Eating Disorder Inventory: A measure of cognitive-behavioral dimensions of anorexia nervosa and bulimia. In Darby, P., Garfinkel, P., Garner, D., & Coscina, D. (Eds.), Anorexia nervosa: Recent developments. New York: Alan R. Liss. Gibbons, R.D., Hedeker, D., Elkin, I., Waternaux, C , Kraemer, H.C., Greenhouse, J.B., Shea, T., Imber, S.D., Sotsky, S.M., & Watkins, J.T. (1993). Some conceptual and statistical issues in analysis of longitudinal psychiatric data: Application to the NIMH Treatment of Depression Collaborative Research Program dataset. Archives of General Psychiatry, 50, 739-750. Goldstein, A.P., & Shipman, W.G. (1961). Patient expectancies, symptom reduction and aspects of the initial psychotherapeutic interview. Journal of Clinical Psychology, 17, 129133. Good, G.E., Dell, D.M., & Mintz, L.B. (1989). Male role and gender role conflict: Relations to help-seeking in men. Journal of Counseling Psychology, 36(3), 295-300.  PSP  in Group  Psychotherapy/page  110  Green, S.B. (1991). How many subjects does it take to do a regression analysis? Multivariate Behavioral Research, 26(3), 449-510.  Gross, JJ., & Levenson, R.W. (1997). Hiding feelings: The acute effects of inhibiting negative and positive emotion. Journal of Abnormal Psychology, 106(1), 95-103.  Grunebaum, H. (1975). A soft-hearted review of hard-nosed research on groups. International Journal of Group Psychotherapy, 25(2), 185-197.  Gurtman, M. B. (1996). Interpersonal problems and the psychotherapy context: The construct validity of the Inventory of Interpersonal Problems. Psychological Assessment, 8, 241-255. Habke, A.M., & Flynn, C.A. (In press). Interpersonal aspects of trait perfectionism. In Flett, G.L., & Hewitt, P.L., (Eds.), Perfectionism: Theory and Research. New York: American Psychological Association. Habke, A.M., & Hewitt, P.L. (2000). Cognitive Appraisals, Affect and Arousal as Related to Perfectionistic Self-Presentation: Implications for Therapy. Manuscript in preparation.  Habke, A.M., Hewitt, P.L., Fehr, B., Callander, L., & Flett, G.L. (1997). Perfectionism and Behavior in Marital Interactions. Paper presented at the annual convention of the Canadian Psychological Association, Toronto, ON, Canada. Hammond, G. (1996). The objections to null hypothesis testing as a means of analyzing psychological data. Australian Journal of Psychology, 48(2), 104-106.  Hellhammer, D.H., Kirschbaum, C , & Belkien, L. (1987). Measurement of salivary Cortisol under psychological stimulation. In J.N. Hingtgen, D.H. Hellhammer, & G. Huppmann (Eds.), Advanced Methods in Psychobiology. Toronto: Hogrefe. Herman, K.C. (1993). Reassessing predictors of therapist competence. Journal of Counseling and Development, 72(1), 29-32.  Hewitt, P.L., & Flett, G.L. (1991a). Dimensions of perfectionism in unipolar depression. Journal ofAbnormal Psychology, 100, 98-101.  Hewitt, P.L., & Flett, G.L. (1991b). Perfectionism in the self and social contexts: Conceptualization, assessment, and association with psychopathology. Journal of Personality and Social Psychology, 60, 456-470.  Hewitt, P.L., & Flett, G.L. (1993). Dimensions of perfectionism, daily stress, and depression: A test of the specific vulnerability hypothesis. Journal ofAbnormal Psychology, 102, 58-65. Hewitt, P.L., & Flett, G.L. (1997). Perfectionism traits, perfectionistic self-presentation and  dyadic adjustment. Paper presented at the annual convention of the American Psychological Association, Chicago, IL. Hewitt, P.L., & Flett, G.L. (2000). Perfectionism: Clinical Implications and Treatment.  Manuscript in preparation.  PSP in Group Psychotherapy/page  111  Hewitt, P.L., & Flett, G.L. (In press). Perfectionism and stress processes. In Flett, G.L. & Hewitt, P.L. (Eds.), Perfectionism: Theory and Research. New York: American Psychological Association. Hewitt, P.L., Flett, G.L., & Ediger, E. (1995). Perfectionism traits and perfectionistic selfpresentation in eating disorder attitudes, characteristics, and symptoms. International Journal of Eating Disorders, 18(4), 317-326.  Hewitt, P.L., Flett, G.L., & Ediger, E. (1996). Perfectionism and depression: Longitudinal assessment of a specific vulnerability hypothesis. Journal ofAbnormal Psychology, 105(2), 276-280. Hewitt, P.L., Flett, G.L., Ediger, E., Norton, G.R., & Flynn, C.A. (1998). Perfectionism in chronic and state symptoms of depression. Canadian Journal of Behavioural Science, 30(4),  234-242. Hewitt, P.L., Flett, G.L., Habke, M., Parkin, M., Lam, R.W., Ediger, E., & Stein, M. (2000). The Perfectionistic Self-Presentation Scale: Development and Psychometric Properties of an  Interpersonal Perfectionism Measure. Unpublished manuscript, University of British Columbia, Vancouver, BC. Hewitt, P.L., Flett, G.L., Han, H., Tomlin, M., & Nielsen, A.D. (2001). Help-seeking and the role ofperfectionistic self-presentation. Manuscript in preparation.  Hewitt, P. L., Flett, G. L., & Mikail, S. F. (1995). Perfectionism and relationship adjustment in pain patients and their spouses. Journal of Family Psychology, 9, 335-347. Hewitt, P.L., Flett, G.L., & Turnbull, W. (1994). Borderline personality disorder: An investigation with the Multidimensional Perfectionism Scale. European Journal of Psychological Assessment, 10, 28-33.  Hewitt, P.L., Flett, G.L., & Turnbull, W. (1992). Perfectionism and Multiphasic Personality Inventory (MMPI) indices of personality disorder. Journal of Psychopathology and Behavioral Assessment, 14(4), 323-335.  Hewitt, P.L., Flett, G.L., & Turnbull-Donovan, W. (1992). Perfectionism and suicide potential. British Journal of Clinical Psychology, 31, 181-190.  Hewitt, P.L., Flett, G.L., & Weber, C. (1994). Dimensions of perfectionism and suicide ideation. Cognitive Therapy and Research, 18(5), 439-460.  Hewitt, P.L., Flynn, C.A., & Low, G. (1999). Manual for the Psychotherapy Self-Disclosure Coding System. Manuscript in preparation. Hewitt, P.L., Newton, J., Flett, G.L., & Callander, L. (1997). Perfectionism and suicide ideation in adolescent psychiatric patients. Journal ofAbnormal Child Psychology, 25(2), 95-101.  Hewitt, P.L., & Norton, G.R. (1993). The Beck Anxiety Inventory: A psychometric analysis. Psychological Assessment, 5(4), 408-412.  PSP in Group Psychotherapy/page 112 Hewitt, P.L., Norton, G.R., Flett, G.L., Callander, L., & Cowan, T. (1998). Dimensions of perfectionism, hopelessness, and attempted suicide in a sample of alcoholics. Suicide and Life Threatening Behavior, 28(4), 395-406.  Hill, C.E., Gelso, C.J., & Mohr, J.J. (2000). Client concealment and self-presentation in therapy: Comment on Kelly (2000). Psychological Bulletin, 126(4), 495-500. Hill, R.W., Zrull, M . C , & Turlington, S. (1997). Perfectionism and interpersonal problems. Journal of Personality Assessment, 69, 81-103.  Hobden, K., & Pliner, P. (1995). Self-handicapping and dimensions of perfectionism: Selfpresentation vs. self-protection. Journal of Research in Personality, 29, 461-474. Horowitz, L.M., Rosenberg, S.E., Baer, B.A., Ureno, G., & Villasefior, V.S. (1988). The Inventory of Interpersonal Problems: Psychometric properties and clinical applications. Journal of Consulting and Clinical Psychology, 56, 885-892.  Horvath, A.O., & Symonds, B.D. (1991). Relation between working alliance and outcome in psychotherapy: A meta-analysis. Journal of Consulting Psychology, 38(2), 139-149.  Hunter, J.E. (1997). Needed: A ban on the significance test. Psychological Science, 8(1), 3-7. Imber, S.D., Pilkonis, P.A., Sotsky, S.M., Elkin, I., Watkins, J.T., Collins, J.F., Shea, M.T., Leber, W.R., & Glass, D.R. (1990). Mode-specific effects among three treatments for depression. Journal of Consulting and Clinical Psychology, 58(3), 352-359.  Janowski, D.S. (1999). Therapist and patient personality characteristics and the nature, quality, and outcome of psychotherapy: Focus on the Myers Briggs Type Indicator. In Janowsky, D.S. (Ed.), Psychotherapy Indications and Outcomes. Washington: American Psychiatric Press, Inc. Jones, E.E., Pittman, T.S. (1993). Toward a general theory of strategic self-presentation. In J.M. Suis (Eds.), The self in social perspective. Psychological perspectives on the self, (Vol.  4, pp. 231-262). Hillsdale, NJ: Lawrence Erlbaum Associates. Kaye, W.H. (1997). Perfectionism in anorexia nervosa: Reply. American Journal of Psychiatry, 154(1), 132.  Kelly, A.E. (2000a). Helping construct desirable identities: A self-presentational view of psychotherapy. Psychological Bulletin, 126(4), 475-494.  