Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Perfectionism, social exclusion, and anorexia nervosa symptoms McGee, Brandy Jennifer 2007

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

Item Metadata

Download

Media
831-ubc_2007-267592.pdf [ 7.24MB ]
Metadata
JSON: 831-1.0100538.json
JSON-LD: 831-1.0100538-ld.json
RDF/XML (Pretty): 831-1.0100538-rdf.xml
RDF/JSON: 831-1.0100538-rdf.json
Turtle: 831-1.0100538-turtle.txt
N-Triples: 831-1.0100538-rdf-ntriples.txt
Original Record: 831-1.0100538-source.json
Full Text
831-1.0100538-fulltext.txt
Citation
831-1.0100538.ris

Full Text

P E R F E C T I O N I S M , S O C I A L E X C L U S I O N , A N D A N O R E X I A N E R V O S A S Y M P T O M S by B R A N D Y J E N N I F E R M C G E E B . S c , The University of Alberta, 1998 M . A . , The University of British Columbia, 2003 A THESIS S U B M I T T E D I N P A R T I A L F U L F I L L M E N T O F T H E R E Q U I R E M E N T F O R T H E D E G R E E O F D O C T O R OF P H I L O S O P H Y in T H E F A C U L T Y OF G R A D U A T E S T U D I E S (Psychology) T H E U N I V E R S I T Y OF B R I T I S H C O L U M B I A Apr i l 2007 © Brandy Jennifer McGee, 2007 11 Abstract Anorexia nervosa affects between 0.5% and 1%> o f the population and kills young women at a rate 10 times higher than the normal population ( D S M - I V - T R ; American Psychiatric Association, 2003; Hoek, 2006). Perfectionism has long been recognized as a risk factor for the disorder, but that relationship is not well understood. We conducted a longitudinal experimental study to examine the effect o f multidimensional perfectionism on state symptoms of anorexia nervosa under conditions of social acceptance, social exclusion, and a control condition. In addition to testing a diathesis-stress model of anorectic symptoms, we also explored whether rejected affect mediated that relationship. The concurrent results indicated that both trait perfectionism and perfectionistic self-presentation were predictive of affective, cognitive, and self-evaluative symptoms of anorexia. However, only perfectionistic self-promotion and nondisplay of imperfection predicted concurrent symptoms beyond the variance accounted for by the B i g Five personality traits. The longitudinal findings suggested that perfectionistic self-presentation predicted a worsening of cognitive and self-evaluative anorectic symptoms over time. B y itself, neither trait perfectionism, nor perfectionistic self-presentation was predictive of concurrent eating behavior or change in eating behavior over time. Tests of the diathesis-stress model indicated that both trait perfectionism and perfectionistic self-presentation interacted with level of belongingness to predict change in state symptoms of anorexia nervosa. However, the different dimensions o f perfectionism were predictive o f different aspects o f the anorectic experience. For example, socially prescribed perfectionism interacted with social feedback condition to predict increasing dietary restriction, whereas self-oriented perfectionism interacted with social feedback condition to predict change in state self-esteem, and perfectionistic self-promotion interacted with level o f belongingness to predict a worsening o f anorectic thoughts. Moreover, with one exception, social acceptance had a negative effect for highly perfectionistic individuals. Tests of the mediated moderation model revealed that rejected affect mediates the relationship between perfectionistic self-promotion and anorectic thoughts about rigid weight regulation in the context of social exclusion. The results are discussed with respect to existing models of anorectic symptom development and implications for treatment and future research. Table of Contents Abstract i i Table of Contents iv List of Tables x List of Figures x i i Acknowledgements x i i i Dedication xiv I N T R O D U C T I O N 1 What is Anorexia Nervosa? 2 Perfectionism and Anorexia Nervosa Symptoms 5 Theoretical Understanding 5 Empirical Evidence 7 Resolving Inconsistencies in the Relationship between Perfectionism and Anorexia Nervosa Symptoms 9 Multidimensional Perfectionism and Anorexia Nervosa 9 Trait Perfectionism and Anorexia Nervosa Symptoms 13 Perfectionistic Self-Presentation and Anorexia Nervosa Symptoms 16 A Diathesis-Stress Model 20 Perfectionism and Social Exclusion 21 Social Exclusion and Anorexia Nervosa Symptoms 23 Multidimensional Perfectionism, Social Exclusion, and Anorexia Nervosa Symptoms 25 Mechanism o f Act ion 30 Mediated Moderation Model 30 V Overview of Methods 32 Hypotheses 33 Time 1 Analyses 34 Time 2 Analyses 34 Mediated Moderation Model 35 M E T H O D S 36 Participants 36 Materials 37 Predictor Variables 37 Multidimensional Perfectionism Scale 37 Perfectionistic Self-Presentation Scale 38 B i g Five Inventory 38 Criterion Variables • • - 39 Profile of M o o d States - Short Form 39 Mizes Anorectic Cognitions Questionnaire - Revised 40 State Self-Esteem Scale 41 Screening Measures 41 Beck Anxiety Inventory 41 Beck Depression Inventory - Second Edition 42 Procedure 42 Overview of Procedure 42 Session 1 43 Session 2 45 VI R E S U L T S 48 Associations Among the Personality and Anorexia Nervosa Symptom Variables at Time 1 49 Zero-Order Correlations 49 Eating Behavior 49 Depressed M o o d 49 Anorectic Cognitions 50 Self-Control 50 Rigid Weight Regulation 50 Weight and Approval 50 State Self-Esteem 51 Social Self-Esteem 51 Appearance Self-Esteem 51 Unique Contributions of Perfectionism 51 Eating Behavior 52 Depressed M o o d 52 Anorectic Cognitions 52 State Self-Esteem 52 Beyond the B i g Five • 53 Eating Behavior 54 Depressed M o o d 54 Anorectic Cognitions 54 State Self-Esteem 55 Vll Associations among Perfectionism, Belongingness, and Time 2 Anorexia Nervosa Symptom Variables 56 Restricting the Sample 56 Dummy Coding 59 Manipulation Check 59 Ma in Effects of Social Feedback 60 Main Effects of Perfectionism 60 Moderation Effects: Testing the Diathesis-Stress Model 62 Eating Behavior 62 Depressed M o o d 63 Anorectic Cognitions 64 Self-Control 64 Rig id Weight Regulation 64 Weight and Approval 65 State Self-Esteem 66 Social Self-Esteem 66 Appearance Self-Esteem 67 Mediated Moderation 67 D I S C U S S I O N 70 Concurrent Relationships 70 Trait Perfectionism 70 Perfectionistic Self-Presentation 73 Unique Contributions of Perfectionism 75 vi i i Trait Perfectionism 75 Perfectionistic Self-Presentation 76 Predicting Change in Anorexia Nervosa Symptoms 78 M a i n Effects 78 Social Feedback Condition 78 Perfectionism 79 Trait Perfectionism 79 Perfectionistic Self-Presentation 79 Diathesis-Stress Mode l 81 Eating Behavior 81 Trait Perfectionism 81 Perfectionistic Self-Presentation 83 Depressed M o o d 83 Trait Perfectionism 83 Perfectionistic Self-Presentation 84 Anorectic Cognitions 84 Trait Perfectionism 84 Perfectionistic Self-Presentation 86 State Self-Esteem 87 Trait Perfectionism 87 Perfectionistic Self-Presentation 89 The Role of Rejected Affect 90 Conclusions about Perfectionism and Anorectic Symptoms 92 ix Strengths and Limitations of the Current Study 94 F O O T N O T E S 163 R E F E R E N C E S 165 A P P E N D I C E S 189 Appendix A 189 Appendix B 190 Appendix C 191 X Lis t of Tables Table 1 Means, standard deviations, and alpha reliability of the time 1 and time 2 variables 97 Table 2 Zero-order correlations of the time 1 personality and anorexia nervosa symptom variables 99 Table 3 Summary of regression analysis for perfectionism dimensions predicting unique variance in time 1 anorexia nervosa symptoms 104 Table 4 Summary of hierarchical regression analysis for perfectionism predicting unique variance in time 1 anorexia nervosa symptoms, controlling for big five personality traits 108 Table 5 Summary of hierarchical regression analyses for social feedback condition predicting time 2 anorexia nervosa symptoms, controlling for time 1 anorexia nervosa symptoms 114 Table 6 Summary of hierarchical regression analyses for perfectionism, social feedback condition and perfectionism x social feedback condition predicting time 2 quantity of food eaten, after controlling for time 1 quantity of food eaten 118 Table 7 Summary of hierarchical regression analyses for perfectionism, social feedback condition and perfectionism x social feedback condition predicting Time 2 P O M S - S F depressed mood, after controlling for time 1 P O M S - S F depressed mood 123 Table 8 Summary of hierarchical regression analyses for perfectionism, social feedback condition and perfectionism x social feedback condition predicting time 2 M A C -R self-control subscale, after controlling for time 1 M A C - R self-control XI subscale 128 Table 9 Summary of hierarchical regression analyses for perfectionism, social feedback condition and perfectionism x social feedback condition predicting time 2 M A C -R rigid weight regulation subscale, after controlling for time 1 M A C - R rigid weight regulation subscale 132 Table 10 Summary of hierarchical regression analyses for perfectionism, social feedback condition and perfectionism x social feedback condition predicting time 2 M A C -R weight and approval subscale, after controlling for time 1 M A C - R weight and approval subscale 137 Table 11 Summary o f hierarchical regression analyses for perfectionism, social feedback condition and perfectionism x social feedback condition predicting time SSES social subscale, after controlling for time 1 SSES social subscale 142 Table 12 Summary of hierarchical regression analyses for perfectionism, social feedback condition and perfectionism x social feedback condition predicting time 2 SSES appearance subscale, after controlling for time 1 SSES appearance subscale.... 146 Table 13 Simple slope regression analyses of significant perfectionism x social feedback interactions predicting time 2 anorexia nervosa symptoms, controlling for time 1 anorexia nervosa symptoms 150 Xll L i s t of Figures Figure 1 Proposed mediated moderation model 153 Figure 2 The interaction of socially prescribed perfectionism with social feedback to predict time 2 amount of food eaten 154 Figure 3 The interaction of nondisplay of imperfection with social feedback to predict square root of time 2 P O M S - S F depressed mood 155 Figure 4 The interaction of perfectionistic self-promotion with social feedback to predict time 2 M A C - R rigid weight regulation 156 Figure 5 The interaction of socially prescribed perfectionism with social feedback to predict time 2 M A C - R weight and approval 157 Figure 6 The interaction of nondisclosure of imperfection with social feedback to predict time 2 M A C - R weight and approval 158 Figure 7 The interaction o f self-oriented perfectionism with social feedback to predict time 2 SSES social self-esteem 159 Figure 8 The interaction of self-oriented perfectionism with social feedback to predict time 2 SSES appearance self-esteem 160 Figure 9 Mediated moderation model: The interaction o f perfectionistic self-promotion with social feedback to predict time 2 M A C - R rigid weight regulation, mediated by rejected affect 161 Figure 10 The interaction of perfectionistic self-promotion with social feedback to predict time 2 rejected affect 162 x m Acknowledgements Above all , I would like to thank my supervisor, Dr. Paul Hewitt, for being a wonderful guide in this process, for challenging me, and for encouraging me. I would also like to sincerely thank my supervisory committee members, Dr. Mark Schaller and Dr. Anita Delongis, for helping to shape and enhance this project and for ensuring that it would be one that I would be able to finish while I still had my youth. This project was supported by doctoral training awards from the Social Sciences and Humanities Research Council , the Michael Smith Foundation for Health Research, and the British Columbia Medical Services Foundation. I am very grateful for their support. To my friends and statistical consultants, Dr. M i k e Papsdorf and Kev in Will iams, thank you for spending countless hours helping me to make sense of my print outs, for being patient with my continuous stream of questions and taking the time to consult your own references when you weren't sure, and mostly for holding my hand through this process. I also owe a debt of gratitude to the volunteers that worked with me on the project—Jasmin Abizadeh, Sara Assadian, A l i a Azab, Al iana Boden, Makenzie Chilton, Carol L i n , and Andrea Penney—thank you for your tireless efforts and insightful feedback, and, most of all , for your humor and enthusiasm. I offer a deep and heartfelt thank you to my closest friends, Nicole Dorfan and Carla Seipp. Y o u kept me going when things were tough, you listened to my anxieties and frustrations and shared your own, and you never wavered in your support or your confidence in me. M y life is enriched by your presence. A n d , finally, to my parents, thank you for seeing me through all the years of school, for standing behind me and encouraging me, for celebrating with me, for always taking my side, and for loving me no matter what. I love you both. xiv Dedication For my partner, Jaimie. I have told you many times how much your support has meant to me, but let me say it again here, publicly. Y o u have helped me with my literature searches, critiqued my work, listened when I needed to talk, been honest with me when I have overreacted, made sure that I always had lots of healthy food to keep me going, stood up for me and stood by me, been my computer guru, and reminded me of how fortunate I am to be doing this at all. Y o u are everything to me. This achievement is as much yours as it is my own. 1 I N T R O D U C T I O N A perfectionistic young woman enrolls in a summer class. The class is small and, knowing no one, she quickly identifies a small group of people with whom she seems to have common interests and hopes to establish a friendship. However, during a lunch break she inadvertently overhears the group talking about her. They complain to one another about how annoying she is, mock her mannerisms and style of dress, and discuss plans to avoid her for the remainder of the course. This rejecting experience sparks a cascade of negative thoughts and feelings. So aversive is this experience that the woman copes by focusing on her appearance, meticulously counting calories, and severely restricting her diet. This scenario illustrates a process that we believe is important to the development of anorexia nervosa. It rests on the idea that the need to belong is one o f the most basic human drives and that attainment of social approval and avoidance of social exclusion propels much of our behavior (Baumeister & Leary, 1995). Indeed there is evidence to suggest that when this drive is thwarted, people respond in a variety of negative, and often self-defeating, ways (Twenge, Catanese, & Baumeister, 2002). However, whereas experiences o f social exclusion are universal, only a small proportion of the population develops symptoms of anorexia nervosa. Thus, it would seem that some people are more vulnerable to rejection than