Kelly, A.E. (2000b). A self-presentational view of psychotherapy: Reply to Hill, Gelso, and Mohr (2000) and to Arkin and Hermann (2000). Psychological Bulletin, 126(4), 505-511. Kirschbaum, C , & Hellhammer, D.H. (1989). Salivary Cortisol in psychobiological research: An overview. Neuropsychobiology, 22, 150-169. Kirschbaum, C , Strasburger, C.J., Jammers, W., & Hellhammer, D.H. (1989). Cortisol and behavior: 1. Adaptation of a radioimmunoassay kit for reliable and inexpensive salivary Cortisol determination. Pharmacology Biochemistry & Behavior, 34, 747-751.  PSP in Group Psychotherapy/page 113  Klein, R.H. & Carroll, R.A. (1986). Patient characteristics and attendance patterns in outpatient group psychotherapy. International Journal of Group Psychotherapy, 36(1), 115-132.  Kowal, A., & Pritchard, D. (1990). Psychological characteristics of children who suffer from headache: A research note. Journal of Child Psychology and Psychiatry, 31, 637-649. Landon, J. (1982). Closing session discussion. Proceedings of the ninth Tenovus workshop. In Read, G.F., Riad-Fahmy, D., Walker, R.F., & Griffiths, K. (eds), Immunoassays of Steroids in Saliva. Cardiff: Alpha Omega Publishing Ltd., 1984, 343-346. Lang, P.J., Levin, D.N., Miller, G.A., & Kozak, M.J. (1983). Fear behavior, fear imagery, and the psychophysiology of emotion: The problem of affective response integration. Journal of Abnormal Psychology, 92(3), 276-306.  Leary, M.R. (1993). The interplay of private self-processes and interpersonal factors in selfpresentation. In J.M. Suls (Eds.), The Self in Social Perspective. Psychological perspectives  on the self, (Vol. 4, pp. 127-155). Hillsdale, NJ: Lawrence Erlbaum Associates, Inc. Leary, M.R., & Kowalski, R.M. (1993). The Interaction Anxiousness Scale: Construct and criterion-related validity. Journal of Personality Assessment, 61(1), 136-146.  Lee, S-J, Quigley, B.M., Nesler, M.S., Corbett, A.B., & Tedeschi, J.T. (1999). Development of a self-presentation tactics scale. Personality and Individual Differences, 26, 701-722. Linden, W., Dadgar, N., & Earle, T.L. (1994). Cortisol responses to acute stress: A selective review. Psychosomatic Medicine, 56, 153-154. Luborsky, L., et al. (1985). Therapist success and its determinants. Archives of General Psychiatry, 42, 602-611. Luborsky, L., Chandler, M., Auerbach, A.H., Cohen, J., & Bachrach, H.M. (1971). Factors influencing the outcome of psychotherapy: A review of quantitative research. Psychological Bulletin, 75(3), 145-185.  Luria, R.E. (1975). The validity and reliability of the Visual Analogue Mood Scale. Journal of Psychiatry Research, 12, 51-57.  MacKenzie, K.R. (1983). The clinical application of a group climate measure. In Dies, R.R., & MacKenzie, K.R., (Eds.), Advances in Group Psychotherapy: Integrating Research and  Practice. New York: International Universities Press. MacKenzie, K.R. (1990). Introduction to Time-Limited Group Psychotherapy. Washington,  DC: American Psychiatric Press, Inc. Maling, M.S., Gurtman, M.B., & Howard, K.I. (1995). The response of interpersonal problems to varying doses of psychotherapy. Psychotherapy Research, 5(1), 63-75. Meleshko, K.G.A., & Alden, L.E. (1993). Anxiety and self-disclosure: Toward a motivational model. Journal of Personality and Social Psychology, 64(6), 1000-1009.  PSP in Group Psychotherapy/page 114 McCallum, M., Piper, W.E., & Joyce, A.S. (1992). Dropping out from short-term group therapy. Psychotherapy, 29(2), 206-215. McCallum, M., Piper, W.E., & Morin, H. (1993). Affect and outcome in short-term group therapy for loss. International Journal of Group Psychotherapy, 43(3), 303-319.  Moore-Ede, M . C , Sulzman, F.M., & Fuller, C A . (1982). The Clocks That Time Us. Cambridge, Massachusetts: Harvard University Press. Morey, L . C (1991). The Personality Assessment Inventory. Odessa: Psychological Assessment Resources. Mussell, M.P., Mitchell, J.E., Crosby, R.D., Fulkerson, J.A., Hoberman, H.M., & Romano, J.L. (2000). Commitment to treatment goals in prediction of group cognitive-behavioral therapy treatment outcome for women with bulimia nervosa. Journal of Consulting and Clinical Psychology, 68(3), 432-437.  Nich, C , & Carroll, K. (1997). Now you see it, now you don't: A comparison of traditional versus random-effects regression models in the analysis of longitudinal follow-up data from a clinical trial. Journal of Consulting and Clinical Psychology, 65(2), 252-261.  Nielsen, A.D., Hewitt, P.L., Han, H., Habke, A.M., Cockell, S.J., Stager, G., & Flett, G.L. (1997). Perfectionistic Self-Presentation and Attitudes Toward Professional Help Seeking.  Paper presented at the annual convention of the Canadian Psychological Association, Toronto, ON, Canada. O'Malley, S.S., Suh, C.S., 8c Strupp, H.H. (1983). The Vanderbilt Psychotherapy Process Scale: A report on the scale development and a process-outcome study. Journal of Consulting and Clinical Psychology, 51(4), 581-586.  Orlinsky, D.E., Grawe, K., & Parks, B.K. (1994). Process and outcome in psychotherapy Noch Einmal. In A.E. Bergin, & S.L. Garfield (Eds.), Handbook of Psychotherapy and Behavior Change. New York: John Wiley & Sons, Inc. Parkin, C M . , Hewitt, P.L., Flynn, C.A., & Flett, G.L. (1999). A validation of the Perfectionistic Self-Presentation Scale. Paper to be presented at the annual convention of the Canadian Psychological Association, Halifax, NS. Paulhus, D.L. (1991). Measurement and control of response bias. In Robinson, J.P., & Shaver, P.R., (Eds.), Measures of Personality and Social Psychological Attitudes. Measures of Social Psychological Attitudes, Vol.1. San Diego, CA: Academic Press, Inc. Paulhus, D. L. (1994). Balanced Inventory of Desirable Responding: Reference Manual for BIDR Version 6. Vancouver, B.C.  Paulhus, D.L., & Reid, D.B. (1991). Enhancement and denial in socially desirable responding. Journal of Personality and Social Psychology, 60(2), 307-317.  Pikus, C.F., & Heavey, C L . (1996). Client preferences for therapist gender. Journal of College Student Psychotherapy, 10(4), 35-43.  PSP in Group Psychotherapy/page 115  Piper, W.E. (1991). Brief group psychotherapy. Psychiatric Annals, 21(7), 419-422. Piper, W.E., Ogrodniczuk, J.S., Joyce, A.S., McCallum, M., Rosie, J.S., O'Kelly, J.G., & Steinberg, P.I. (1999). Prediction of dropping out in time-limited, interpretive individual psychotherapy. Psychotherapy, 36(2), 114-122. Piper, W.E., Perrault, E.L. (1989). Pretherapy preparation for group members. International Journal of Group Psychotherapy, 31(1), 17-34.  Rampling, D.J., & Williams, R.A. (1977). Evaluation of group processes using visual analogue scales. Australian and New Zealand Journal of Psychiatry, 11, 189-191.  Rector, N.A., Bagby, R.M., Segal, Z.V., Joffe, R.T., & Levitt, A. (2000). Self-criticism and dependency in depressed patients treated with cognitive therapy or pharmacotherapy. Cognitive Therapy and Research, 24(5), 571-584.  Riad-Fahmy, D., Read, G.F., Walker, R.F., & Griffiths, K. (1982). Steroids in saliva for assessing endocrine function. Endocrine Reviews, 3, 367-395. Roth, D.L., Snyder, C.R., & Pace, L.M. (1986). Dimensions of favorable self-presentation. Journal of Personality and Social Psychology, 51(4), 867-874.  Saltzman, C., Luetgert, M.J., Roth, C.H., Creaser, J., & Howard, L. (1976). Formation of a therapeutic relationship: Experiences during the initial phase of psychotherapy as predictors of treatment duration and outcome. Journal of Consulting and Clinical Psychology, 44(4),  546-555. Sexton, H.C., Hembre, K., & Kvarme, G. (1996). The interaction of the alliance and therapy microprocess: A sequential analysis. Journal of Consulting and Clinical Psychology, 64(3),  471-480. Schlenker, B.R., & Weigold, M.F. (1992). Interpersonal Processes Involving Impression Regulation and Management. Annual Review of Psychology, 43, 133-168. Shay, J.J. (1996). "Okay, I'm here, but I'm not talking": Psychotherapy with the reluctant male. Psychotherapy, 33(3), 503-513. Shea, M.T., Elkin, I., & Sotsky, S.M. (1999). Patient characteristics associated with successful treatment: Outcome findings from the NIMH Treatment of Depression Collaborative Research Program. In Janowsky, D.S. (Ed.), Psychotherapy Indications and Outcomes. Washington, DC: American Psychiatric Press, Inc. Sotsky, S.M., Elkin, I., Watkins, J.T., Collins, J.F., Shea, M.T., & Leber, W.R. (1990). Modespecific effects among three treatments of depression. Journal of Consulting and Clinical Psychology, 58(3), 352-359.  Spangler, D.L., & Burns, D.D. (1999). Is it true that men are from Mars and women are from Venus? A test of gender differences in dependency and perfectionism. Journal of Cognitive Psychotherapy: An International Quarterly, 13(4), 339-357.  PSP in Group Psychotherapy/page 116  Spanier, G. (1976). Measuring dyadic adjustment: New scales for assessing the quality o f marriage and similar dyads. Journal of Marriage and the Family, 38, 15-28. Spence, J.T., & Robbins, A.S. (1992). Workaholism: Definition, measurement, and preliminary results. Journal of Personality Assessment, 58, 160-178.  