others. We propose that perfectionism embodies the attempt to secure social acceptance and avoid social exclusion, and that highly perfectionistic individuals are, therefore, at greater risk for anorectic symptoms from social exclusion experiences. There are at least two avenues by which the volatile combination of perfectionism and social exclusion may lead to anorexia symptoms. On one hand, physical attractiveness is a means of obtaining social approval (Gilbert, 1997), and dieting may be an active attempt to favorably 2 affect the opinions of the group. Alternatively, social exclusion is known to produce aversive self-awareness (Twenge, Catanese, & Baumeister, 2003), and theory suggests that dietary restriction may ameliorate such painful states (Heatherton & Polivy, 1992). To investigate these relationships, we conducted a study in which female university students were randomly assigned to receive feedback indicating social acceptance or social exclusion, or to receive no feedback at all (control). We measured perfectionism and state anorexia nervosa symptoms in session 1 and then measured state anorexia nervosa symptoms again one week later in session 2, immediately following the administration of the experimental manipulation. We hypothesized that, in the context of social exclusion, highly perfectionistic women would experience state cognitive, affective, and behavioral symptoms consistent with anorexia nervosa (AN) . Moreover, we predicted that this relationship would be mediated by rejected affect. We sought to add to existing research in five ways. First, we used multidimensional measures of perfectionism to clarify the differential relations of the separate dimensions to anorectic symptoms. Second, we examined a diathesis-stress model of A N development, using social exclusion as a novel putative moderator. Third, we tested whether rejected affect mediated the diathesis-stress model. Fourth, we used state measures for our dependent variables to assess immediate fluctuations in anorectic symptoms, rather than chronic changes. And , fifth, we expanded the traditional range of anorectic symptoms beyond eating behavior and anorectic cognitions, and included both depressed mood and self-esteem. W h a t is Anorex ia Nervosa? Anorexia nervosa is a disorder that affects between 0.5% and 1%> of the population, most commonly girls and women beginning in mid- to late adolescence ( D S M - I V - T R ; American Psychiatric Association, 2003). It is characterized by a constellation of affective, physiological, 3 cognitive, and behavioral symptoms. A t present, critical diagnostic symptoms include a refusal to maintain a minimally normal body weight, intense fear of gaining weight, disturbance in the way one's body is perceived, undue influence of body weight or shape on self-evaluation, or denial of the seriousness of the abnormally low weight, and, for postmenarcheal women, amenorrhea ( D S M - I V - T R ; American Psychiatric Association, 2003). However, symptoms of the disorder extend beyond the current D S M - I V - T R definition, and some argue that existing diagnostic criteria do not adequately capture the cardinal symptoms of A N (Hebebrand, Casper, Treasure, & Schweiger, 2004). For example, in addition to the D S M - I V - T R criteria, other diagnostic schemes emphasize the importance of depressed mood and low self-esteem in the internal anorectic experience (Psychodynamic Diagnostic Manual; Silver Spring, MD:Al l iance of Psychoanalytic Organizations; 2006). Research and clinical observation suggest that anorectic patients exhibit a variety of cognitive and behavioral symptoms. For example, they may engage in deliberate dietary restriction, consume food in a ritualized manner (e.g., unusual food combinations, cutting food into small pieces, limited variety in food selections), and be persistently preoccupied with food, as evidenced by ruminative calorie counting, mental imaging of food, and eager interest in food preparation for others (Bruch, 1978; Rothenberg, 1990). In addition, cognitive behavioral conceptualizations of A N have emphasized dysfunctional cognitions as a central feature of the disorder (Garner & Bemis, 1982). Anorectic patients suffer from cognitive disturbances such as all-or-nothing thinking and extreme fear of weight gain (Mizes, Christiano, Madison, Post, Seime, & Varnado, 2000), and show signs of cognitive impairments that are unrelated to depressed mood (McDowel l , Moser, Ferneyhough, Bowers, Andersen, & Paulsen, 2003). Thus, we included both cognitive and behavioral symptom measures as outcome variables. In addition to commonly studied anorectic attitudes and behaviors, there are other related constructs that are closely linked to the presence of eating disturbance, most notably depressed mood and self-esteem. For instance, numerous researchers have reported high rates of comorbidity between A N and major depressive disorder (Braun, Sunday, & Halmi , 1994; Laessle, Ki t t l , Fitcher, & Wittchen, 1987). Others have suggested that it may be more than simply comorbidity, that there is substantial overlap between A N and depression, including shared genetic factors (Wade, Bul ik , Neale, & Kendler, 2000), common serotonergic abnormalities (Kaye, Gendall, & Strober, 1998), and higher lifetime prevalence rates for both major depressive disorder and dysthymia in women with A N compared to control women (Pearlstein, 2002). Moreover, there may be an interactive relationship between depressed mood and other anorectic symptoms. For instance, Corte and Stein (2004) found that negative mood predicted disordered eating behavior in a clinical sample of anorectic women, and other research revealed that low perceived dieting success predicts an increase in depression in university women over a 4-week diet trial (Oates-Johnson & Clark, 2004). Similarly, participating in a treatment group designed to help women stop dieting leads to a decrease in depression (Polivy & Herman, 1992). Given high levels of comorbidity (Braun et al., 1994), questions about the degree to which A N and depression are variants of the same underlying psychopathology (Kaye et al., 1998; Wade et al., 2000), and evidence for the immediate effects of negative mood on A N symptoms and vice versa, we chose to include depressed mood as an outcome variable in the present study. Self-esteem has also been closely linked to anorectic symptomatology. In fact, some have described A N as a "disorder of the sense of self-esteem and self-worth" (Sassaroli & Ruggiero, 2005, p. 139), whereas others have noted that core characteristics of anorexia are negative self-evaluation (Vitousek & Hollon, 1990) and undue influence of body weight or shape on self-evaluation ( D S M - I V - T R ; American Psychiatric Association, 2003). Empirically, women with A N tend to have lower self-esteem than their non-eating disordered counterparts (Shisslak, Pazda, & Crago, 1990; Wagner, Halmi , & Maguire, 1987) and, among college students, self-esteem is negatively correlated with severity of eating difficulties (Mintz & Betz, 1988). However, recent work suggests that low self-esteem is only associated with thoughts about dietary restriction in the presence of stress (Sassaroli & Ruggiero, 2005). Therefore, in addition to anorectic cognitions, eating behavior, and depressed mood, we chose to include state self-esteem as a dependent variable in this study. We focused on state variables because we were interested in the immediate effects of social exclusion in highly perfectionistic women, rather than on enduring traits or chronic mood. In addition, self-esteem has been conceptualized in both unidimensional (e.g., Rosenberg, 1965) and multidimensional terms (e.g., Heatherton & Polivy, 1991). For our purposes, we chose to examine rapid shifts in self-esteem in both appearance and social domains, domains that are closely related to A N , with its consummate focus on appearance and high need for social approval (Bruch, 1973; P D M , Silver Spring, M D : Alliance of Psychoanalytic Organizations, 2006). Perfectionism and Anorexia Nervosa Symptoms Theoretical Understanding Perfectionism is widely viewed as a clinical feature of A N (see Bruch, 1978) and theory suggests that perfectionism may play an etiological role in the disorder (e.g., Slade, 1982). In her seminal work, Bruch (1973) commented on a pattern that was ubiquitous in the premorbid histories of her anorectic patients. Prior to the onset of the disorder, these girls were described as compliant, submissive, and popular individuals who excelled in all pursuits, most notably 6 academics and athletics. Bruch (1988) hypothesized that extreme dietary restriction was actually a late symptom in the overall development of the disorder and that, paradoxically, the "superperfectionistic" behavior about which parents spoke with so much pride was often an early indicator of psychological problems. Others have also noted that anorectic patients were described by their parents as having perfectionistic personalities prior to illness onset (Halmi, Goldberg, Eckert, Casper, & Davis, 1977), and that childhood perfectionism is one of only two factors that distinguished A N patients from healthy controls (Fairburn, Cooper, D o l l , & Welch, 1999). From a psychodynamic perspective, Bruch (1973, 1978) ascribed such perfectionistic drives to a deep and pervasive sense of ineffectiveness arising from problems in early child development, particularly problems with separation and individuation. Slade's (1982) functional analysis of A N provides a different theoretical model and suggests that perfectionism interacts with general dissatisfaction with life and the self to produce a need to control and achieve in some aspect of life. If dieting is triggered (e.g., by comments from peers or cultural pressure), then the desire for control may become focused on dieting behavior. Dieting behavior is, in turn, positively reinforced by feelings of success and negatively reinforced via fear of weight gain and avoidance of other stressors. This creates a downward spiral of dieting and weight loss that eventually leads to A N . Heatherton and Baumeister's (1991) escape from self-awareness model may also provide some insight into the relationship between perfectionism and A N . Although the model was developed to explain binge behavior more typical of bulimia nervosa, it may also be relevant to the binge-purge subtype of A N . The model posits that individuals with high (i.e., perfectionistic) standards who fall short of those expectations w i l l enter a process of negative self-awareness. 7 This unpleasant state of self-awareness generates an escape response in which cognitive focus is narrowed, resulting in disinhibition and binge eating. Empi r i c a l Evidence Numerous studies have tied perfectionism to A N in cross-sectional samples ranging from currently i l l and recovered anorectic patients (e.g., Strober, 1980; Sullivan, Bulik, Fear, & Pickering, 1998) to university students (e.g., Davis, Claridge, & Fox, 2000, Pearson & Gleaves, 2006), and other work has indicated that perfectionism is predictive of more profoundly disturbed behaviors among anorectic patients (Halmi et al., 2000). Although fewer studies have examined the relationship longitudinally, those that have done so prospectively have demonstrated that perfectionism is a risk factor for A N development (Tyrka, Waldron, Graber, & Brooks-Gunn, 2002), and that perfectionism is transmitted independently of eating disorders and is among the most potent of vulnerability factors (Lilenfeld et al., 2000). More recent work has clarified the link between perfectionism and specific eating disorder symptoms. Forbush, Heatherton, and Keel (2007) examined the relationship of perfectionism (as measured by the Eating Disorders Inventory Perfectionism subscale) to an array of eating disordered behaviors including fasting, bingeing, and purging behaviors. They found that, amongst women, perfectionism showed the strongest association with fasting and purging behaviors, and that the association between perfectionism and binge eating was mediated by fasting. Amongst men, only fasting was significantly associated with perfectionism. Together these results suggest that perfectionism is closely tied to dietary restriction, and that the apparent relationship between perfectionism and binge eating is due to an underlying third variable, fasting. Unfortunately, because the authors used the perfectionism subscale of the Eating Disorders Inventory, which is known to be a composite of two distinct dimensions of trait perfectionism (Bardone, Vohs, 8 Abramson, Heatherton, & Joiner, 2000; Joiner, Heatherton, Rudd, & Schmidt, 1997; Joiner & Schmidt, 1995, Sherry, Hewitt, Besser, McGee, & Flett, 2004), it is not possible to know whether only one or both of these dimensions is involved in the relationships demonstrated by Forbush and colleagues. A s with the more traditional measures of disordered eating, depressed mood and self-esteem are also correlated with perfectionism. For example, perfectionism has been associated with depression in both student (Enns, Cox, Sareen, Freeman, 2001; Flett, Hewitt, Blankstein, & O'Brien, 1991; Minarik & Ahrens, 1996) and psychiatric samples (Hewitt & Flett, 1991a), and with postpartum depression in a population-based sample of women (Mazzeo et al., 2006). Further, recent longitudinal research on a clinical sample demonstrated that perfectionism predicted an increase in depressive symptoms over time, after controlling for initial level of depression and neuroticism (Dunkley, Sanislow, Gri lo, & McGlashan, 2006). Finally, perfectionism has been implicated in diathesis-stress models of depression (e.g., Dunkley, Blankstein, Halsall, Wil l iams, & Winkworth, 2000; Hewitt & Flett, 1993a; Hewitt, Flett, & Ediger, 1996) and research supports a mediational role for perfectionism in the relationship between harsh parenting and depression (Enns, Cox, & Clara, 2002). Self-esteem has also been associated with perfectionism. Researchers have demonstrated that perfectionism is negatively associated with trait self-esteem in university students (Ashby & Rice, 2002; Pearson & Gleaves, 2006). Moreover, there appears to be a developmental path in which maladaptive perfectionism mediates the relationship between perceived parental psychological control and low global self-esteem (Soenens, Vansteenkiste, Luyten, Duriez, & Goossens, 2005). 9 Resolving Inconsistencies in the Relationship between Perfectionism and Anorexia Nervosa Symptoms Although recent reviews suggest that the weight of the evidence points strongly to a predispositional role for perfectionism in eating disorder development (Lilenfeld, Wonderlich, Riso, Crosby, & Mitchel l , 2006), inconsistencies remain in the perfectionism and A N literature. For example, not all studies support a link between perfectionism and anorectic symptoms (Calam & Waller, 1998) and others suggest that perfectionism persists following recovery from A N (Bastiani, Rao, Weltzin, & Kaye, 1995). There are at least two possible explanations for this. First, perfectionism is a multidimensional construct (Hewitt & Flett, 1991b) and the use of various unidimensional scales has obscured the differential relations between A N and the different perfectionism dimensions and, thereby, reduced our ability to find meaningful predictive relationships. A n d , second, perfectionism may act as a general risk factor that predicts A N only under certain (e.g., stressful) conditions (McGee, Hewitt, Sherry, Parkin, Flett, 2005; Sassaroli & Ruggiero, 2005; Slade, 1982). We w i l l address each possibility in turn. Multidimensional Perfectionism and Anorexia Nervosa Symptoms In contrast to unidimensional conceptualizations of perfectionism, which tend to focus only on self-directed perfectionistic cognitions or attitudes, multidimensional perfectionism encompasses important motivational and interpersonal aspects as well as perfectionistic beliefs (Flett, Hewitt, Blankstein, & Pickering, 1998; Hewitt & Flett, 1991b; Hewitt et al., 2003). Weak, null, or inconsistent findings for the relationship between perfectionism and A N symptoms (e.g., Calam & Waller, 1998) may be clarified by using multidimensional measures of perfectionism so that we might explore the how each distinct dimension relates to eating problems. For example, multidimensional work has revealed that some, but not al l , o f the perfectionism 10 dimensions are predictive of A N symptoms (e.g., Hewitt, Flett, & Ediger, 1995; Minarik & Ahrens, 1996). More recently, Pearson and Gleaves (2006) conducted a confirmatory factor analysis of perfectionism using several perfectionism scales, and examined the relationship between the resulting factors and eating disorder features. Their results revealed that each factor had a different relationship with eating disturbance, suggesting that perfectionism must be assessed multidimensionally. Thus, unidimensional conceptualizations of perfectionism may conceal the true relationship between perfectionism and A N symptoms or provide inconsistent results. This problem is underscored by studies showing that the commonly used Eating Disorders Inventory Perfectionism subscale (EDI-P; Garner, Olmstead, & Polivy, 1983) is actually a composite of two distinct dimensions of perfectionism (Bardone et al., 2000; Joiner et al., 1997; Joiner & Schmidt, 1995, Sherry et al., 2004). Despite this, the EDI -P is usually treated as a unidimensional measure of perfectionism, which may obscure the differential relations of its perfectionism components to eating pathology. Perfectionism has been conceptualized in a number o f different ways, but in the past decade there has been increasing interest in multidimensional models of perfectionism (Frost, Marten, Lahart, & Rosenblate, 1990; Hewitt & Flett, 1991b; Hewitt et al., 2003). From Hewitt and Flett's perspective, there are two key components: trait perfectionism and perfectionistic self-presentation. Both components are understood as stable personality features that involve cognitive, motivational, and interpersonal aspects. However, whereas trait perfectionism entails ; stable and consistent need to be perfect, perfectionistic self-presentation involves a desire to appear to be perfect in the eyes of others. Trait perfectionism consists of three dimensions, only two of which are germane to the prediction of A N symptoms (Hewitt & Flett, 1991b). Self-oriented perfectionism (SOP) entails a strong drive for oneself to be perfect, and may involve 11 unrealistic self-expectations and a tendency to censure one's own behavior and focus on one's own flaws. In contrast, socially prescribed perfectionism (SPP) reflects the individual's belief that others (e.g., parents, society) hold unrealistically high standards for one's own behavior and that these others w i l l only be satisfied when such standards are met. The self-oriented and socially prescribed dimensions, though related, are fundamentally different with respect to motivation, controllability, and coping (Hewitt & Flett, 1991b, Flett, Russo, & Hewitt, 1994). Self-oriented perfectionists maintain an inward, personal focus consistent with intrinsic motivation and, because their standards and goals may be changed in a proactive manner, they have greater perceived control (Hewitt & Flett, 1991b) and more constructive thoughts about action oriented behavioral coping (Flett et al., 1994). However, their harsh self-evaluations are indicative of a style of emotional coping consistent with reduced self-acceptance (Flett et al., 1994). Self-oriented perfectionism is related to conscientiousness (Hi l l , Mclntire, & Bacharachi, 1997), and individuals high on this dimension are particularly sensitive to achievement failures (Enns & Cox, 2005; Hewitt & Flett, 1993a). In contrast, socially prescribed perfectionism is associated with an outward focus, diminished intrinsic motivation, and higher levels of extrinsic motivation (Hewitt & Flett, 1991b) consistent with an intense interpersonal sensitivity and fear of negative evaluation (Flett, Hewitt, & De Rosa, 1996). Because they typically take an interpersonal focus, such persons are thought to be particularly vulnerable to interpersonal stressors (e.g., Hewitt & Flett, 1993a). Moreover, because socially prescribed perfectionists derive their goals from the perceived expectations of others, they lack control over these standards (Hewitt & Flett, 1991b). They have a markedly negative coping style characterized by a tendency to overgeneralize negative outcomes and to ruminate about unpleasant events, an absence of behavioral coping, and a form of superstitious thinking by 12 which individuals believe that good outcomes wi l l necessarily be followed by bad events (Flett et al., 1994). Finally, unlike self-oriented perfectionism, socially prescribed perfectionism is associated with neuroticism rather than conscientiousness (Dunkley, Blankstein, & Flett, 1997, H i l l etal., 1997). Perfectionistic self-presentation, or the desire to appear perfect to others, is understood as an entrenched and deceptive interpersonal style that has as its goal the presentation of a flawless facade (Hewitt et al., 2003). Like trait perfectionism, it is comprised of three distinct facets. Perfectionistic self-promotion (PSP) involves actively asserting one's strengths or achievements. For example, boasting about a high grade on an exam would be consistent with this facet. Conversely, nondisplay of imperfection (NDP) and nondisclosure of imperfection (NDC) are defensive styles geared toward concealing imperfections. However, whereas nondisplay of imperfection entails a desire to avoid behavioral displays of imperfection, like tripping in public, nondisclosure of imperfection involves a reluctance to verbally admit perceived flaws, such as confiding to a friend that one was passed over for a promotion. The facets of perfectionistic self-presentation, though related, differ from each other and are differentially related to outcomes. For example, whereas nondisplay of imperfection is consistently associated with decrements in self-esteem beyond the effects of trait perfectionism, perfectionistic self-promotion is associated with higher levels of self-esteem after controlling for the variance associated with trait perfectionism (Hewitt et al., 2003). However, it may be that the elevated self-esteem that is associated with perfectionistic self-promotion reflects narcissitic tendencies (Sorotzkin, 1985) rather than more adaptive and healthy levels of self-esteem. Indeed, of the three self-presentational facets, only perfectionistic self-promotion was significantly correlated with narcissism (Hewitt et al., 2003). In addition, recent work revealed that both 13 perfectionistic self-promotion and nondisplay of imperfection were related to thoughts about having cosmetic surgery performed, though the authors suggested that different motivations underlie the desire for cosmetic surgery in each of these groups (Sherry, Hewitt, Lee-Baggley, Flett, & Besser, 2004). Perfectionistic self-promoters were thought to be motivated by a desire to garner admiration or attention, whereas individuals high on nondisplay of imperfection were thought to be inspired to cover up or rectify physical flaws associated with aging. Although the self-presentational styles are related to the trait dimensions, they are distinct aspects of perfectionism and are predictive of different maladaptive outcomes (Hewitt et al. 2003). Perfectionistic self-presentation is the interpersonal expression of perfectionistic tendencies. Although it is possible that a person high on one of the trait dimensions may also exhibit elevated levels of perfectionistic self-presentation, it is just as possible that the person would be low on perfectionistic self-presentation (Hewitt et al., 2003). For example, imagine two adolescent girls, each of whom feels the burden of extremely high perceived parental expectations (i.e., they have high levels of socially prescribed perfectionism). One girl might respond by creating a public image of perfection. She may tell others only of her successes or she may avoid situations in which she fears her inadequacies w i l l be revealed. In contrast, the other girl might respond to the perceived pressure with anger and resentment. She may clash with her parents and perhaps even flaunt her displeasure by dressing or behaving in ways that elicit her parents' disapproval. The latter girl is low on perfectionistic self-presentation, while the former, with her high need for approval and public self-consciousness, is high on perfectionistic self-presentation. Trait Perfectionism and Anorexia Nervosa Symptoms Clinical observations of A N patients suggest the presence of both self-oriented and 14 socially prescribed perfectionism. For example, anorectics are characterized by rigid and harsh self-evaluations, and Bruch (1988) noted that although their unrealistic aspirations may have originated with perceived parental expectations, they are later internalized as the patient's own goals. In the words of one such patient: "I feel that I can't live on just an ordinary scale of human endeavour. I feel that I have to make this world better and do as much as a human being is capable of doing. What I have to achieve is something that absolutely squeezes the last drop out of me, otherwise I haven't given enough." (Bruch, 1978, p. 53) The commonly reported tendency for eating disordered women to view achievements in all-or-nothing terms may indicate the presence of self-oriented perfectionism (Bruch, 1988). Self-oriented perfectionism engenders rigid and unrealistically high self-imposed standards that extend to all areas of functioning, including appearance and eating behavior (Geller, Cockell , Hewitt, Goldner, & Flett, 2000; Hewitt et al., 1995). However, in addition to the need to be perfect for oneself, A N patients also believe that others expect them to be perfect. They are plagued by a need to outguess others and to do what they think others expect them to do (Bruch, 1978). B y their own descriptions, they are "constantly concerned with being found wanting, not being good enough, not l iving up to 'expectations', in danger of losing their parents love and consideration" (Bruch, 1978, p. 39). Thus, based on clinical vignettes, both self-oriented and socially prescribed perfectionism are evident in women who struggle with A N . Consistent with this, evidence from empirical study supports a link between trait perfectionism and anorectic pathology. Bastiani and her colleagues (1995) showed that restricting subtype anorectic subjects are perfectionistic and that this perfectionism persists even after weight restoration. Moreover, their results suggested that this perfectionism is generally 15 experienced as self-imposed (i.e., SOP) and not as a response to others' expectations (i.e., SPP). In addition, self-oriented perfectionism was found to be a risk factor associated with disordered eating in adolescent females (McVey , Pepler, Davis, Flett, & Abdole l l , 2002). However, in a more stringent test of these relationships, Cockell and colleagues (2002) found that after controlling for self-esteem, depression, and overall psychiatric severity, women with A N had substantially higher levels of both self-oriented and socially prescribed perfectionism than did a matched sample of women with mood disorders. Further work on the relationship of trait perfectionism to anorexia nervosa in both clinical and non-clinical populations has confirmed the relationship of both SOP and SPP to anorectic symptomatology (e.g., Geller et al., 2000; Hewitt et al., 1995; Pliner & Haddock, 1996). Trait perfectionism has also been linked to both depression and low self-esteem. Self-oriented perfectionism and socially prescribed perfectionism are consistently linked to depression in both psychiatric and student samples (Chang & Sanna, 2001; Flett, Besser, Davis, & Hewitt, 2003, Hewitt et al., 1996). For example, Flett, Hewitt, Blankstein, and Mosher (1995) demonstrated that both self-oriented and socially prescribed perfectionism predicted depression at Time 1, but that only self-oriented perfectionism predicted an increase in depression three months later. Moreover, a number of studies have supported a diathesis-stress conceptualization of perfectionism and depression. Hewitt and Flett (1993a) demonstrated that the relationship between socially prescribed perfectionism and depression was moderated by stress in both psychiatric and student samples, and Chang and Rand (2000) showed that socially prescribed perfectionism interacted with perceived stress to predict distress symptoms one month later. Similarly, socially prescribed perfectionism interacted with life stress to predict an increase in depressive symptoms in medical students over a 5 month period (Enns, Cox, & Clara, 2005). In 16 contrast, in a longitudinal study of patients Hewitt and colleagues (1996) found that socially prescribed perfectionism predicted an increase in depression over a 4 month period, but that it did not interact with either achievement or interpersonal stressors to predict such changes. Self-oriented perfectionism has also been implicated in moderational models with a number of studies showing that self-oriented perfectionism interacts with stress to predict increases in depression (Enns & Cox, 2005; Enns et a l , 2005). Work by Hewitt and colleagues with depressed patients has demonstrated that self-oriented perfectionism interacts with achievement stress to predict both concurrent depressive symptoms (Hewitt & Flett, 1993a) as well as an increase in depressive symptoms over time (Hewitt, et al., 1996). Trait perfectionism has also been linked to self-esteem. Past work has indicated that socially prescribed perfectionism is negatively associated with general, academic, appearance, and social self-esteem (Flett et al., 1991; Flett et a l , 1996; Hewitt et al., 2003), and with a fragile or unstable self-concept (Flett et al., 1991). More recent research revealed