Spielberger, C.D., Gorsuch, R.C., & Lushene, R.E. (1970). Manual for the State-Trait Anxiety Inventory. Palo Alto, CA: Consulting Psychologists. Stein, D.M., & Lambert, M.J. (1995). Graduate training in psychotherapy: Are therapy outcomes enhanced? Journal of Consulting and Clinical Psychology, 63(2), 182-196.  Stone, W.N., Blaze, M., & Bozzuto, J. (1980). Late dropoutsfromgroup psychotherapy. American Journal of Psychotherapy, 34(3), 401-413.  Sullivan, H.S. (1953). The interpersonal theory of psychiatry. New York: Norton. Szabo, CP., & Terre Blanche, M.J. (1997). Perfectionism in anorexia nervosa. American Journal of Psychiatry, 154(1), 132.  Tardy, C H . (1988). Self-disclosure: Objectives and methods of measurement. In Tardy, C H . (Ed.), A Handbook for the Study of Human Communication: Methods and Instruments for Observing, Measuring, and Assessing Communication Processes. Communication and  Information Science. Norwood, NJ: Ablex Publishing Corp., 324-346. Teyber, E. (2000). Interpersonal process in psychotherapy: a relational approach. Belmont,  CA: Brooks/Cole. Toseland, R.W., & Siporin, M. (1986). When to recommend group treatment: A review of the clinical and the research literature. International Journal o f Group Psychotherapy, 36(2), 171-201. Tschuschke, V., & Dies, R.R. (1994). Intensive analysis of therapeutic factors and outcome in long-term inpatient groups. International Journal of Group Psychotherapy, 44(2), 185-208.  Tschuschke, V., MacKenzie, K.R., Haaser, B., & Janke, G. (1996). Self-disclosure, feedback, and outcome in long-term inpatient psychotherapy groups. Journal of Psychotherapy Practice and Research, 5(7),35-44.  Umeda, T., Hiramatsu, R., Iwaoka, T., Shimada, T., Miura, F., & Sato, T. (1981). Use o f saliva for monitoring unboundfreeCortisol levels in serum. Clinica Chimica Acta, 110, 245-253. Venables, P.H., & Christie, M.J. (1980). Electrodermal activity. In Martin, I. & Venables, P.H. (Eds.), Techniques in Psychophysiology. Chichester: John Wiley & Sons. Vining, R.F., & McGinley, R.A. (1985). Hormones in saliva. CRC Critical Reviews in Clinical Laboratory Sciences, 23, 95-146.  Walsh, J.J., Wilding, J.M., & Eysenck, M.W. (1994). Stress responsivity: The role o f individual differences. Personality and Individual Differences, 16(3), 385-394.  PSP in Group Psychotherapy/page 117  Waring, E.M., & Chelune, G.J. (1983). Marital intimacy and self-disclosure. Journal of Clinical Psychology, 39(2), 183-190.  Waring, E.M., Schaefer, B., & Fry, R. (1994). The influence of therapeutic self-disclosure on perceived marital intimacy. Journal of Sex and Marital Therapy, 20(2), 135-146. Weissman, A., & Beck, A.T. (1978, December). Development and validation of the dysfunctional attitude scale. Paper presented at the 12th Annual Meeting of the Association for the Advancement of Behavior Therapy, Chicago, IL. As cited in Imber et al., (1990). Mode-specific effects among three treatments for depression. Journal of Consulting and Clinical Psychology, 58(3), 352-359.  Weitzman, E.D., Fukushima, D., Nogeire, C. Roffwarg, H., Gallagher, T.F., & Heilman, L. (1971). Twenty-four hour pattern of the episodic secretion of Cortisol in normal subjects. Journal of Clinical Endocrinology, 33, 14-22.  Wiggins, J.S., Trapnell, P., & Phillips, N. (1988). Psychometric and geometric characteristics of the Revised Interpersonal Adjective Scale (IAS-R). Multivariate Behavioral Research, 23, 517-530. Wolfe, R.N., & Lennox, R.D. (1986). Getting Along and Getting Ahead: Empirical support for a theory of protective and acquisitive self-presentation. Journal of Personality and Social Psychology, 50(2), 356-361.  Yalom, I.D. (1967). Prediction of improvement in group therapy. Archives of General Psychiatry, 17, 159-168. Yalom, I.D. (1995). The Theory and Practice of Group Psychotherapy. Fourth Edition. New  York: BasicBooks. Zimpfer, D.G. (1991). Pretraining for group work: A review. The Journal for Specialists in Group Work, 16(4), 264-269.  Zuckerman, D.M., Prusoff, B.A., Weissman, M.M., & Padian, N.S. (1980). Personality as a predictor of psychotherapy and pharmacotherapy outcome for depressed outpatients. Journal of Consulting and Clinical Psychology, 48(6), 730-735.  Zuroff, D.C., Blatt, S.J., Sotsky, S.M., Krupnick, J.L., Martin, D.J., Salislow III, C.A., & Simmens, S. (2000). Relation of therapeutic alliance and perfectionism to outcome in brief outpatient treatment of depression. Journal of Consulting and Clinical Psychology, 68(1),  114-124.  PSP in Group Psychotherapy/page  118  Table 1 Means, standard deviations and alphas of predictor variables and dependent variables. MEASURE (items)  MEAN  ST. DEVIATION  ALPHA  Predictors: Perfectionistic Self-Presentation Scale (27) Nondisplay (10) Nondisclosure (7)  55.37 30.31  9.32 8.95  .88 .90  87.99 72.50 69.04  9.15 13.92 16.38  .84 .82 .85_  Total -- pre-session (12) Total - post-session (12) Early - pre-session (2) Early -- post-session (2) Late - pre-session (2) Late -- post-session (2)  31 23 28 34 35 43 35 02 25 83 26 40  16 45 17 39 24 52 23 06 21 13 23 26  Assessment Anticipation of 1 session Post-pregroup session Post first actual session Post second to last session  6.51 6.83 5.65 7.14 9.06  6.04 6.80 4.59 4.44 8.54  No show with excuse No show without excuse Late arrival Incomplete questionnaires Attendance (actual sessions)  1.15 0.46 1.34 1.01 8.46  1.99 0.81 1.62 1.27 1.95  4.63 4.71 4.01 5.49  1.61 1.76 1.85 2.14  Multidimensional Perfectionism Scale (45) Self-oriented (15) Other-oriented (15) Socially Prescribed (15) Dependent Variables: Mood Ratings - Anxiety  Cortisol (ng/ml) st  Resistant Behaviours  Patient Outcome Ratings Satisfaction Benefited Goal Achievement Ratings (midpoint) Goal Achievement Ratings (post-treatment)  PSP in Group Psychotherapy/page  MEAN  M E A S U R E (items)  ST. DEVIATION  Close Others' Ratings of Patient Pre-Treatment Self-oriented perfectionism (15) Other-oriented perfectionism (15) Socially prescribed perfectionism (15) Nondisplay of imperfections (15) Nondisclosure of imperfections (15) Post-Treatment Self-oriented perfectionism (15) Other-oriented perfectionism (15) Socially prescribed perfectionism (15) Nondisplay of imperfections (15) Nondisclosure of imperfections (15)  89.12 69.42 67.26 50.27 32.53 82.11 66.95 66.56 47.47 30.18  10.09 15.91 17.16 11.49 8.93 13.18 12.63 14.73 11.02 8.06  Benefited  4.77  1.58  4.60  1.47  Therapists' Post-Treatment Ratings  Benefited  119  ALPHA  .86 .89 .90 .89 .84 .89 .85 .88 .88 .83  Beck Depression Inventory  Pre-Treatment (21) Post-Treatment (21)  17.44 11.10  8.44 8.10  .86 .91  Beck Anxiety Inventory  Pre-Treatment (21) Post-Treatment (21)  15.36 10.36  10.39 7.50  .91 .88  1.69 1.40  0.63 0.63  .96 .99  3.74 0.74  4.02 3.55  0.38 0.25  0.64 0.68  3.39 -1.12  4.65 3.83  0.41 0.26  0.46 0.43  Inventory of Interpersonal Problems Pre-Treatment: Mean Post-Treatment: Mean Autonomic Reactivity Pre-treatment  HEART RATE  Elevation during the test Decline relative to baseline post-test SKIN CONDUCTANCE LEVEL Elevation during the test Decline relative to baseline post-test Post-treatment  HEART RATE  Elevation during the test Decline relative to baseline post-test SKIN CONDUCTANCE LEVEL Elevation during the test Decline relative to baseline post-test Control Variable: Interaction Anxiousness Scale (15)  47.77  12.15  .91  PSP in Group Psychotherapy/page  Table 1 (cont'd)  Late Session  Early Session Mean  SD  Mean  SD  3.15 12.18 5.28  3.30 9.06 3.89  4.59 13.78 7.50  4.26 11.14 6.10  10.37 0.43 0.13  7.23 0.97 0.52  11.58 0.41 0.06  7.33 1.02 0.24  3.41 3.99 2.44 1.09  3.64 3.50 2.05 1.84  4.68 3.83 2.50 1.06  3.22 4.32 1.98 1.93  2.03 2.16 1.25  1.80 2.22 0.90  2.02 2.08 1.06  2.00 2.35 0.35  Disclosure Valence Positive Negative Neutral  Personal Content Personal 1 Personal 2 Personal 3  Focus In group Out group Both Neither  Mood Ratings Anxiety Sadness Happiness  120  PSP in Group Psychotherapy/page  121  Table 2: Correlations between protective forms of perfectionistic self-presentation, trait perfectionism and other control variables. NONDISPLAY OF IMPERFECTION r  P  NONDISCLOSURE OF IMPERFECTION r  Trait Perfectionism Self-oriented Other-oriented Socially prescribed  jg*** .30** .67***  <.001 .010 <.000  .48*** .22-  Related Variables Depression (BDI) Social anxiety (IAS) Use of medications  42*** .54*** .23*  <.001 <.001 .049  .30**  59***  2g***  .22-  P <.001 .072 <.001  .010 <.001 .062  ~p<.10; * p<.05; **p<.01; ***p<.001 Bonferroni correction: .10/12 p < . 