that, whereas socially prescribed perfectionism was associated with low global self-esteem, self-oriented perfectionism was not (Klibert, Langhinrichsen-Rohling, Saito, 2005). After controlling for socially prescribed perfectionism, self-oriented perfectionism was positively correlated with self-esteem. Perfectionistic Self-Presentation and Anorexia Nervosa Symptoms In addition to trait perfectionism, evidence also suggests that perfectionistic self-presentation plays a role in A N . Clinical observation has implicated all three self-presentation dimensions—perfectionistic self-promotion, nondisplay of imperfection, and nondisclosure of imperfection—in A N symptomatology. For example, Bruch (1978, 1988) noted that anorectics feel safe from blame and criticism only when they can maintain the image o f perfection in the eyes of others, and that they have been praised for presenting this facade from an early age. 17 Consistent with a high level of perfectionistic self-promotion, some withdraw behind a "mask of superiority" (p. 146) whenever they experience self-doubt or encounter disagreement (Bruch, 1978), and may even use their appearance as a means to convey an image o f perfection to others. Others, however, adopt a more defensive, concealing approach consistent with nondisplay of imperfection or nondisclosure of imperfection. For such a patient, all her efforts are directed towards "hiding the fatal flaw of her fundamental inadequacy" (Bruch, 1988, p. 6). Perfectionistic self-presentation may evolve from the familial environment of individuals with anorexia (Humphrey, 1992). In a family that presents a public image of perfection while masking underlying problems, promoting an image of perfection or hiding perceived faults reflects conformity to established familial norms. Convergence in levels of perfectionism may be a factor with other reference groups as well . Meyer and Waller (2001) found that social proximity promotes convergence in levels of perfectionism in college roommates who were randomly assigned to housing groups. Alternatively, theory suggests that perfectionistic self-presentation may arise from a disturbed identity development (Bruch, 1973). Weinrich, Doherty, and Harris (1985) compared the identity development of female anorectic and bulimic patients with female patients with other psychiatric disorders and with normal controls. They found that anorectic and bulimic patients exhibited significantly lower self-evaluations than the other two groups. Moreover, anorectics' current self-evaluations were lower than their past self-evaluations, suggesting an "anti-developmental" trend. These women may attempt to construct a less fragile self-concept by focusing on their physical appearance (Striegel-Moore, Silberstein, & Rodin, 1993). They may concentrate on their public image to compensate for a weakened identity. 18 There is also empirical support that ties perfectionistic self-presentation to A N symptoms. For example, using anorectic patients and psychiatric controls, Cockel l et al. (2002) demonstrated that anorectic subjects had higher scores on nondisclosure of imperfection than did other psychiatric patients, suggesting that anorectic patients are more reluctant to admit their imperfections than are other psychiatric patients. Additional work on women with A N revealed that all three self-presentational perfectionism facets (perfectionistic self-promotion, nondisplay of imperfection, and nondisclosure of imperfection) were associated with the anorectics' tendency to suppress negative feelings and to give priority to the feelings of others (Geller et al., 2000). Research using female university students demonstrated that the self-presentational facets of perfectionism were related to anorectic symptoms as well as body image avoidance and self-esteem (Hewitt, Flett, & Ediger, 1995). Finally, recent work revealed that all three self-presentational facets predicted anorectic symptoms in undergraduate women who were dissatisfied with their appearance, but not in women who were satisfied with the way they look (McGee et al., 2005). A s with trait perfectionism, perfectionistic self-presentation has also been associated with depression and low self-esteem. There are few studies on the relationship between perfectionistic self-presentation and depression, but available data indicates that all three dimensions are positively correlated with depressive symptoms (Cockell et al., 2002). Moreover, both nondisplay of imperfection and nondisclosure of imperfection predict depression severity scores in both students and psychiatric patients, even after controlling for levels of trait perfectionism and other facets of personality (Hewitt et a l , 2003). In addition, perfectionistic self-presentation has been associated with self-esteem. For example, nondisplay of imperfection and nondisclosure of imperfection were negatively 19 correlated with general, appearance, academic, and social self-esteem (Hewitt et al., 2003), although after controlling for levels of trait perfectionism only nondisclosure of imperfection was uniquely predictive of lower social self-esteem. The relationship between perfectionistic self-promotion and self-esteem appears to be more complex. Although zero-order correlations revealed a negative association with self-esteem, once the variance associated with trait perfectionism was removed, perfectionistic self-promotion uniquely predicted increased general, appearance, and social self-esteem. Thus, whereas most dimensions of perfectionism are associated with lower levels of self-esteem, perfectionistic self-promotion, and to a lesser degree self-oriented perfectionism, may be associated with a somewhat more positive self-concept, though only after controlling for all other dimensions of perfectionism. Overall, there is abundant evidence that both trait perfectionism and perfectionistic self-presentation are associated with anorexia nervosa symptoms. However, recent reviews have also suggested that, in addition to perfectionism, A N symptoms are associated with other personality variables, notably neuroticism (Cassin & von Ranson, 2005). Further, prospective studies indicated that neuroticism was a strong predictor of A N development in a population based sample of twins (Bulik, Sullivan, Tozzi , Furberg, Lichtenstein, & Pedersen, 2006). This begs the question: Does perfectionism offer incremental explanatory power in the prediction of A N symptoms beyond that provided by the higher order trait neuroticism? This is an important question because both trait and perfectionistic self-presentation dimensions are correlated with neuroticism (Enns et al., 2005, Hewitt et al., 2003, H i l l et al., 1997). Although this concern has not yet been addressed in the A N literature, it has been examined with respect to other outcome variables. For example, Haring, Hewitt, and Flett (2003) found that socially prescribed perfectionism predicted maladaptive marital coping and poorer marital adjustment for both the 20 self and the partner, even after controlling for neuroticism. In addition, other work has shown that nondisclosure of imperfection is associated with depression after controlling for the B i g Five traits (Hewitt et al., 2003). However, not all studies have confirmed the incremental predictive validity of perfectionism. For example, in a sample of suicidal adolescent inpatients, neither self-oriented nor socially prescribed perfectionism significantly predicted posttreatment depression, hopelessness, or suicidal ideation after controlling for level of neuroticism (Enns, Cox, & Inayatulla, 2003). Therefore, to enhance our understanding of the role that personality plays in shaping A N , we examined whether perfectionism could offer anything unique, beyond the B i g Five traits, in the prediction of A N symptoms. A Diathesis-Stress M o d e l In addition to the multidimensional assessment of perfectionism, inconsistencies in the literature, such as high levels of perfectionism in recovered anorectics (Bastiani et al, 1995), could also be explained by the existence of moderator variables that change the strength or direction of the perfectionism-AN relationship. Connan and Treasure (1998) noted that stress negatively affects eating in both animal and human models and may trigger A N in vulnerable persons, persons with maladaptive coping skills such as emotional coping or reluctance to seek help, persons such as perfectionists (Ey, Henning, & Shaw, 2000; Hewitt & Flett, 2002). Consistent with this, many have observed that individuals who recover from A N are more likely to exhibit perfectionism than controls (Bastiani et al., 1995; Pla & Toro, 1999; Srinivasagam, Kaye, Plotnicov, Greeno, Weltzin, & Rao, 1995; Stein et al., 2002; Sutandar-Pinnock, Woodside, Carter, Olmsted, & Kaplan, 2003), suggesting that perfectionism may act as a continuing vulnerability that prompts eating disturbance only under certain conditions. In this vein, researchers have delineated a variety of diathesis-stress models for A N in which perfectionism 21 serves as an underlying trait vulnerability that interacts with environmental stress (Sassaroli & Ruggiero, 2005; Slade, 1982) or intrapersonal factors such as body dissatisfaction (McGee et al., 2005) to predict anorectic symptoms. Thus, it appears that whereas perfectionism is a risk factor for disordered eating, its negative potential may only be realized in the presence of stress or negative life events. So what kind of experiences might precipitate A N symptoms in highly perfectionistic individuals? We believe that social exclusion may be such an experience. Perfectionism and Social Exclus ion A number of studies suggest that there is reason to believe that perfectionists may be particularly susceptible to social exclusion. For example, early work revealed that socially prescribed perfectionism, but not self-oriented perfectionism, was associated with fear of negative evaluation in university students (Flett, et al, 1996; Hewitt & Flett, 1991b) and with dependent attitudes, such as need for admiration and desire to please others, in psychiatric patients and university students (Sherry, Hewitt, Flett, & Harvey, 2003). Thus, people who believe that others have perfectionistic demands for them fear a loss of social approval. Further, all three dimensions of perfectionistic self-presentation—perfectionstic self-promotion, nondisplay of imperfection, and nondisclosure of imperfection—are also positively correlated with fear o f negative evaluation and need for approval (Hewitt et al., 2003). Similarly, other studies demonstrated that perfectionism (as measured by the EDI) was associated with a need for approval (Belangee, Sherman, & Kern, 2003; Moulton, Moulton, & Roach, 1998). In addition, researchers have found that both trait perfectionism and perfectionistic self-presentation are positively correlated with public self-consciousness (i.e., the tendency to focus on outwardly observable aspects of the self) and social anxiety (Hewitt et al., 2003; Saboonchi & Lundh, 1997; 22 Saboonchi, Lundh, & Ost, 1999), and that perfectionism is predictive of greater symptom severity within socially phobic patient populations (Juster, Heimberg, Frost, & Holt, 1996). There is also evidence that both self-oriented and socially prescribed perfectionism are positively associated with sociotropy (commonly understood as the need for approval and fear of social rejection) in nonclinical samples (Bhar & Kyrios, 1999; Flett, Hewitt, Garshowitz, & Martin, 1997; Hewitt & Flett, 1993a). This suggests that individuals who have excessively high expectations for themselves, or who believe that others have unrealistic expectations for them, are plagued by a high need for approval and fear of social rejection. Moreover, the need for social approval is considered to be so central to the experience of those high on perfectionism, that a recently developed perfectionism inventory incorporated a 'need for approval' subscale (Hi l l , Huelsman, Furr, Kibler , Vicente, & Kennedy, 2004). Based on existing data, we can conclude that perfectionists have a heightened fear of social rejection and strong desire for social acceptance and, thus, are likely to be particularly susceptible to social exclusion. Recently, Hewitt, Flett, Sherry, and Caelian (2006) explored this issue in the context of suicide. They argued that perfectionistic individuals may founder in their quest for social approval and, consequently, experience real or imagined social disconnection, which leads to suicide behaviors. Although this is a mediational, rather than a moderational model, it highlights the interpersonal sensitivity of perfectionists, particularly socially prescribed perfectionists, and their strong need for approval. Further, Hewitt and Flett (2002) have noted that perfectionism has a complex relationship with stress, including interpersonal stress. They suggested that perfectionism influences stress generation and then interacts with the resulting stress to produce psychopathology. Thus, although perfectionists abhor social rejection, they may inadvertently create interpersonal disharmony by eschewing personal disclosures that would allow them to 23 connect with others (Geller et al., 2000; Hewitt et al., 2003) or by responding to disagreements with hostility or neglect (Flett, Hewitt, Shapiro, & Rayman, 2001; Haring et al., 2003; Hewitt & Flett, 1991b). Their hypersensitivity to social exclusion and poor coping skills may then, in turn, trigger psychological symptoms, such as those associated with A N . Social Exclusion and Anorex ia Nervosa Symptoms What is the relationship of social exclusion to A N ? The clinical literature indicates that anorectic patients harbor an intense fear of disapproval and rejection and a heightened concern about how they are viewed by others (Bruch, 1978, 1988). Moreover, it appears that fear of rejection is not an effect of the eating disorder itself, but is linked to eating disturbance in nonclinical populations as well . For example, Atlas (2004) found that sensitivity to criticism and rejection sensitivity predicted eating restraint in a sample of female undergraduates after controlling for body mass index. In addition, placing high importance on social acceptance predicted disordered eating in adolescent girls (McVey et al., 2002), and recent use of structural equation modelling revealed that fear of negative evaluation had both a direct and indirect effect on eating disorder symptoms in students (McClintock & Evans, 2001). Research has generally shown that poor peer acceptance is positively associated with weight concerns, including anorectic risk (Thomsen, M c C o y , Gustafson, & Will iams, 2002) and body dissatisfaction (Graham, Eich, Kephart, & Peterson, 2000), and that women with A N typically report low perceived social support (Tiller, Sloane, Schmidt, Troop, Power, & Treasure, 1997). Researchers have also explored this issue experimentally through the use of social rejection manipulations in the lab. Baumeister, DeWal l , Ciarocco, and Twenge (2005) demonstrated that undergraduate students who were told that no one else in their group wanted to work with them ate more cookies than other participants. However, this study did not consider 24 personality vulnerabilities of participants and how those predispositions could alter response to social exclusion. In a study by Rezek and Leary (1991), socially rejected women with a high drive for thinness (predisposition) ate less sweetened cereal and indicated that they planned to eat less at dinner that day, compared to socially accepted women with the same anorectic tendencies. The literature also indicates that social exclusion is associated with higher levels of depression and negative affect and with decrements in self-esteem. Across two studies, Buckley, Winkel , and Leary (2004) found that participants who were told that another participant did not want to work with them experienced greater negative affect and lower state self-esteem. Further, participants who were increasingly rejected (i.e., were initially accepted by a fictitious participant, but then increasingly rejected over time) generally felt worse than those who were constantly rejected. Others have noted that ostracism is associated with depression and suicide (Williams & Zadro, 2001) and that social exclusion typically leads to negative emotional experiences such as depression, loneliness, and feelings of isolation (Baumeister & Leary, 1995; Gardner, Pickett, & Brewer, 2000; Leary, 1990; Will iams, Cheung, & Choi , 2000). However, not all studies have found that social exclusion leads to negative affect (Baumeister et al., 2005), and some have suggested instead that rejection experiences prompt emotional numbing (Twenge et al., 2003). Although the bulk of the evidence indicates that social exclusion causes negative emotions, there are inconsistencies that warrant further attention. Both theory and research tie social exclusion to lower self-esteem (Baumeister & Leary, 1995; Leary, Schreindorfer, & Haupt, 1995; Leary, Tambor, Terdal, & Downs, 1995). Content analysis of recollections about being the target of ostracism indicated that those who felt ostracized for no discernible reason reported stronger threats to self-esteem than those who attributed the ostracism to a specific cause (Sommer, Will iams, Ciarocco, & Baumeister, 2001). 25 Moreover, recent work indicated that reductions in state self-esteem mediate the relationship between social ostracism and negative mood (Williams et al. 2000), suggesting a possible mechanism by which ostracism exerts its effects. Thus, not only are individuals with A N highly attuned to signs of potential social exclusion, but social rejection can itself spur eating disturbance as well as related mood and self-esteem difficulties. However, given that virtually everyone w i l l experience social rejection at some point, but only a small fraction of these wi l l go on to develop A N , it suggests that the risk is greatest for those with a personality predisposition, such as perfectionism. Multidimensional Perfectionism, Social Exclusion, and Anorexia Nervosa Symptoms Some of the most interesting questions about social exclusion and perfectionism pertain to the function that each is assumed to play from a developmental or evolutionary perspective. Although much of this work is speculative, it provides a theoretical backdrop for the hypotheses tested here. Baumeister and Leary (1995) have argued that group life was central to the survival of early humans. Social affiliation facilitated numerous activities from mating to defense to food acquisition that were adaptive for survival. Thus, it is likely that only those individuals who maintained social bonds were able live long enough to procreate. Given the importance of sustaining group membership, it is logical to think that a genetically driven mechanism may have evolved to avoid social exclusion. Leary and Downs (1995) hypothesized that self-esteem is such a mechanism. They suggested that self-esteem evolved to function as a sociometer calibrated to detect cues indicating social exclusion, and to alert the individual v ia negative affective reactions. In addition to theory, there is also empirical support for the idea that we are predisposed to avoid social exclusion. For example, Leary and colleagues (1995) found that behaviors or situations that are associated with social exclusion are also correlated with increases 26 in negative affect and decrements in self-esteem. Overall, theory and research suggest that we are inclined to avoid social exclusion and maintain public esteem, and that self-esteem may be a mechanism by which we monitor our progress toward that goal. While perfectionism has not been viewed from an evolutionary stance, it has certainly been conceptualized from a developmental perspective. Flett and colleagues (2002) have outlined four theoretical models for the development of perfectionism: the social expectations model, the social learning model, the social reaction model, and the anxious rearing model. Although each developmental model is unique, they are similar in their relationship to social approval. In each case, it is supposed that the child is motivated to maintain or acquire social approval (usually from the parents) and that it is this force that shapes her behavior and, ultimately, her self-concept. Perfectionism, then, is the means by which the child seeks to realize this drive in a less than ideal environment, one characterized by threat of social exclusion. In accord with this, Pacht (1984) suggested that perfectionists attempt to prove that they are loveable by setting excessively high goals. This is consistent with psychodynamic models that view perfectionism as a compensatory mechanism that serves to temporarily soothe feelings of rejection (Bruch, 1978). In addition, perfectionism has been strongly associated with indicators of insecure attachment, such as need for approval (e.g., Andersson & Perris, 2000), and with fear of negative evaluation (e.g., Hewitt et al., 2003), highlighting the perfectionist's drive to avoid social exclusion. Thus, perfectionism might be seen as a means by which to avoid (or attempt to avoid) social exclusion. In fact, understanding this need for acceptance and the resulting impact on interpersonal behaviors forms the basis of psychodynamic interpersonal treatment of perfectionism (Tasca, M i k a i l , & Hewitt, 2005). 27 Unfortunately, evidence suggests that perfectionism is not a particularly effective means of securing social approval. For instance, socially prescribed perfectionism is associated with marital dissatisfaction both in the self and in the partner (Habke, Hewitt, & Flett, 1999; Haring et al., 2003; Hewitt, Flett, & M i k a i l , 1995) and with more self-reported negative social interactions (Flett et al., 1997). In addition, it appears that others can detect the inauthentic front erected by perfectionistic self-presenters (Hewitt et al., 2003), and such behavior may be viewed negatively (Hewitt, Habke, Lee-Baggley, Sherry, & Flett, 2006). Most tellingly, socially prescribed perfectionism has been linked to the generation and maintenance of social disconnection (Blankstein & Dunkley, 2002; Flett et al., 1996; Hewitt, Flett, Sherry, & Caelian, 2006). Therefore, it appears that perfectionism may, in fact, exacerbate the problem of social exclusion, leading to a vicious circle of hypervigilance to signs of rejection and a resulting increase in perfectionistic efforts. How is this process related to eating behavior? In his theory of social attractiveness, Gilbert (1992; 1997) proposed that one of the most effective strategies by which humans gain status and secure social approval is to display qualities that others w i l l find attractive and to, thereby, draw positive attention. He refers to this as social attention holding power (SAHP) . One may increase positive social attention through a variety of different talents including, intellectual ability, artistic ski l l , altruism, and, most notable for our purposes, physical attractiveness. Precisely which attributes are considered valuable depends on the particular social and cultural context. In North American culture, studies have noted both an increasing focus on physical attractiveness as a sign of prestige and a shift toward a more unachievable ideal physique over the past forty years (Spitzer, Henderson, & Zivian, 1999). In fact, physical attractiveness has been described as a gifted status that confers social advantages throughout life (Cash, 1990), and 28 being thin is seen as a route to social acceptance (Leary, Tchividjian, & Kraxberger, 1994). Moreover, because physical appearance is one of the first points of social contact, it can have an enormous impact on social behavior. Thus, for a woman concerned with social approval, modifying her physical appearance may be a very valid means by which to increase her S A H P and secure social acceptance. Consistent with this, some have noted that the efforts of women with eating disorders to achieve a socially desirable appearance reflect their attempt to gain social approval (Button, 1983; Franco-Paredes et al., 2005; Striegel-Moore et al., 1993). Moreover, the belief that thinness w i l l have a positive impact on friendships predicted restrained eating in adolescents (Allen, Thombs, Mahoney, & Daniel, 1993; Gerner & Wilson, 2005). Unfortunately, the thin ideal that is currently in favor is unattainable for all but a very small percentage of the populace (Fallon, 1990). So, it is likely that attempts to mimic or achieve the appearance ideal w i l l be met with failure. In sum, we agree that human beings are predisposed to desire social approval and to avoid social exclusion. Moreover, we argue that perfectionism is a means by which individuals attempt to secure social approval in an environment characterized by threat of social exclusion. Given the importance of appearance in obtaining social approval in modern Western culture, it is unsurprising that a highly perfectionistic person might choose to focus her perfectionistic efforts in that domain. And , in light of current standards of beauty, those efforts are likely to focus on weight reduction. However, this explanation takes a macroscopic view of anorectic symptom development. We are also interested in examining these processes at a more immediate, microscopic level. H o w can we understand the effects of perfectionism and social exclusion on anorectic symptoms in the moment? 29 In their escape from self-awareness model, Heatherton and Baumeister (1991) suggested that in the context of perfectionistic standards, a perceived failure experience (such as social exclusion) would result in aversive self-perceptions accompanied by negative affect. They hypothesized that binge eaters would attempt to escape this aversive state by narrowing their cognitive field, focusing on the immediate stimulus environment to the exclusion of more broadly meaningful thought. One of the consequences of this type of low-level thought is to disengage normal inhibitions against overeating and permit binge behavior. In a series of experiments, Baumeister and his colleagues (2005) confirmed that social exclusion and social rejection produce decrements in self-regulation. For example, undergraduate students who were told that no one else in their group wanted to work with them ate more than other participants. Although this work suggests that binge eating is a consequence of cognitive deconstruction, other research has indicated that the binge actually facilitates cognitive deconstruction rather than resulting from it (Schupak-Neuberg & Nemeroff, 1993). That is, binge behaviors, such as repetitive chewing and swallowing, with a focus on the taste or texture of the food, may help to narrow the cognitive focus and exclude painful thoughts and feelings about the self (Schupak-Neuberg & Nemeroff, 1993). We believe that this same escape process may also explain restricting subtype anorectic behaviors such as ritualistic food preparation (e.g., cutting food into tiny pieces), compulsive calorie counting, or repeated weighing (Rothenberg, 1990), to the extent that such behaviors facilitate cognitive narrowing. Consistent with this, Heatherton and Polivy 's (1992) spiral model suggests that negative self-awareness triggers dieting, which, in turn, prompts a downward spiral of increasingly disordered eating. 30 Mechanism of Action In light of the impact o f social exclusion on mood (e.g., Buckley et al., 2004), researchers have investigated whether negative affect mediates the relationship between social exclusion and subsequent behavior. In general, the results have indicated that mood does not mediate these effects (Baumeister et al., 2005; Baumeister, Twenge, & Nuss, 2002; Wil l iams et al., 2000). However, these studies have typically used single-item measures of mood or brief general measures that tap overall positive and negative affect. Thus, the failure to find mediation effects may be due to a lack o f specificity regarding the type o f mood involved or to the use o f an unreliable or invalid measure. Moreover, no studies have yet examined the role of mood in a mediated moderational model of anorectic eating disturbance. Therefore, we chose to explore whether rejected affect mediated the link between perfectionism and anorectic behaviors and cognitions in the context of social exclusion. Mediated Moderation Model The model that we tested is one in which rejected affect mediates the relationship between perfectionism and subsequent A N symptoms in the context o f social rejection (Figure 1). We suggested that in the context of social exclusion, perfectionism—specifically the social dimensions of perfectionism (i.e., socially prescribed perfectionism, perfectionistic self-promotion, nondisplay o f imperfection, and nondisclosure o f imperfection)—would prompt feelings of rejection, which would in turn lead to A N symptoms. We expected that the dimensions of perfectionism with an interpersonal focus would make people particularly vulnerable to negative social feedback and that individuals high on these dimensions would be at greatest risk following our social exclusion manipulation. We also felt that certain perfectionism dimensions would be more likely to precipitate particular types of A N symptoms following 31 social exclusion. For example, given the strong relationship between socially prescribed perfectionism and depression (e.g., Enns et al., 2005; Hewitt et al., 1996) we would expect that individuals high on this dimension would be particularly l ikely to experience an increase in depressed mood in response to an interpersonal stressor, such as social exclusion. In contrast, the proactive and somewhat narcissistic style of a person high on perfectionistic self-promotion suggests that such a person would be more likely to experience elevations in prescriptive thoughts about how to achieve weight control. And following an experience of social rejection, the defensive interpersonal style of persons high on nondisclosure of imperfection and nondisplay of imperfection is likely to prompt an increase in ruminative thoughts about how weight and appearance affect social acceptance and what factors might be important in securing approval. This model is consistent with developmental and evolutionary theory on the importance of social affiliation and the consequences of social exclusion (Baumeister & Leary, 1995, Flett et al., 2002). Moreover, it fits with empirical evidence on the relationship between perfectionism and A N (e.g., Cockel l et al., 2002), perfectionism and fear of rejection (e.g., Hewitt et al., 2003; Sherry et al., 2003), and social exclusion and A N symptoms (e.g., Rezek & Leary, 1991). However, as it would not make sense to use rejected mood as a mediator in predicting concurrent depressed mood, and because other work has found that self-esteem mediates the link between social exclusion and negative mood (Williams et a l , 2000), we tested only the diathesis stress model for the depressed mood and state self-esteem outcome variables. The full mediated moderational model was tested for eating behavior and anorectic cognitions. 