0 0 4  Note: Scores on nondisplay of imperfection can range from 10 to 70, nondisclosure of imperfections from 7 to 49, trait perfectionism subscales from 15 to 105, BDI scores from 0 to 21, and IAS scores from 15 to 75. For each of these scales, higher scores reflect greater difficulties with the variable in question. Use of medications is a dummy coded variable with 0 reflecting no use of medications and 1 reflecting use of medications.  Table 3: Correlations between protective forms of perfectionistic self-presentation and mood rating variables. NONDISPLAY OF IMPERFECTION Anxiety Total Early Sessions Late Sessions  Pre-session r p .21- .092 .31** .010 .16 .225  Post-session r p .30* .018 .33** .006 .18 .152  NONDISCLOSURE OF IMPERFECTION Pre-session r P .450 .10 .13 .303 .724 .05  Post-session r p .29* .022 .21.086 .25* .044  ~p<.10; * p<.05; ** p<.01 Bonferroni correction: .10/8 p < . 0 1 2 5  Note: Scores on nondisplay of imperfection can range from 10 to 70, nondisclosure of imperfections can range from 7 to 49, and anxiety ratings were scored using a 100-point scale. For each of these measures, higher scores reflect greater difficulties with the variable in question.  PSP in Group Psychotherapy/page Table 4: Multiple regression analyses predicting post-session anxiety at various points in treatment.  R Change 2  F change  Predicting: Post-session Anxiety - Across All Sessions Step 1: Pre-session .464 .464 51.84*** anxiety Step2:PSPS  .510  .047  Nondisplay Nondisclosure Predicting: Post-session Anxiety - Early Sessions Step 1: Pre-session .184 .184 anxiety Step2:PSPS .231 .047 Nondisplay Nondisclosure Predicting: Post-session Anxiety - Late Sessions Step 1: Pre-session .313 .313 anxiety Step2:PSPS .405 .093 Nondisplay Nondisclosure ~ p<.10; * p<.05; **p<.01; *** p<.001  Beta  .68***  rj  <.001  2.76.05 .19-  14.67***  .629 .092  .43***  <.001  .18 07  .181 .608  .56***  <.001  1.93 1  26.83***  4.43* .04 .28*  .738 .017  122  PSP in Group Psychotherapy/page  123  Table 5: M u l t i p l e regression analyses predicting post-session anxiety following all sessions, controlling for trait perfectionism and related variables.  F change R Change R Predicting: Post-session Anxiety - Across All Sessions 2  Stepl: Pre-session anxiety  .460  .460  Step 2: Social anxiety Depression Use o f M e d s  .537  .078  Step 3: M P S SOP OOP SPP  .548  50.17***  Beta  _ *=M 68  3.14* .04 .07  .691 .015 .489  .10 .05 .04  .367 .635 .746  .01 .14  .939 .312  .26*  .557 Step 4: P S P S DISP DISC * p<.05; **p<.01; *** p<.001 SOP=Self-oriented perfectionism; SPP=Socially-prescribed perfectionism; DISC=Nondisclosure of imperfection  .011  .009  <.001  0.43  0.52  OOP=Other-oriented perfectionism; DISP=Nondisplay of Imperfection;  PSP in Group Psychotherapy/page  Table 6: Multiple regression analyses predicting post-session anxiety following sessions late in therapy, controlling for trait perfectionism and related variables. R  2  R Change 2  F change  Beta  26.20***  .56***  p  Predicting: Post-session Anxiety - Late Sessions  Step 1: Pre-session anxiety  .311  .311  Step 2: Social anxiety Depression UseofMeds  .368  .056  Step 3: MPS SOP OOP SPP  .376  Step 4: PSPS DISP DISC  .465  * p<05; **p<.01; * * * p<001 SOP=Self-oriented perfectionism; SPP=Socially-prescribed perfectionism; DISC=Nondisclosure of imperfection  .008  .089  <.001  1.64 .12 .17 .00  .324 .148 .977  .06 .07 -.08  .616 .618 .573  .11  .471 .007  0.22  4.17*  OOP=Other-oriented perfectionism; DISP=Nondisplay of Imperfection;  124  PSP in Group Psychotherapy/page  125  Table 7:  Correlations between perfectionism and salivary C o r t i s o l . NONDISCLOSURE OF IMPERFECTION  NONDISPLAY OF IMPERFECTION Cortisol Assessment Anticipation Post-pregroup Post V' actual Post 9 actual th  * p<.05 Bonferroni correction: .10/10  r  P  r  p  -.01  .923  .27*  mi  -.11  .376 .584 .651  .07  .563  -.07 -.06  .01 -.11  .964 .405  .00  1.000  -.00  .989  p<.01  Note: Scores on nondisplay of imperfection can range from 10 to 70, nondisclosure of imperfections from 7 to 49, and salivary Cortisol concentrations from .05 to 38.89 ng/mL. For each of these measures, higher scores reflect greater levels of the variable in question.  Table 8: Multiple regression analyses predicting salivary C o r t i s o l concentrations during the assessment session, controlling for trait perfectionism and other related variables. R  2  R Change 2  F change  Beta  P  -.22-  .074  -.24.31* .11  .072 .026 .392  .16 .07 -.03  .248 .620 .865  -.14 .34*  .431 .044  Predicting: Anxiety- Assessment Session Step 1: Time of testing  .049  .049  3.29-  Step 2: Social anxiety Depression Use of Meds  .157  .108  2.60-  Step 3: M P S SOP OOP SPP  .187  Step 4: PSPS DISP DISC  .251  ~ p<.10; * p<.05 SOP=Self-oriented perfectionism; SPP=Socially-prescribed perfectionism; DISC=Nondisclosure of imperfection  .030  .064  0.72  2.41-  OOP=Other-oriented perfectionism; DISP=Nondisplay of Imperfection;  PSP in Group Psychotherapy/page  126  Table 9: Correlations between perfectionism and resistant therapy behaviours.  NONDISPLAY OF IMPERFECTION r  P  r  P  -.02  .889  -.05  .709  .09  .448  .26*  .032  -.13  .270  .01  .943  .03  .825  -.08  .491  -.03  .838  -.13  .304  Resistant Behaviours No show with excuse No show without excuse Late arrival Incomplete questionnaires Attendance  NONDISCLOSURE OF IMPERFECTION  * p<.05 Bonferroni correction: .10/10 p<.01  Note: Scores on nondisplay of imperfection can range from 10 to 70, nondisclosure of imperfections can range from 7 to 49, and disruptive behaviours were monitored for all 12 therapy sessions. Scores on these measures ranged from 0 to 11. Attendance was assessed for the 'actual' therapy sessions and ranged from 1 to 10 sessions. For each of all of these measures, with the exception of attendance, higher scores reflect greater problems.  PSP in Group Psychotherapy/page  127  Table 10: Multiple regression analyses predicting not showing up to sessions without providing an excuse, controlling for trait perfectionism and ofher.related variables. R .054 2  Step 1: Social anxiety Depression Use of Meds  R Change .054 2  F change 1.25  Step 2: MPS SOP OOP SPP  .084  .030  0.69  Step 3: PSPS DISP DISC  .139  .055  1.97  ~ p<.10; SOP=Self-oriented perfectionism; SPP=Socially-prescribed perfectionism; DISC=Nondisclosure of imperfection  Beta  P  .13 .04 -.21-  .338 .758 .094  .06 -.12 .16  .669 .395 .285  -.05 .32-  .798 .053  OOP=Other-oriented perfectionism; DISP=Nondisplay of Imperfection;  PSP in Group Psychotherapy/page  128  Table 11: Correlations between perfectionism and observational data from an early session.  NONDISPLAY OF IMPERFECTION  Disclosure Valence Positive Negative Neutral  NONDISCLOSURE OF IMPERFECTION r  r  P  -.03 .07 -.09  .828 .573 .446  .05 .29* .13  .689 .015 .275  -.12  .331  .21-  .087  .069  .29*  .015  P  Focus In group or both  Mood Ratings Sadness  .22-  ~p<10; *p<.05 Bonferroni correction: .10/10 p<.01  Note: Scores on nondisplay of imperfection can range from 10 to 70, and scores on nondisclosure of imperfections can range from 7 to 49, with higher values reflecting greater difficulty with perfectionism. The frequency of disclosures of varying valences and foci were noted and scores ranged from 0 to 40 with higher numbers reflecting a greater frequency of that type of disclosure. Ratings of sadness were made on a 10-point scale with higher values reflecting greater sadness.  PSP in Group Psychotherapy/page  129  Table 12: Correlations between perfectionism and observational data from a late session.  NONDISPLAY OF IMPERFECTION Disclosure Valence Positive Negative Neutral Focus In group or both Mood Ratings Sadness  NONDISCLOSURE OF IMPERFECTION r  r  P  -.02 .21.01  .893 .092 .926  -.03 -.02 -.09  .815 .876 .480  .09  .490  .16  .189  .10  .405  .10  .392  P  ~p<10; *p<05 Bonferroni correction: .10/10 p<.01 Note: Scores on nondisplay o f imperfection can range from 10 to 70, and scores on nondisclosure o f imperfections can range from 7 to 49, with higher values reflecting greater difficulty with perfectionism. The frequency o f disclosures o f varying valences and foci were noted and scores ranged from 0 to 63 with higher numbers reflecting a greater frequency o f that type o f disclosure. Ratings o f sadness were made on a 10-point scale with higher values reflecting greater sadness.  PSP in Group Psychotherapy/page  130  Table 13: Correlations between perfectionistic self-presentation, self-disclosure totals, valence and focus, and observed sadness scores across all coded sessions. NONDISPLAY OF IMPERFECTIONS Total Disclosures  Overall Early Middle Late  NONDISCLOSURE OF IMPERFECTIONS  r  P  r  .04 .01 -.05 .12  .748 .916 .679 . .309  .02 .25* -.10 -.05  .869 .041 .439 .700  -.04 .12 -.12  .747 .326 .321  -.00 .06 -.06  .985 .648 .620  -.03  .828  .19  .135  P  Valence Ratio  Positive Negative. Neutral Focus Overall  In group or both  -  Mood Overall  .15  Sadness ~p<.10;  .227  .20  .102  *p<05  Bonferroni correction: .10/18 p<.005  Note: Scores on nondisplay of imperfection can range from 10 to 70, and scores on nondisclosure of imperfections can range from 7 to 49, with higher values reflecting greater difficulty with perfectionism. The frequency of disclosures of varying valences and foci were noted and scores ranged from 0 to 183 with higher numbers reflecting a greater frequency of that type of disclosure. Ratings of sadness were made on a 10-point scale with higher values reflecting greater sadness.  PSP in Group Psychotherapy/page  131  Table 14: Regression analysis of perfectionistic self-presentation predicting disclosure late in therapy, after controlling for level of disclosure early in therapy, trait perfectionism and other related variables. R  R Change  F change  Beta  Early disclosures  .127  .127  9.20*  .36*  .004  Step 2: Social anxiety Depression Use of Meds  .142  .014  0.34 .11 .00 .06  .401 .989 .605  Step 3: MPS SOP OOP SPP  .166  .024  0.54 .14 .13 .06  .358 .379 .687  Step 4: PSPS DISP DISC  .252  .11 .39*  .566 .016  Step 1:  * p<.05; **p<.01 SOP=Self-oriented perfectionism; SPP=Socially-prescribed perfectionism; DISC=Nondisclosure of imperfection  .086  3.17*  OOP=Other-oriented perfectionism; DISP=Nondisplay of Imperfection;  PSP in Group Psychotherapy/page  132  Table 15: Regression analysis of perfectionistic self-presentation predicting negative self-disclosures late in therapy, controlling for disclosures of this valence early in therapy, trait perfectionism and other related variables. R^  R Change  F change  Beta  .193  .193  15.05***  44***  Step 2: Social anxiety Depression Use of Meds  .233  .040  Step 3: MPS SOP OOP SPP  .242  Step4:PSPS DISP DISC  .356  2  Step 1:  2  p_  Early  negative  <.001  disclosures  ~ p<.10; * p<05; ** p<01 SOP=Self-oriented perfectionism; SPP=Socially-prescribed perfectionism; DISC=Nondisclosure of imperfection  .010  .113  1.04 .21~ -.05 .01  .087 .684 .943  .09 -.09 .03  .559 .507 .814  .20 -.44**  .271 .004  0.25  4.83*  OOP=Other-oriented perfectionism; DISP=Nondisplay of Imperfection;  PSP in Group Psychotherapy/page 133  Table 16: Regression analysis of perfectionistic self-presentation predicting ingroup and combined ingroup/outgroup self-disclosures, controlling for trait perfectionism and other related variables. R  2  R Change 2  Step 1: Social anxiety Depression Use of Meds  .029  Step 2: MPS SOP OOP SPP  .166  .136  Step 3: PSPS DISP DISC  .198  .032  ~ p<.10; * p<05; ** p<.01 SOP=Self-oriented perfectionism; SPP=Socially-prescribed perfectionism; DISC=Nondisclosure of imperfection  .029  F change  Beta  P  0.62 -.78 .05 .06  .195 .717 .658  .33* -.06 .15  .019 .688 .317  -.26 .11  .178 .503  3.16*  1.12  OOP=Other-oriented perfectionism; DISP=Nondisplay of Imperfection;  PSP in Group Psychotherapy/page  134  Table 17: Correlations between perfectionism and ratings of treatment satisfaction and global benefit. NONDISPLAY OF IMPERFECTIONS  NONDISCLOSURE OF IMPERFECTIONS  r  P  r  .23.12 -.08  .065 .343 .547  .11 .05 -.16  .394 .710 .202  .10  .480  -.14  .308  CLOSE OTHERS' RATINGS Benefit  .09  .553  .08  .606  THERAPIST RATINGS Benefit  .28*  .026  .33**  .008  PATIENT RATINGS Satisfaction Benefit Goal Achievement Ratings (midpoint) Goal Achievement Ratings (post-tx)  P  ~p<.10; *p<.05; ** p<01 Bonferroni correction: .10/12 p<.008 Note: Scores on nondisplay o f imperfection can range from 10 to 70, and scores on nondisclosure o f imperfections can range from 7 to 49, with higher values reflecting greater difficulty with perfectionism. Ratings o f satisfaction and benefit were made on a seven-point scale ranging from " d i d not benefit" to "benefited significantly". Higher scores reflect greater perceived benefit from the program. Goal achievement ratings were indicated on a 10-point scale anchored by the statements " n o progress" and " g o a l achieved". Higher ratings indicate greater success i n meeting personal therapy goals.  PSP in Group Psychotherapy/page  135  Table 18: Multiple regression analyses of therapist ratings of patient benefit from treatment, controlling for trait perfectionism and other related variables. R .088 2  Step 1: Social Anxiety Depression Use of meds  R Change .088 2  Step 2: MPS SOP OOP SPP  .171  .084  Step 3: PSPS DISP DISC ~ p<10  .184  .013  SOP=Self-oriented perfectionism; SPP=Socially-prescribed perfectionism; DISC=Nondisclosure of imperfection  F change 1.93  Beta  P  .08 .25.01  .535 .063 .930  .14 .14 .16  .333 .327 .314  .03 .15  .883 .369  1.92  0.42  OOP=Other-oriented perfectionism; DISP=Nondisplay of Imperfection;  PSP in Group Psychotherapy/page  136  Table 19: Correlations between perfectionistic self-presentation and post-treatment outcome measures. NONDISPLAY OF IMPERFECTIONS TRAIT PERFECTIONISM Self-oriented Other-oriented Socially-prescribed  r .21 .28* 44***  P  NONDISCLOSURE OF IMPERFECTIONS r  P  .116 .029 <.001  .28* .12 .36**  .029 .382 .005  .096 .084 .064  .33** .09 .17  .010 .518 .198  SYMPTOMS Depression (BDI) Anxiety (BAI) Interpersonal Problems (mean IIP)  .22.23.24-  ~ p < 1 0 ; *p<.05; * * p < 0 1 ; *** p<001 Bonferroni correction: .10/12 p<.008  Note: Scores on nondisplay of imperfection can range from 10 to 70, nondisclosure of imperfections from 7 to 49, trait perfectionism subscales from 15 to 105, BDI and BAI scores from 0 to 21, and IIP scores from 0 to 4. In each case, higher values reflect greater psychopathology.  PSP in Group Psychotherapy/page  137  Table 20: Partial correlations between perfectionistic self-presentation and post-treatment outcome measures, controlling for pre-treatment levels of these variables.  NONDISPLAY OF IMPERFECTIONS r  P  r  .06 .06 .16  .633 .656 .223  .10 -.02 .09  .445 .886 .487  .04 .11 -.11  .745 .419 .404  .23-.00 -.08  .078 .991 .551  TRAIT PERFECTIONISM Self-oriented Other-oriented Socially-prescribed  NONDISCLOSURE OF IMPERFECTIONS P  SYMPTOMS Depression (BDI) Anxiety (BAI) Interpersonal Problems (mean IIP) ~p<10 Bonferroni correction: .10/12 p<.008  Note: Scores on nondisplay of imperfection can range from 10 to 70, nondisclosure of imperfections from 7 to 49, trait perfectionism subscales from 15 to 105, BDI and BAI scores from 0 to 21, and IIP scores from 0 to 4. In each case, higher values reflect greater psychopathology.  PSP in Group Psychotherapy/page  138  Table 21: Multiple regression analyses predicting post-treatment depression, controlling for pre-treatment depression, trait perfectionism and other related variables R  2  R Change  F change  Beta  13.24***  42***  .001  .12 .12  .371 .332  .10 .24 -.20  .447 .101 .192  -.02 .33*  .930 .042  2  Step 1: BDI pre-treatment  .191  .191  Step 2: Social Anxiety Use of Meds  .219  .028  0.96  Step 3: MPS SOP OOP SPP  .282  .063  1.50  Step 4: PSPS DISP DISC  .341  .059  * p<.05; **p<.01; *** p<.001 SOP=Self-oriented perfectionism; SPP=Socially-prescribed perfectionism; DISC=Nondisclosure of imperfection  2.18  OOP=Other-oriented perfectionism; DISP=Nondisplay of Imperfection;  P  PSP in Group Psychotherapy/page  139  Table 22: Correlations between perfectionistic self-presentation and others' ratings of perfectionism preand post-treatment NONDISPLAY OF IMPERFECTIONS  Member ratings >  NONDISCLOSURE OF IMPERFECTIONS  Close other ratings PRETREATMENT Self-oriented Perfectionism Other-oriented Perfectionism Socially-prescribed Perfectionism Nondisplay of Imperfections Nondisclosure of Imperfections POSTTREATMENT Self-oriented Perfectionism Other-oriented Perfectionism Socially-prescribed Perfectionism Nondisplay of Imperfections Nondisclosure of Imperfections ' - ./ ~p<10;  r  P  r  .05  .677  .01  .927  .11  .407  -.14  .288  .23-  .069  .14  .272  .24-  .061  -.04  .749  .26*  .041  .19  .149  .00  .997  -.05  .739  .16  .277  .03  .834  .18  .216  -.01  .937  .12  .406  -.08  .578  .30*  .035  .34*  P  .018  I  *p<.05  Bonferroni correction:  .10/20 p<.005  Note: Scores on nondisplay of imperfection can range from 10 to 70, nondisclosure of imperfections from 7 to 49, and trait perfectionism subscales from 15 to 105. In each case, higher values reflect greater psychopathology.  