32 Overview of Methods In our study, female university students attended two sessions held one week apart. In session 1, they completed a questionnaire package and a behavioral eating assessment. In session 2, they were randomly assigned to receive personality feedback indicating social acceptance or social exclusion, or to receive no feedback at all (control). Then they were again asked to complete a series of measures and a behavioral eating assessment. Our predictor variables were perfectionism and social feedback condition. Our outcome variables were depressed mood, state self-esteem, quantity of food eaten, and anorectic cognitions. We measured demographic characteristics, B i g Five personality traits, perfectionism, and all outcome variables in session 1 and then measured the outcome variables again in session 2, immediately following the experimental manipulation. In light of ethical concerns regarding the impact of the social exclusion manipulation on vulnerable persons, participants were screened for high levels of anxiety and depression in session 1. Those participants with elevated anxiety or depression scores were screened out of the study in session 1 and did not return for session 2. A s noted earlier, we were particularly interested in more immediate, state symptom changes rather than chronic or enduring traits. Although both are important in understanding anorectic disturbance, the bulk o f the work to date has employed trait-like measures of eating pathology and neglected more immediate symptom changes. Moreover, the model that we proposed best fits with an episodic understanding of A N development. In this regard, it is akin to the escape from self-awareness model of binge eating (Heatherton & Baumeister, 1991), with an orientation towards momentary fluctuations in symptom levels occurring in predisposed individuals in response to stress. Therefore, we chose state measures of our affective, behavioral, and cognitive symptom variables. 33 Undergraduate women were specifically selected because eating problems are predominantly a female health concern (Heatherton, Nichols, Mahamedi, & Kee l , 1995), and because an estimated 64% of university women exhibit some degree of eating disordered behavior (Mintz & Betz, 1988). Moreover, many researchers view eating disorders as a continuum of symptom severity (Butler, Slade, & Newton, 1990; Stice, Ki l l en , Hayward, & Taylor, 1998), suggesting that female undergraduates represent a rich source of information on the relationship between personality and eating behavior. Based on work by Green (1991), 108 participants should ensure that the regression procedures we used would have power statistics of .80 for a medium effect size with six predictors. Data were analyzed in three broad groupings. First, Time 1 data were analyzed to provide a cross-sectional picture of the relationship of perfectionism to the outcome variables prior to any experimental manipulation, and to allow us to address secondary questions about the unique effects of perfectionism beyond other personality traits. This was both a replication and extension of previous work, allowing us to examine the role of other personality traits and the effects of perfectionism on state rather than trait A N symptoms. Second, we examined the effect of social feedback condition, perfectionism, and the perfectionism x social feedback interactions on Time 2 outcome variables, after controlling for the levels of the outcome variables at Time 1. Third, we investigated whether rejected affect mediated the diathesis-stress models for quantity of food eaten or anorectic cognitions. 34 Hypotheses Time 1 Analyses 1. Consistent with past findings, we expected to find that all o f the perfectionism dimensions would be significantly positively correlated with anorectic cognitions and depressed mood, and negatively associated with appearance and social self-esteem at Time 1. 2. Based on work by Forbush et al. (2007), we expected that perfectionism would be significantly negatively correlated with the amount of food consumed at Time 1 (i.e., that perfectionism would be associated with dietary restriction). 3. We expected that perfectionism would remain a significant predictor of Time 1 A N symptoms, after controlling for the effect of the B i g Five personality traits. Time 2 Analyses 1. , Based on past findings, we expected to find a main effect o f social feedback condition on change in depressed mood, state self-esteem, and eating behavior. Specifically, we expected that participants in the rejected group would experience a greater increase in depressed mood and anorectic cognitions, larger decrements in state self-esteem, and more food consumed than participants in either the accepted or control groups. 2. We also expected to find a main effect of perfectionism on change in the dependent variables. a) Amongst the trait dimensions, we hypothesized that socially prescribed perfectionism and self-oriented perfectionism would predict an increase in depressed mood and anorectic cognitions, and a decrease in state self-esteem and amount of food eaten over time. b) With respect to perfectionistic self-presentation, we anticipated that perfectionistic self-promotion, nondisplay o f imperfection and nondisclosure of imperfection would predict 35 an increase in depressed mood and anorectic cognitions, and a decrease in state self-esteem and amount of food eaten over time. 3. Based on existing literature, we hypothesized that the trait dimension of socially prescribed perfectionism and all three perfectionistic self-presentation facets (perfectionistic self-promotion, nondisplay of imperfection and nondisclosure of imperfection) would interact with social feedback condition to predict change in the dependent variables over time. We expected that highly perfectionistic women who were faced with social exclusion would experience greater increases in depressed mood and anorectic cognitions, and greater decreases in state self-esteem compared to perfectionistic women who were accepted or who did not receive any social feedback. Although social exclusion was expected to cause and increase in food consumption as a main effect, based on past work (Forbush et al., 2007; Rezek & Leary, 1991) we expected that in the context of high levels of perfectionism, it may lead to a decrease in food consumption. Mediated Moderation Model 1. We expected rejected affect to mediate the diathesis-stress models for anorectic cognitions and amount of food eaten. Although past work has indicated that mood does not mediate the effects of social exclusion (e.g., Baumeister et al., 2005; Wil l iams et al., 2000), we hypothesized that by using a mood measure that was more specific to rejected affect, we would find support for a mediated moderational model. 36 METHODS In order to explore these relationships we conducted an experimental study using university students in which perfectionism was measured and sense of belongingness manipulated. We examined the effect on mood, anorectic cognitions, state self-esteem, and eating behavior both before an after the experimental manipulation. Participants A sample of 149 female undergraduate students in psychology courses at the University of British Columbia participated in the study and provided basic demographic information. Participants were recruited from the undergraduate participant pool. In exchange for participating this study, each participant received a 2% bonus added to her final course grade. Participants ged 19.78 years of age (SD = 2.82) with 2.10 years of university education (SD = 1.06); 96% of the sample listed their relationship status as single or dating, while 2.7%> indicated that they were cohabitating and 1.3% reported that they were married. Thirty-five percent of the sample were in their first year of university; 32%> were in their second year; 22%> were in their third year; 8% were in their fourth year; and 3%> were in their fifth year of university education. Forty-one percent of participants reported their ethnic identity as Caucasian of European or North American descent; 52% as Asian; 4% as East Indian; 1.3% as Middle Eastern, 0.7% as African, and 0.7%> as Latino. The average number of years that participants in this sample had resided in Canada at the time of the study was 16.49 (SD = 6.26). This sample is comparable to other samples of university students recruited at the University of British Columbia (e.g., Sherry, Hewitt, Besser et al., 2004). in averat 37 Materials Predictor Variables Multidimensional Perfectionism Scale The Multidimensional Perfectionism Scale (MPS; Hewitt & Flett, 1991b) is a 45-item scale composed of three 15-item subscales designed to measure self-oriented perfectionism, other-oriented perfectionism, and socially prescribed perfectionism. However, as neither theory nor past research postulated an association between OOP and A N symptoms, only SPP and SOP were included in this study. Participants made 7-point ratings to reflect their agreement with statements, with higher scores indicating increased levels of trait perfectionism. Trait perfectionism dimensions are stable over time (e.g., Hewitt, Flett, Turnbull-Donovan, & Mika i l , 1991; Hewitt & Flett, 1991b) and have a high degree of internal consistency (e.g., Hewitt & Flett, 1991b). Additional studies have established incremental, predictive, convergent, and discriminant validity, as well as the multidimensionality of the M P S in psychiatric patients, community members, and university students (e.g., Hewitt et al., 1991; Hewitt & Flett, 1991a). For this study, the M P S was administered both as a self-report scale and also as an informant-report scale that was completed by a person close to the participant who responded as he or she felt the participant would respond. This was done to provide support for the validity of the self-report measure. In this study, the highest zero-order correlations between the self and informant-report versions of the M P S occurred between the corresponding dimensions of perfectionism (SOP: r = A10,p < .001; SPP: r = 253, p = .004), suggesting that the M P S self-report taps distinct aspects of perfectionism that are detectable by others. 38 Perfectionistic Self-Presentation Scale The Perfectionistic Self-Presentation Scale (PSPS; Hewitt et al., 2003) is a 27-item measure composed of three perfectionistic self-presentation subscales: perfectionistic self-promotion (10 items), nondisplay of imperfection (10 items), and nondisclosure of imperfection (7 items). Participants made 7-point ratings to reflect their agreement with statements, with higher scores indicating increased levels of perfectionistic self-presentation. The PSPS possesses good internal consistency, test-retest reliability, and adequate convergent and discriminant validity. Coefficient alpha values generally range between .78 and .86 for the three subscales, supporting their internal consistency, and test-retest reliability estimates range between .74 and .84, indicating a high level of stability in both clinical and student samples (Hewitt et al., 2003). Further work has documented the factorial stability, construct validity, convergent validity, and predictive validity of the PSPS in both psychiatric patients and university students (Habke et al., 1999; Hewitt, Flett, & M i k a i l , 1995; Hewitt et a l , 2003). A s with the M P S , the PSPS was administered both as a self-report scale and also as an informant-report scale that was completed by a person close to the participant who responded as he or she felt the participant would respond. The highest zero-order correlations between the self- and informant-reports occurred between the corresponding dimensions of perfectionistic self-presentation (PSP: r = .413,/? < .001; N D P : r = .407, p < .001, N D C : r = 315, p < .001), suggesting that the PSPS self-report measures distinct and outwardly detectable aspects of perfectionism. Big Five Inventory The B i g Five Inventory (BFI; John & Srivastava, 1999) is a 44-item measure consisting of short phrases designed to assess the B i g Five dimensions of personality. It generates five subscale scores corresponding to the five personality factors: Emotional Stability (reverse scored 39 Neuroticism), Extraversion, Openness to Experience, Agreeableness, and Conscientiousness. Items are rated on a 5-point scale from A (disagree strongly) to E (agree strongly), with higher scores indicating higher levels of the trait in question. Research has demonstrated that the BFI has good internal consistency (mean subscale coefficient alpha = .83), good convergent validity with other B i g Five scales, and test-retest reliabilities that range from .80 to .90 (John & Srivastava, 1999). Criterion Variables Profile of Mood States - Short Form To assess state levels of mood, participants completed the short form of the Profile of Mood States ( P O M S - S F ; Shacham, 1983). The P O M S - S F is a 37-item measure of psychological distress in which respondents indicate the degree to which each adjective describes them within the past week on a 5-point scale. For the purposes of this study, scale instructions were altered so that participants rated each adjective on the degree to which it describes them right now. The P O M S - S F yields an overall Total M o o d Disturbance score as well as scores for six subscales: Fatigue-Inertia, Vigor-Act ivi ty , Tension-Anxiety, Depression-Dejection, Anger-Hostility, and Confusion-Bewilderment. For this study, only scores on the depression subscale were used. Higher scores on a subscale reflect more intense feelings on that mood. The P O M S - S F has good psychometric properties, with internal consistencies ranging from .80 to .91, and is highly correlated with the original P O M S (Curran, Andrykowski, & Studts, 1995). For this study, five additional adjectives were embedded in