PSP in Group Psychotherapy/page  140  Table 23: Correlations between perfectionistic self-presentation and autonomic reactivity in response to a stress test. NONDISPLAY O F IMPERFECTIONS  r  PRETREATMENT  Heart rate elevation during testing Heart rate decline post-testing SCL elevation during testing SCL decline posttesting  P  NONDISCLOSURE O F IMPERFECTIONS  r  P  .030  .24-  .055  -.07  .572  .14  .254  .02  .848  -.05  .677  -.01  .928  -.07  .575  -.05  .691  .07  .600  -.17  .200  -.06  .631  .12  .357  .21  .109  .10  .464  .18  .181  .27*  POSTTREATMENT  Heart rate elevation during testing Heart rate decline post-testing SCL elevation during testing SCL decline posttesting ~p<10; *p<.05  Bonferroni correction: .10/16 p<.006  Note: Scores on nondisplay of imperfection can range from 10 to 70, nondisclosure of imperfections from 7 to 49, with higher values reflecting greater psychopathology. Heart rate elevation and decline (or return to baseline) scores ranged from -12.72 indicating a decline in heart rate between the time points, to 22.71 indicating a strong rise in heart rate. Similarly, SCL elevation and decline scores ranged from -1.95 to 2.86. Again, higher scores reflected a stronger stress response.  PSP in Group Psychotherapy/page  141  Table 24: Partial correlations between perfectionistic self-presentation and post-treatment autonomic reactivity, controlling for pre-treatment reactivity. NONDISPLAY OF IMPERFECTIONS  Heart rate elevation during testing Heart rate decline post-testing SCL elevation during testing SCL decline posttesting  NONDISCLOSURE OF IMPERFECTIONS  r  P  r  -.04  .761  .15  .296  -.15  .278  .03  .820  .08  .587  .28*  .041  .04  .778  .19  .181  P  * p<.05 Bonferroni correction: .10/8 p<.013 Note: Scores on nondisplay of imperfection can range from 10 to 70, nondisclosure of imperfections from 7 to 49, with higher values reflecting greater psychopathology. Heart rate elevation and decline (or return to baseline) scores ranged from -12.72 indicating a decline in heart rate between the time points, to 22.71 indicating a strong rise in heart rate. Similarly, SCL elevation and decline scores ranged from -1.95 to 2.86. Again, higher scores reflected a stronger stress response.  PSP in Group Psychotherapy/page 142  Table 25: Multiple regression analyses predicting post-treatment SCL reactivity in response to a stress test, controlling for pre-treatment SCL reactivity, trait perfectionism and other related variables. R^ R Change F change Predicting: Post-treatment SCL elevation during the test Stepl: .064 .064 3.49SCL elevation pre-treatment 2  Step 2: Social Anxiety Depression UseofMeds  .083  Step 3: MPS SOP OOP SPP  .113  Step 4: PSPS DISP DISC  .274  ~ p<10; * p<.05; ** p<.01 SOP=Self-oriented perfectionism; SPP=Socially-prescribed perfectionism; DISC=Nondisclosure of imperfection  2  .019  .030  .160  Beta  p_  .25-  .068  0.33 -.06 .01 -.12  .704 .972 .409  .12 ..10 .15  .454 .585 .408  0.51  4.75* .12 .56** OOP=Other-oriented perfectionism; DISP=Nondisplay of Imperfection;  .566 .005  PSP in Group Psychotherapy/page  Table 26: Differing ratings for male and female group members. Female Group Members X Disruptive Therapy Behaviours: No show without excuse Ratings of Benefit by Group Members: Treatment Satisfaction Global Benefit Goal Achievement Ratings (midtx) Goal Achievement Ratings (posttx) SCLelevation during the test Pre-treatment Post-treatment SCL - return to baseline Pre-treatment Post-treatment ~ p<.10;  Male Group Members x a  P  1.43  2.29  0.63*  1.06  .042  4.91  1.43  4.09*  1.82  .043  5.10  1.56  4.00*  1.93  .016  4.25  1.72  3.57-  2.03  .100  5.91  2.14  4.78*  1.97  .026  0.43 0.33  0.71 0.32  0.28 0.57-  0.50 0.65  .067  0.35 0.19  0.63 0.31  0.08 0.42-  0.73 0.59  .066  * p<.05  NB: Where pre- and post-treatment data are given, gender differences are noted in posttreatment ratings, controlling for pre-treatment ratings.  143  PSP in Group Psychotherapy/page  144  Table 27: Interactions between gender and perfectionistic self-presentation in predicting disclosures late in therapy, controlling for the frequency of disclosures early in therapy. R .126  R Change .126  F change 9.27**  .211  .084  2.18-  2  Step 1: Early disclosures Step 2: Main effects Nondisplay Nondisclose Gender  2  .057  Step 3: .268 Interactions Nondisplay x gender Nondisclosure x gender ~p<10; *p<05; ** p<.01  Beta .36**  P .003  .28* -.32* .10  .038 .027 .381  .06  .944  -1.02-  .089  2.28  Table 28: Interactions between gender and perfectionistic self-presentation in predicting negative disclosures late in therapy, controlling for the frequency of negative disclosures early in the R .197  R Change .197  F change 15.68***  .317  .272  3.57*  2  Step 1: Early negative disclosures Step 2: Main effects Nondisplay Nondisclose Gender  .369 Step 3: Interactions Nondisplay x gender . Nondisclosure x gender ~p<10; *p<.05; ** p<01  2  .052  Beta  P <.001  .36** -.37** .05  .005 .007 .677  -.37  .655  -.78  .157  2.43-  PSP in Group Psychotherapy/page  145  Table 29: Interactions between gender and perfectionistic self-presentation in predicting positive disclosures late in therapy, controlling for the frequency of positive disclosures early in therapy. R .047  R Change .047  .077  .030  2  Step 1: Early positive disclosures Step 2: Main effects Nondisplay Nondisclose Gender  2  .114  .191 Step 3: Interactions Nondisplay x gender Nondisclosure x gender ~p<10; *p<05; ** p<01  F change 3.16-  Beta .22-  P .080  .03 -.08 .17  .838 .579 .176  -.07  .939  -1.38*  .030  0.66  4.15*  Table 30: Interactions between gender and perfectionistic self-presentation in predicting ingroup disclosures late in therapy, controlling for the frequency of ingroup disclosures early in ther R .116  R Change .116  .136  .019  2  Step 1: Early ingroup disclosures Step 2: Main effects Nondisplay Nondisclose Gender  .216 Step 3: Interactions Nondisplay x gender Nondisclosure x gender ~p<10; *p<05; ** p<01  2  .080  F change 8.43**  Beta 34**  P .005  0.45 .11 .04 -.04  .446 .779 .770  -.19  .837  -1.10-  .076  3.03-  PSP in Group Psychotherapy/page  Table 31: Interactions between gender and perfectionistic self-presentation in predicting ratings of perceived benefit by close others. B}  R Change  Step 1: Main effects Nondisplay Nondisclose Gender  .024  .024  Step 2: Interactions Nondisplay x gender Nondisclosure x gender ~p<.10; *p<05  .135  2  .111  F change  Beta  £  .07 .05 -.12  .715 .795 .437  -2.59*  .040  0.34  2.51-  .76  J  .339  Table 32: Interactions between gender and perfectionistic self-presentation in predicting post-treatment skin conductance return to baseline following a stress test, controlling for post-stress return to baseline pre-treatment. R}  R Change 2  F change  Step 1: Pre-tx SCL return to baseline  027  .027  1.40  Step 2: Main effects Nondisplay Nondisclose Gender  131  .105  1.93  Step 2: Interactions Nondisplay x gender Nondisclosure x gender ~p<10; *p<05  227  .096  Beta  p_  .16  .242  -.12 .29-.26-  .473 .083 .066  3.39*  .021  -1.20-  .099  2.85-  146  PSP in Group Psychotherapy/page  Table 33: Interactions between gender and perfectionistic self-presentation in predicting change in nondisclosure of imperfections as rated by close others. R^ .330  R Change .330  .372  .041  2  Step 1:  2  F change 22.70***  Beta .58***  £ <.001  Pre-tx other-rated nondisclose Step 2: Main effects Nondisplay Nondisclose Gender  .048 .420 Step 2: Interactions Nondisplay x gender Nondisclosure x gender ~p<10; *p<05; **p<.01; *** p<.001  0.94 .02 .18 -.12  .883 .228 .319  1.67-  .082  -.95  .137  1.69  PSP in Group Psychotherapy/page 148 Table 34: Tests of the mediational model. R  R change  F change  Beta  BECK DEPRESSION INVENTORY Model: nondisclosure of imperfection ^ post-session anxiety Step 1: .200 .200 13.49** Pre-treatment depression Step 2: .204 .004 0.30 Pre-session anxiety (late sessions) Step 3: .234 .030 2.04 Post-session anxiety (late sessions) Step 4: .257 .023 1.55 Nondisclosure  l  decrease in depression .45** .001  ess  .07  .587  .21  .159  .17  .218  SKIN CONDUCTANCE LEVELS (Elevation during the stress test) less decrease in SCL reactivity Model: nondisclosure of imperfection post-session anxiety' Stepl: .061 .061 .25.081 3.17SCL elevation pre-treatment .848 0.04 Step 2: .061 .001 .03 Pre-session anxiety (late sessions) .33* .043 Step 3: .141 .079 4.34* Post-session anxiety (late sessions) .20 .175 Step 4: .175 .034 1.89 Nondisclosure FEMALE GROUP MEMBERS ONLY: Model: nondisclosure of imperfectio, less self-disclosure over time—•/ess decrease in SCL reactivity 160 6.49* .40* .016 Step 1: .160 SCL elevation pre-treatment .989 ,00 .000 0.00 Step 2: .160 Self-disclosure (early session) -.40* .018 Step 3: .297 .136 6.21* Self-disclosure (late session) Step 4: .355 .104 2.81 .28 .058 Nondisclosure p<.10;  * p<.05; ** p<.01  ON  IP a,  I ^>  o -s;  TJ C  ai i_  +5 3 O (A O t O a> (A  Q.  >^  I  o  E ">5> "kl  o ra a> a> i_ 3 o <A  I I  lev  iscl  O "53  co  a>  in  c o  !E (A  o  rel  ra  CO  3  o u w  TJ  ra w  CNJ  0) J Z  c o z  /  C  /  lat  •—  in  I  0)  re C  E o «  o  py  o  £ 0) Q.  O  Q. C  U  «»G> C  on  etw  c  ort  noi  e JZ  nc  s  on  (A  I  in  ra  k. P> t_ 3 (A  O •o C a>  O 0  k. 3 D) Li.  