the P O M S - S F to serve as items in the manipulation check. Studies on the effects of social exclusion vary greatly with respect to the type of manipulation check used. They range from single item mood measures (Baumeister et al., 2002) to questions that simply establish that the participant attended to the 40 manipulation (Buckley et al., 2004). We were interested in establishing that our manipulation created a significant difference between the groups on feelings of rejection. In general, past studies have used adjectives related to depressed, rejected, or negative affect, but none have provided evidence for the validity o f those measures. We compiled a list o f 13 adjectives, comprised of depression items from the P O M S - S F and an additional five adjectives thought to reflect rejected affect (alone, rejected, included, accepted, and excluded). A sample of 72 individuals (graduate students and community members) independently rated these items and the P O M S - S F depression items on the degree to which they described how one would feel following an experience of social exclusion. Mean item scores on a five-point scale ranging from 0 (not at all) to 4 (extremely) were as follows: alone (Af= 3.53, SD = .67), rejected (Af= 3.42, SD = .75), included (reversed; Af= 3.85, SD = .36), accepted (reversed; Af = 3.85, SD - .36), and excluded (Af = 3.51, SD = .61), unhappy (Af = 3.31, SD = .66), sad (Af= 3.31, SD = .71), worthless (Af = 1.76, SD = 1.16), hopeless (Af= 1.78, SD = 1.02), discouraged (Af = 3.22, SD = .81), helpless (Af= 1.86, SD = 1.07), miserable (Af= 2.44, £0=1.10), and blue (Af = 3.26, SD = .93). To construct the manipulation check, we adopted only those items with a mean score above 3.0 on the 0 to 4 scale. Thus, our manipulation check consisted o f the following items: alone, rejected, included (reversed), accepted (reversed), excluded, unhappy, sad, discouraged, and blue. Internal consistency for the nine manipulation check items was good with a coefficient alpha of .82. Mizes Anorectic Cognitions Questionnaire - Revised The Mizes Anorectic Cognitions Questionnaire - Revised ( M A C - R ; Mizes et al., 2000) is a 24-item self-report questionnaire that measures three dimensions o f anorectic cognitions: self-41 control as the basis for self-esteem, rigid weight regulation and fear of weight gain, and weight and eating behavior as the basis for approval. Items may also be combined for a total score. Items were rated on a 5-point Likert scale from 1 (strongly disagree) to 5 (strongly agree), with higher scores indicating more dysfunctional cognitions. The M A C - R has excellent psychometric properties including strong internal consistency (coefficients alpha range from .82 to .90) and good criterion-related validity (Mizes et al., 2000). For example, it discriminates between anorectic and bulimic patients on all but the rigid weight regulation subscale, and offers good convergent validity with other eating disorder scales (Mizes et al., 2000). State Self-Esteem Scale The State Self-Esteem Scale (SSES; Heatherton & Polivy, 1991) is a 20-item measure of momentary fluctuations in self-esteem that consists of three subscales: performance, social, and appearance self-esteem. It has good internal consistency (coefficient alpha = .92; Crocker, Cornwell, & Major, 1993), and is separable from mood effects. Moreover, it has been used in a number of eating disorder studies (e.g., Polivy & Herman, 1992; Saftner & Crowther, 1998). Only the appearance and social self-esteem subscales were used in this study. Screening Measures Beck Anxiety Inventory In order to screen for individuals with elevated levels of anxiety, we used the Beck Anxiety Inventory ( B A I ; Beck, Epstein, & Brown, 1988), a 21-item measure of severity of anxiety symptoms. Items were rated on a 4-point scale from "Not at a l l " to "Severely, I could barely stand it". Higher scores reflect greater anxiety. The B A I has high internal consistency and test-retest reliability, and good concurrent and discriminant validity (Beck et al., 1988). 42 Beck Depression Inventory - Second Edition The Beck Depression Inventory - Second Edition (BDI-II; Beck, Steer, & Brown, 1996) is a 21-item measure designed to assess severity o f depressive symptoms. Items were rated on a 4-point scale from 0 to 3, with higher scores indicating greater depressive pathology. Internal reliability of the B D I - I I is good (alpha = .93), and there is support for both convergent and discriminant validity (Beck et al.,' 1996), as well as criterion-related validity (Arnau, Meagher, Norris, & Bramson, 2001). Procedure Overview of Procedure Participants signed up to participate in a study that purportedly investigated the influence of personality on food preferences and how such preferences might change over time. They were informed that participation required attendance at two sessions, held exactly one week apart. Participants were run individually for all parts of the study. In session 1, participants completed a questionnaire package consisting o f demographic questions as well as measures for perfectionism, the B i g Five personality traits, state mood, anorectic cognitions, state self-esteem, and screening measures for anxiety and depression. They also completed a behavioral eating assessment that assessed the quantity o f food consumed across a variety o f snack foods. A t the end of the first session, participants with high levels of either anxiety or depression were thoroughly debriefed and did not return for session 2. For the remaining participants, they were provided with a package containing instructions and the informant-report versions of the M P S and PSPS and asked to return the completed package at their second session the following week. A t the outset of session 2, participants who were randomly assigned to either the accepted or excluded experimental groups received false feedback on the meaning o f their extraversion 43 s c o r e , a s c a l c u l a t e d f r o m t h e p e r s o n a l i t y m e a s u r e c o m p l e t e d a t s e s s i o n 1 ( i . e . , e x p e r i m e n t a l m a n i p u l a t i o n ) . P a r t i c i p a n t s i n t h e c o n t r o l g r o u p r e c e i v e d n o f e e d b a c k . T h e p r o c e d u r e f o r t h e e x p e r i m e n t a l m a n i p u l a t i o n w a s a d a p t e d f r o m p r e v i o u s w o r k ( T w e n g e , B a u m e i s t e r , T i c e , & S t u c k e , 2001). A l l p a r t i c i p a n t s t h e n c o m p l e t e d a q u e s t i o n n a i r e p a c k a g e t o a s s e s s l e v e l s o f s t a t e m o o d ( i n c l u d i n g i t e m s f o r t h e m a n i p u l a t i o n c h e c k ) , a n o r e c t i c c o g n i t i o n s , a n d s t a t e s e l f - e s t e e m , a n d t h e y a g a i n u n d e r w e n t t h e b e h a v i o r a l e a t i n g a s s e s s m e n t . F i n a l l y , p a r t i c i p a n t s c o m p l e t e d a s e r i e s o f q u e s t i o n s t o p r o b e f o r s u s p i c i o n o f h y p o t h e s e s a n d t o a s s e s s b e l i e v a b i l i t y o f t h e f e e d b a c k , a n d w e r e t h e n t h o r o u g h l y d e b r i e f e d . Session 1 W h e n p a r t i c i p a n t s a r r i v e d f o r t h e f i r s t s e s s i o n , t h e y w e r e t a k e n t o a s m a l l e r p r i v a t e r o o m a n d a s k e d t o c o m p l e t e a q u e s t i o n n a i r e p a c k a g e t h a t i n c l u d e d m e a s u r e s f o r b a s i c d e m o g r a p h i c i n f o r m a t i o n , p e r f e c t i o n i s m ( M P S a n d P S P S ) , B i g F i v e p e r s o n a l i t y t r a i t s ( B F I ) , a n o r e x i a n e r v o s a s y m p t o m v a r i a b l e s ( P O M S - S F , M A C - R , a n d S S E S ) , a s w e l l a s s c r e e n i n g m e a s u r e s t o i d e n t i f y t h o s e p a r t i c i p a n t s w i t h e l e v a t e d l e v e l s o f a n x i e t y o r d e p r e s s i o n ( B A I a n d B D I ) . P a r t i c i p a n t s w e r e t h e n a s k e d t o s a m p l e a v a r i e t y o f f o o d p r o d u c t s . F o u r d i f f e r e n t , c o m m o n l y a v a i l a b l e s n a c k f o o d s ( M i s s V i c k i e s s a l t a n d v i n e g a r c h i p s , R i t z B i t z c r a c k e r s , S u n m a i d r a i s i n s , a n d S m a r t i e s c a n d y -c o a t e d c h o c o l a t e ) w e r e p r e s e n t e d i n i n d i v i d u a l , f a c t o r y - s e a l e d p a c k a g e s t h a t w e r e o p e n e d i n f r o n t o f t h e p a r t i c i p a n t a n d e m p t i e d i n t o s e p a r a t e b o w l s w i t h n u m b e r e d p l a c a r d s i n f r o n t o f e a c h b o w l . P a r t i c i p a n t s w e r e i n f o r m e d t h a t w e w e r e i n t e r e s t e d i n f a c t o r s t h a t d e t e r m i n e a p e r s o n ' s f o o d p r e f e r e n c e s a n d h o w t h e s e p r e f e r e n c e s c h a n g e o v e r t i m e ; t h u s , s h e w o u l d c o m p l e t e a " t a s t e t e s t " t o e v a l u a t e t h e s n a c k f o o d s . E a c h p a r t i c i p a n t w a s t h e n a s k e d t o r a t e e a c h f o o d i t e m o n a s e r i e s o f g e n e r a l d i m e n s i o n s ( e . g . , s w e e t n e s s , c r u n c h i n e s s , t e x t u r e , s a l t i n e s s , e t c ) , a n d t h e n i n t e r m s o f g e n e r a l l i k i n g . A s w e l l , a l l p a r t i c i p a n t s w e r e t o l d t h a t " w e h a v e t o n s o f t h i s s t u f f , s o 44 please feel free to sample as much as you like and let us know i f you need more in order to make your ratings". In order to ensure privacy for the participants, the experimenter set a timer for 10 minutes and told the participant that she would be next door in the adjacent room and would return when the timer went off. The food items were weighed before and after each session. The amount eaten constituted the behavioral measure of disordered eating, and the session 1 quantity was used as a covariate in all analyses in which session 2 quantity was a criterion variable. Finally, because there was concern that some individuals may be more vulnerable to a social exclusion manipulation, we screened participants on levels of anxiety and depression.1 While the participant completed the taste test, the experimenter scored the B A I and B D I . If the participant's scores were at the mid-point of the moderate range or above on either scale (i.e., greater than 22 on the B A I or greater than 24 on the B D I ; Beck et al., 1988; Beck et al., 1996) she was excluded from the study at that point. Such a participant was told that we were interested in pre-selecting participants with specific psychological traits: those whose scores on two of the questionnaires completed in session 1 fell within a specific range. She was told that her scores fell outside of this range, and so she would receive 1 credit for her participation and that she would be excused from the remainder of the study. We did not identify these traits of interest and no participants inquired as to what they were. The participant was then thoroughly debriefed as to the true nature of the study, was provided with a list of related research articles and referral information for psychological services, and was offered the opportunity to ask questions or express concerns. She did not return for the second session. For those participants whose anxiety and depression levels were within the acceptable range, they were provided with a questionnaire package consisting of the informant-report versions of the M P S and PSPS , and a short cover letter explaining how to complete the scales. They were asked to take the package with them and 45 to have someone who knows them very well (e.g., parent, spouse, or close friend) complete the questionnaires over the next week. The letter instructed the reader to fdl-out the questionnaires as i f he or she was the target person that he or she was rating (see Hewitt & Flett, 1991b). In other words, to rate each statement as he or she believed the participant would rate that item. The participant was asked to return the completed questionnaires to the lab when she arrived for the second session. Session 2 When the participant arrived for the second session, she was seated in the same private room as was used in session 1. For participants who had been randomly assigned to the accepted or excluded group using a random number table, the experimenter explained that, in addition to providing course credit, the Psychology department also asks that study participants be provided with a learning experience whenever possible. In that vein, the experimenter noted that she had had an opportunity to score some of the participants' questionnaires from session 1 and would provide feedback on those scores.2 The experimenter then placed what appeared to be a computer generated N E O - P I report on the desk in front of the participant. She explained what the N E O - P I measures and explained to the participant what is meant by the term "extraversion". The experimenter then opened the false report to the second page to reveal a line graph with scores for each of the five personality dimensions plotted. She gave the participant accurate feedback about her extraversion score (high, medium, or low) and pointed to this region of the graph. The experimenter commented that this score was a good (or bad, depending on the experimental condition) thing for relationships and said that she would read the personality description to the participant. While still leaving the report in front of the participant, she turned to the final page of the report and read aloud the personality description designed to invoke a 46 s e n s e o f s o c i a l a c c e p t a n c e o r s o c i a l e x c l u s i o n , d e p e n d i n g o n t h e e x p e r i m e n t a l c o n d i t i o n ( s e e A p p e n d i x A f o r p e r s o n a l i t y d e s c r i p t i o n s ) . T h e p a r t i c i p a n t w a s f r e e t o r e a d t h e d e s c r i p t i o n a l o n g w i t h t h e e x p e r i m e n t e r . A f t e r r e a d i n g t h e d e s c r i p t i o n , t h e e x p e r i m e n t e r p a u s e d t o a l l o w t h e p a r t i c i p a n t t o d i g e s t t h e i n f o r m a t i o n a n d t h e n p r o c e e d e d t o c o l l e c t t h e m a t e r i a l s f o r t h e r e s t o f t h e s e s s i o n . P a r t i c i p a n t s i n t h e c o n t r o l c o n d i t i o n d i d n o t r e c e i v e a n y f e e d b a c k r e g a r d i n g a n y s c a l e s c o r e s o r t h e m e a n i n g s o f t h o s e s c o r e s . F r o m t h i s p o i n t o n , p r o c e d u r e s w e r e i d e n t i c a l f o r p a r t i c i p a n t s i n a l l t h r e e e x p e r i m e n t a l c o n d i t i o n s . A s i n s e s s i o n 1 , p a r t i c i p a n t s c o m p l e t e d a q u e s t i o n n a i r e p a c k a g e , t h i s t i m e c o m p o s e d o f t h e o u t c o m e m e a s u r e s o n l y ( P O M S - S F , M A C - R , a n d S S E S ) . T h i s w a s f o l l o w e d b y a t a s t e t e s t a s o u t l i n e d f o r s e s s i o n 1. W h e n t h e t i m e r w e n t o f f , s i g n a l i n g t h e e n d o f t h e t a s t e t e s t , t h e e x p e r i m e n t e r r e t u r n e d t o t h e r o o m . A t t h a t p o i n t , t h e p a r t i c i p a n t w a s a s k e d t o c o m p l e t e a g r a d u a t e d q u e s t i o n n a i r e t o a s s e s s b e l i e v a b i l i t y o f t h e f e e d b a c k a n d t o p r o b e f o r s u s p i c i o n . T h e q u e s t i o n n a i r e b e g a n w i t h a g e n e r a l q u e s t i o n a b o u t h o w t h e p a r t i c i p a n t u n d e r s t o o d t h e p u r p o s e o f t h e s t u d y a n d g r a d u a l l y p r o c e e d e d t o m o r e s p e c i f i c q u e s t i o n s a b o u t w h e t h e r t h e p a r t i c i p a n t g u e s s e d t h e s t u d y ' s h y p o t h e s e s o r f e l t t h a t h e r b e l i e f s r e g a r d i n g w h a t t h e s t u d y w a s a b o u t m i g h t h a v e a f f e c t e d h e r r e s p o n s e s . N o p a r t i c i p a n t s g u e s s e d t h e s t u d y ' s h y p o t h e s e s . I n a d d i t i o n , p a r t i c i p a n t s i n t h e a c c e p t e d a n d e x c l u d e d c o n d i t i o n s r a t e d t h e b e l i e v a b i l i t y o f t h e p e r s o n a l i t y f e e d b a c k o n a s c a l e f r o m 1 ( n o t a t a l l b e l i e v a b l e ) t o 1 0 ( v e r y b e l i e v a b l e ) . T h e p a r t i c i p a n t w a s t h e n i n f o r m e d a s t o t h e t r u e n a t u r e o f t h e s t u d y , w i t h a n e m p h a s i s o n t h e r a n d o m a n d f a l s e n a t u r e o f t h e s o c i a l f e e d b a c k , a n d w a s c a r e f u l l y a n d t h o r o u g h l y d e b r i e f e d . S h e w a s o f f e r e d t h e o p p o r t u n i t y t o a s k q u e s t i o n s o r e x p r e s s c o n c e r n s . P a r t i c i p a n t s w e r e a l s o b e p r o v i d e d w i t h a l i s t o f a r t i c l e s r e g a r d i n g p e r f e c t i o n i s m , s o c i a l e x c l u s i o n , a n d e a t i n g c o n c e r n s , w e r e g i v e n a l i s t o f 47 university-based counselling services, and provided with contact information for the principal and co-investigators should they wish to contact the lab in the future regarding this study. 