O (A  O  I  "35  >  I o  CM  Adejaiu UJ a}E| sejnsopsia  m  CNJ  o  CO  m  CO  CS  o o  3- ° CD +£ C  c (0 CD o tf) CD C  O  •<= u  il 3  «  o  £ t  o (A  .i  I  o o  «) o  (A C  = e-°  o u  o  O O) tf>  c o *r  O  E  •E  c  a E  o *-  *#o ©  O o  3 (A O O W  o > ro o . Q. re  S  c o  .1 o a>  £  c  ©  £ .E </) ro 0} o </) c 22 2! = o tf> 0) "O •O H! CD  CO H3  3  °  u_  o  oo  CD  CM  O  Adejain u; a\e\ sajnsopsjp-uas aAjusod  60 CJ  o c  I  "35  o  o c  tf)  co  re  D <-> s  fe  <D  O c  <n o  .E ra Q. ° -  tf) ® C ._ .2  o  o O) >*.E re = Q.  4  2  TJ c C o u  o c C  4  Q)  E-> tf)< o re o a> c  TJ C  o CD <£>  0)  i_ 3  O)  LO  iq o  O  aunsseq oj ujnjaj -|3S  IT)  I -g o  o c "55 <A re  -s;  O *i (0 c  c  E re  . _ *•• +J  a> i  Li  c a> .2 c O o  €5 E  £  o £ g> o> •5 .E w = o o 0  i:  (A £  1 8 o *r c c •s 1 a> (A o c 0) (A O  I a. TJ C a> O 0)  O)  e m p s e q oj u-injaj -|og  PSP in Group Psychotherapy/page  157  Figure 9: Illustrations of mediational models for nondisclosure of imperfections predicting problems in group therapy process and negative treatment outcome. A). Post-session- anxiety late in therapy  .21  Nondisclosure of imperfections  w.  Post-treatment depression, , controlling for pre-treatment levels  (.22-) .17  B) Post-session anxiety late in therapy  .33*  .30*>  Nondisclosure of imperfections  w  Post-treatment SCL reactivity, controlling for pre-treatment levels  (.27*) .20  C) Women only: Self-disclosure late in therapy, controlling for earlier levels  Post-treatment SCL reactivity, controlling for pre-treatment levels  Nondisclosure of imperfections (.35*) .28  N B : A l l values are standardized regression coefficients (P's). Values in parentheses are P's for the unmediated model. In each case, these decreased when the mediator was added to the model. Therefore, the association between nondisclosure o f imperfection and negative therapy outcome can be partially explained by anxiety and a lack o f increase i n self-disclosure over the course o f therapy. ~ p<.10; * p<.05; ** p<.01  PSP in Group Psychotherapy/page  APPENDIX A:  Participant recruitment poster  158  PSP in Group Psychotherapy/page  CO  *  &o  CD CO  £  p. o sr cr P cr < CD CO ^  CD  CD  o  CD  *< o ^ o g.  CD  CD CD CD >-t  < < CD a* o »-+>  CD  co  2- CTQ § »—t pi. cr S> n O p  P  § o  CO CO CD  P-  CO  O  CD X  co  CD  i—*•  ftj  P  >t>  X  CD O  O  3 o o  r-t-  FT  CO  p *** £ ©  »  8P M I S  ?I$ S£• e  "3  2  B  »  2  S  M  a; | S 1 ^ ft  S ©  1  T 3  n ©  as  3  B  i  5  BJ S L S . x — © !—! ? p" I , Ci 2 © sr _  e 5«  of 2. S* g g se v© ^ * 2 2 ©" * © 0 0  * a, a S' » «s  cr 52 © cr » »  c r  D  as  e  ft  e S » S » ?2.« '< c 3  2*° ® ® 3 a 2 2 s 2. B " » © . g r^S.^ © a -i © w vo i  ^ »  B  3  {•Si  ft  O o ft  << O  o  a  CO  CTQ  &  CD  O  C*  CD  §  »-*•  o.  CO  Q  CD O  OP  Ct> CD  •o  d  CD  r+  CD  * cr * C f Q cr CD o 13 o O O o g CD <->  a o o D  jjjjjjjl  159  PSP in Group Psychotherapy/page  160  APPENDIX B Study Measures PERFECTIONISTIC SELF-PRESENTATION SCALE Listed below are a group of statements. Please rate your agreement with each of the statements using the following scale. If you strongly agree, circle 7; if you strongly disagree, circle 1; if you feel somewhere in between, circle any one of the numbers between 1 and 7. If you feel neutral or undecided the midpoint is 4. 1 2 Disagree Strongly  3  4 5 Neutral  6  7 Agree Strongly  1. It is okay to show others that I am not perfect 1 2. I judge myself based on the mistakes I make in front of other people 1 3. I will do almost anything to cover up a mistake 1 4. Errors are much worse if they are made in public rather than in private 1 5. I try always to present a picture of perfection 1 6. I would be awful if I made a fool of myself in front of others.... 1 7. If I seem perfect, others will see me more positively 1 8. I brood over mistakes that I have made in front of others 1 9. I never let others know how hard I work on things 1 10. I would like to appear more competent than I really am 1 11. It doesn't matter if there is a flaw in my looks 1 12. I do not want people to see me do something unless I am very good at it 1 13. I should always keep my problems to myself 1 14. I should solve my own problems rather than admit them to others 1 15. I must appear to be in control of my actions at all times 1 16. It is okay to admit mistakes to others 1 17. It is important to act perfectly in social situations 1 18. I don't really care about being perfectly groomed 1 19. Admitting failure to others is the worst possible thing 1 20. I hate to make errors in public 1 21. I try to keep my faults to myself 1 22. I do not care about making mistakes in public 1 23. I need to be seen as perfectly capable in everything I do 1 24. Failing at something is awful if other people know about it 1 25. It is very important that I always appear to be "on top of things" 1 26. I must always appear to be perfect 1 27. I strive to look perfect to others 1  2 3 4 5 6 7 2 3 4 5 6 7 2 3 4 5 6 7 2 2 2 2 2 2 2 2  3 3 3 3 3 3 3 3  4 4 4 4 4 4 4 4  5 5 5 5 5 5 5 5  6 7 6 7 6 7 6 7 6 7 6 7 6 7 6 7  2 3 4 5 6 7 2 3 4 5 6 7 2 2 2 2 2 2 2 2 2 2 2  3 3 3 3 3 3 3 3 3 3 3  4 4 4 4 4 4 4 4 4 4 4  5 5 5 5 5 5 5 5 5 5 5  6 7 6 7 6 7 6 7 6 7 6 7 6 7 6 7 6 7 6 7 6.7  2 3 4 5 6 7 2 3 4 5 6 7 2 3 4 5 6 7  PSP in Group Psychotherapy/page  161  MULTI-DIMENSIONAL PERFECTIONISM SCALE Listed below are a number of statements concerning personal characteristics and traits. Read each item and decide whether you agree or disagree and to what extent. If you strongly agree, circle 7; if you strongly disagree, circle 1; if you feel somewhere in between, circle any one of the numbers between 1 and 7. If you feel neutral or undecided the midpoint is 4. 1 2 Disagree Strongly  3  4 5 Neutral  6  7 Agree Strongly  1. When I am working on something, I cannot relax until it is perfect 1 2 2. I am not likely to criticize someone for giving up too easily. 1 2 3. It is not important that the people I am close to are successful... 1 2 4. I seldom criticize my friends for accepting second best 1 2 5. I find it difficult to meet others'expectations of me 1 2 6. One of my goals is to be perfect in everything I do 1 2 7. Everything that others do must be of top-notch quality 1 2 8. I never aim for perfection in my work 1 2 9. Those around me readily accept that I can make mistakes too.... 1 2 10. It doesn't matter when someone close to me does not do their absolute best 1 2 11. The better I do, the better I am expected to do 1 2 12.1 seldom feel the need to be perfect 1 2 13. Anything I do that is less than excellent will be seen as poor work by those around me 1 2 14.1 strive to be as perfect as I can be 1 2 15. It is very important that I am perfect in everything I attempt 1 2 16.1 have high expectations for the people who are important to me 1 2 17.1 strive to be the best at everything I do 1 2 18. The people around me expect me to succeed at everything I I do 1 2 19.1 do not have very high standards for those around me 1 2 20.1 demand nothing less than perfection of myself 1 2 21. Others will like me even if I don't excel at everything 1 2 22.1 can't be bothered with people who won't strive to better themselves 1 2 23. It makes me uneasy to see an error in my work 1 2 24.1 do not expect a lot from my friends 1 2 25. Success means that I must work even harder to please others 1 2 26. If I ask someone to do something, I expect it to be done flawlessly 1 2 27.1 cannot stand to see people close to me make mistakes 1 2 28.1 am perfectionistic in setting my goals 1 2 29. The people who matter to me should never let me down 1 2 30. Others think I am okay, even when I do not succeed 1 2 31.1 feel that people are too demanding of me 1 2  3 3 3 3 3 3 3 3 3  4 4 4 4 4 4 4 4 4  5 5 5 5 5 5 5 5 5  6 6 6 6 6 6 6 6 6  7 7 7 7 7 7 7 7 7  3 4 5 6 7 3 4 5 6 7 3 4 5 6 7 3 4 5 6 7 3 4 5 6 7 3 4 5 6 7 3 4 5 6 7 3 4 5 6 7 3 3 3 3  4 4 4 4  5 5 5 5  6 7 6 7 6 7 6 7  3 3 3 3  4 4 4 4  5 5 5 5  6 6 6 6  3 3 3 3 3 3  4 4 4 4 4 4  5 5 5 5 5 5  6 7 6 7 6 7 6 7 6 7 6 7  7 7 7 7  PSP in Group Psychotherapy/page  32.1 must work to my full potential at all times 33. Although they may not show it, other people get very upset with me when I slip up 34.1 do not have to be the best at whatever I am doing 35. My family expects me to be perfect 36.1 do not have very high goals for myself 37. My parents rarely expected me to excel in all aspects of my life 38.1 respect people who are average 39. People expect nothing less than perfection from me 40.1 set very high standards for myself 41. People expect more from me than I am capable of giving 42.1 must always be successful at school or work 43. It does not matter to me when a close friend does not try their hardest 44. People around me think I am still competent even if I make a mistake 45.1 seldom expect others to excel at whatever they do  162  1 2 3 4 5 6 7 1 1 1 1  2 2 2 2  3 3 3 3  4 4 4 4  5 5 5 5  6 6 6 6  7 7 7 7  1 1 1 1 1 1  2 2 2 2 2 2  3 3 3 3 3 3  4 4 4 4 4 4  5 5 5 5 5 5  6 6 6 6 6 6  7 7 7 7 7 7  1 2 3 4 5 6 7 1 2 3 4 5 6 7 1 2 3 4 5 6 7  Copyright (©) Paul L. Hewitt, Ph.D., & Gordon L. Flett, Ph.D., 1988  MOOD  S C A L E S  Please indicate how strongly you are feeling each of these emotions right now by placing a tick at the appropriate place on each line. Sad/Depressed  not at all  extremely Tense/Anxious  not at all  extremely Frustrated/Angry  not at all  extremely  PSP in Group Psychotherapy/page  163  DISRUPTIVE THERAPY BEHAVIOUR CHECKLIST  Session Date: No show with excuse  No show without excuse  Late arrival  Incomplete homework  Subject  #  PSYCHOTHERAPY SELF-DISCLOSURE CODING SYSTEM Hewitt, Flynn, & Low, G. (1999) Self-Disclosure: Defined as a self-reference that tells us something about who the speaker is (e.g., "I am not a verbal person, I'm a visual person"), how the speaker relates to others (e.g., "I feel angry with my mother, but I love her and miss her"), or the affect (emotions) of the speaker (e.g., "I feel badly"). Each self-disclosure may contain several statements related to the same theme of affect (e.g., "I'm kind of anxious. My stomach is quaking and my hands are shaking."), core beliefs (e.g., "I feel others expect too much of me."), or self-concept descriptors (e.g., "I'm an open person."). Self-disclosures may be found in statements of different weight or directness. For example, selfdisclosures may be: Direct, first person references, independently initiated, which go to core beliefs, self-concept or affect: "I feel guilty." "I feel others expect too much of me." "I find it difficult to speak out." "It makes me feel anxious." "My life is in complete flux." "I tend to be excessively private." "I don't have an identity of my own." "I would wake up in the middle of the night, crying." "I tend to get panicky." "That sets up the scenario that they disappoint me."  PSP in Group Psychotherapy/page  164  Indirect statements which still go to core beliefs, self-concept or affect, or responsive agreements: "I feel the same way." "One day, you wake up and feel resentful." "I agree." "It opened a comfort level for me." "It's so frustrating." "It came out of the pain and fear at the time." "It's a big deal by the time you set up this boundary." "I went down this wonderful path of anorexia." "There's a big sense of loss of things wanted that I didn't get." "It's hard to let people know my inner side." Vague self-references, using "you" in general or "possibility" terms such as "might" or "maybe", disclosures involving greater distancing, a more abstract way of thinking or not owning the feeling or belief: "Maybe, I'm asking too much." "You feel as though you're being pressured." (said in the general sense) "You have to take responsibility for your own life." "I seem to have, perhaps, set certain expectations of myself." "I sort of felt left out." "It's really hard to deal with the expectations of other people." References to states of knowledge are not considered to be self-disclosures. For example: "I find that interesting." "It's important." "I understand what you're saying." Descriptions of other people without reference to the self, and plans also are not considered to be self-disclosures. For example: "I'm trying to overcome that." "My father was very demanding." "My parents are very perfectionistic." Valence: Refers to the tone of the content of each self-disclosure. Positive:  "I'm really looking forward to it." "It opened a comfort level for me." Negative: "One day, you wake up and feel resentful." "You feel as though you are being pressured." Note: these disclosures may reflect something therapeutic, but still have negative content E.g., "I am angry with you." "That makes me sad." Neutral: "You have to take responsibility for your own life." "I'm not a verbal person, I'm a visual person."  PSP in Group Psychotherapy/page  165  Here and Now Focus: Also coded for each speech segment rather than each self-disclosure. Is the person self-disclosing in relation to people or events that happen within the group, outside the group, or more abstractly. Codes: In-group, Out-group, Both, Neither.  Depth of Revealing: How difficult is it for the person to reveal this information? Consider: content, how often this person has disclosed this information in the past, how much importance this person places on it. This is coded for each speech segment (i.e., the entire period in which the person speaks or, if briefly interrupted, continues with the same theme or focus) rather than each statement or individual self-disclosure. See video anchored examples 1 - marginally personal 2 - moderately personal 3 - very personal  Affect: Coded at the end of viewing the entire tape Defined as: an overall rating of expressed mood: Anxiety:  a state of being uneasy, apprehensive, or worried about what may happen; concern about a possible future event. Depression: low spirits, gloominess, dejection. An emotional condition, characterized by feelings of hopelessness and inadequacy. Hostility: a feeling of enmity, ill will, unfriendliness, antagonism. Happiness: having, showing or causing a feeling of great pleasure, contentment, joy.  PSP in Group Psychotherapy/page  166  THERAPY PROGRESS RATING - PATIENT Please answer the following questions about your experiences in this therapy program by circling the appropriate number on each scale. 1. In general, how satisfied were you with this group treatment program?  1 not at all satisfied  2  3  4  5  6  7 completely satisfied  6  7 benefited significantly  2. How much do you think you benefitedfromthis program?  1 did not benefit  2  THERAPY PROGRESS RATING - THERAPIST Please rate how much each member benefitedfromgroup therapy on the following scale: 1 = not at all to 7 = very much Subject Number or Name  Benefit Rating  THERAPY PROGRESS RATING - OTHER  In general, how much do you think your friend or family member benefited from this program?  1 did not benefit  2  7 benefited significantly  PSP in Group Psychotherapy/page 167  DAVICON IQ QUESTIONS Question set 1: 1. Which of the following is different? Bay, Christmas, Water, Pear 2. Solve the following problem: 5 + 3-3*4 3. Spell the name of the capital city of Hungary 4. Rearrange the following letters to form a word: GSRU A Question set 2: 5. Which of the following is different? Knot, Eye, Tooth, Grain 6. Solve the following problem: 12-4 + 8/2 7. Name the capital city of New Zealand 8. Rearrange the following letters to form a word: OLDME  PSP in Group Psychotherapy/page  168  INTERACTION ANXIOUSNESS SCALE Please indicate the extent to which each statement is characteristic or true of you, on a scale from 1 to 5: 1 2 3 4 5 not at all slightly moderately very extremely characteristic of you characteristic of you 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15.  I often feel nervous even in casual get-togethers I usually feel uncomfortable when I am in a group of people I don't know I am usually at ease when speaking to a member of the opposite sex I get nervous when I must talk to a teacher or boss Parties often make me feel anxious and uncomfortable I am probably less shy in social interactions than most people I sometimes feel tense when talking to people of my own sex if I don't know them very well I would be nervous if I were being interviewed for a job I wish I had more confidence in social interactions I seldom feel anxious in social situations In general, I am a shy person I often feel nervous when talking to an attractive member of the opposite sex I often feel nervous when calling someone I don't know very well on the telephone I usually get nervous when I speak to someone in a position of authority... I usually feel relaxed around other people, even people who are quite differentfromme  1 1 1 1 1 1  2 2 2 2 2 2  3 3 3 3 3 3  4 4 4 4 4 4  5 5 5 5 5 5  1 1 1 1 1  2 2 2 2 2  3 3 3 3 3  4 4 4 4 4  5 5 5 5 5  1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5  PSP in Group Psychotherapy/page 169  APPENDIX C:  Sobel's (1982) Test of the Significance of the Indirect Effect.  a = standardized regression coefficient of the independent variable predicting the mediator b = standardized regression coefficient of the mediator predicting the dependent variable s = standard error of a a  s = standard error of b b  ab t= SQRT(b s + a s + 2  2  a  2  2  b  s v) a  NB: In SPSS, the standard error of the standardized regression coefficient can be obtained by adding the syntax "SES" to the standard regression statistics.  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0090569/manifest

Comment

Related Items