48 R E S U L T S Prior to analyses, the data were examined to ensure that the assumptions of the statistical tests were adequately met (Tabachnick & Fidel l , 1996). First, missing data were identified and replaced by the subject item mean for the given subscale of that measure at that time point, unless more than 10% of the data were missing in which case that subscale was considered missing. Second, distributions were examined for the presence of univariate outliers and considerable skew or kurtosis. One participant was identified as an outlier on the Time 1 P O M S -SF. Analyses were executed both with and without this participant in the sample and the results were not substantially different, so she was left in the sample. The P O M S - S F depressed mood subscale exhibited significant positive skewness and positive kurtosis. Although this is not a formal assumption of regression, it did result in violations of the assumptions of normality and heteroscedasticity of the residuals. Thus, the P O M S - S F depressed mood subscale was transformed using a square root transformation and the transformed variable was used in all analyses. Next, multicollinearity and singularity were assessed by examining the bivariate correlations between predictor variables. The magnitude of all correlations were within accepted limits, but, in accordance with published procedures for detecting moderation effects (Aiken & West, 1991), the perfectionism predictor variables were centered by subtracting the mean from each score, and dummy coded variables for the experimental manipulation were left uncentered. The advantage of centering is to reduce the correlation between the interaction term and its constituent variables without changing the overall interaction or any aspect of the interaction that is subsequently examined (e.g., simple slope analysis). Also in accordance with Aiken and West (1991), criterion variables were left in their original, uncentered form. 49 The means, standard deviations and coefficients alpha for all measures are presented in Table 1 and were consistent with previous reports using nonclinical samples (e.g., Curran et al., 1995; Heatherton & Polivy, 1991; Hewitt & Flett, 1991b, Hewitt et al., 2003; John & Srivastava, 1999; Mizes et al., 2000). Further, all scales showed adequate internal consistency, though the coefficient alpha for B F I Agreeableness was somewhat lower than expected. Associations Among the Personality and Anorexia Nervosa Symptom Variables at Time 1 Zero-Order Correlations Zero-order bivariate correlations between all Time 1 variables are presented in Table 2 and were used to assess the degree of association of perfectionism with concurrent A N symptoms. We were specifically interested in the relationship of perfectionism to the amount of food consumed, depressed mood ( P O M S - S F depression subscale), anorectic cognitions ( M A C - R self-control, rigid weight regulation, and weight and approval subscales), and state self-esteem (SSES social and appearance self-esteem subscales). A multistage Bonferroni procedure (Larzelere & Mulaik, 1977) was used to control the family-wise Type I error rate in all analyses.3 Eating Behavior With respect to eating behavior, none of the trait perfectionism or perfectionistic self-presentation dimensions were significantly associated with the amount of food eaten at Time 1, suggesting that perfectionism does not, by itself, predict the quantity of food a person eats. Depressed Mood The results indicated that socially prescribed perfectionism, nondisplay of imperfection, and nondisclosure of imperfection all significantly predicted Time 1 depressed mood. This is 50 consistent with the literature on perfectionism and depression (e.g., Flett et al., 2003; Hewitt et a l , 1996; Hewitt et al., 2003). Anorectic Cognitions Self-Control. Consistent with the notion that certain facets of perfectionism are associated with A N symptoms (Bastiani et al., 1995, Cockell et al., 2002), perfectionistic self-promotion and nondisplay of imperfection both exhibited significant positive correlations with Time 1 measures of thoughts about self-control as the basis for self-esteem. This suggests that a desire to present a facade of perfection is associated with beliefs about rigid self-control of food intake as the basis for self-esteem. Rigid Weight Regulation. A s expected, the zero-order correlations also indicated that the social dimensions of perfectionism—socially prescribed perfectionism, perfectionistic self-promotion, nondisplay of imperfection and nondisclosure of imperfection—all exhibited significant positive correlations with cognitions about strict weight regulation and fear of weight gain. Thus, perfectionists who are concerned with the meeting the perceived expectations of others or with presenting a veneer of perfection to the world tend to think in all-or-nothing terms regarding routine and the likelihood of weight gain (e.g., "If I don't establish a daily routine, everything w i l l be chaotic and I won't accomplish anything" or "If I eat a sweet, it w i l l instantly be converted to stomach fat"). Weight and Approval. A s predicted, the social dimensions of perfectionism—socially prescribed perfectionism, perfectionistic self-promotion, nondisplay of imperfection and nondisclosure of imperfection—all exhibited significant positive correlations with Time 1 thoughts about weight and eating behavior as the basis for social approval. This suggests that 51 perfectionists with an interpersonal focus tend to think that being thin is associated with being accepted. State Self-Esteem Social Self-Esteem. The results for correlations with social self-esteem indicated that the social dimensions of perfectionism (socially prescribed perfectionism, perfectionistic self-promotion, nondisplay of imperfection and nondisclosure of imperfection) were all significantly negatively associated with confidence that one presents well socially and is generally accepted. Appearance Self-Esteem. A s for the social self-esteem subscale, the social dimensions of perfectionism (socially prescribed perfectionism, perfectionistic self-promotion, nondisplay of imperfection and nondisclosure of imperfection) were each significantly negatively associated with confidence in one's appearance. Unique Contributions of Perfectionism We were also interested in determining which perfectionism dimensions predicted unique variance in Time 1 outcome variables, after controlling for the effects of the other dimensions of perfectionism. Therefore, we conducted a series of multiple regression analyses in which all five perfectionism dimensions (self-oriented perfectionism, socially prescribed perfectionism, perfectionistic self-promotion, nondisplay of imperfection and nondisclosure of imperfection) were entered into the regression in a single step to predict the Time 1 outcomes. A multistage Bonferroni correction was applied, as outlined for the zero-order correlations, and all results are presented in Table 3. 52 Eating Behavior With respect to eating behavior, none of the trait perfectionism or perfectionistic self-presentation dimensions were significantly associated with the amount of food eaten at Time 1, after controlling for the variance associated with the other perfectionism dimensions. This is consistent with the zero-order correlations and suggests that perfectionism does not, by itself, predict the quantity o f food a person eats. Depressed Mood We confirmed a main effect for perfectionism on Time 1 P O M S - S F depressed mood. However, none of the perfectionism dimensions uniquely predicted greater P O M S - S F depressed mood over and above what is accounted for by the other perfectionism dimensions. Anorectic Cognitions When we examined anorectic cognitions, the results indicated that the perfectionism block was a significant predictor for each of the M A C - R subscales (self-control, rigid weight regulation, and weight and approval). Further, we found that perfectionistic self-promotion was a unique predictor o f maladaptive thoughts across all three o f these M A C - R subscales. This suggests that among perfectionism dimensions, perfectionistic self-promotion is uniquely and positively related to thoughts that are typical of anorectic psychopathology. State Self-Esteem Finally, our analyses predicting Time 1 state self-esteem confirmed a significant main effect of perfectionism on both o f the SSES subscales: social and appearance self-esteem. Moreover, nondisplay o f imperfection emerged as a significant predictor of SSES social self-esteem beyond the effects o f the other perfectionism dimensions, such that higher levels of 53 nondisplay of imperfection predicted a higher degree of concern about how one is viewed by others. The results for the appearance subscale of the SSES indicated that self-oriented perfectionism, socially prescribed perfectionism, and nondisplay of imperfection all predict unique variance in this aspect of state self-esteem. There was a negative association for the socially prescribed perfectionism and nondisplay of imperfection. That is, higher levels of perfectionism predict lower levels of appearance self-esteem. The opposite was true for self-oriented perfectionism: higher levels of self-oriented perfectionism predict higher levels of appearance self-esteem. This suggests that, once all other dimensions of perfectionism are controlled, self-oriented perfectionism predicts a positive evaluation of one's appearance. Although this conflicts with conceptualizations of self-oriented perfectionism, it is consistent with recent research showing that, in the absence of socially prescribed perfectionism, self-oriented perfectionism is positively associated with global self-esteem (Klibert et al., 2005). Beyond the B i g Five In light of some concerns that perfectionism may not be substantially different from other personality constructs, particularly neuroticism (Enns et al., 2005), we attempted to show that perfectionism is able to predict unique variance in the Time 1 outcome variables beyond that accounted for by the B i g Five personality traits: emotional stability (reverse scored neuroticism), extraversion, openness to experience, agreeableness, and conscientiousness. We conducted a series of hierarchical regressions in which the B i g Five personality traits were entered into the regression together in Step 1, followed by the five perfectionism dimensions in Step 2. Once again, a multistage Bonferroni correction was applied to control family-wise error, and the results are presented in Table 4. 54 Eating Behavior Among the B i g Five traits, only extraversion remained a significant positive predictor of the amount of food eaten at Time 1, after controlling for the other B i g Five subscales. However, none of the trait perfectionism or perfectionistic self-presentation dimensions were significantly associated with the amount of food eaten at Time 1, after controlling for the variance associated with the B i g Five traits and the other perfectionism dimensions. This is consistent with the zero-order correlations and suggests that perfectionism does not, by itself, predict the quantity of food a person eats. Depressed Mood Our results indicated that emotional stability was a unique negative predictor of Time 1 P O M S - S F depressed mood, beyond the variance associated with the other B i g Five traits. However, none of the perfectionism dimensions uniquely predicted P O M S - S F depressed mood over and above what was accounted for by the B i g Five traits and the other perfectionism dimensions. This is consistent with recent longitudinal work on trait perfectionism, neuroticism, and depression in adolescent inpatients (Enns et al., 2003), but it is at odds with other research showing that nondisclosure of imperfection provides incremental predictive power beyond neuroticism in a cross-sectional sample of university students (Hewitt et al., 2003). Anorectic Cognitions The results for the analyses of Time 1 anorectic cognitions revealed that emotional stability was significantly negatively associated with all three of the M A C - R subscales (self-control, rigid weight regulation, and weight and approval), after controlling for the influence of the other B i g Five traits. In addition, agreeableness emerged as a unique predictor of anorectic thoughts about the need to be thin to be accepted, over and above the variance accounted for by 55 the other B i g Five traits. However, most interesting was the finding that, across all three M A C - R subscales, perfectionistic self-promotion provides incremental explanatory power in the prediction of anorectic cognitions beyond that accounted for by the B i g Five traits and the other perfectionism dimensions. Thus, it appears that perfectionistic self-promotion has a strong and unique relationship to thoughts characteristic of A N . State Self-Esteem With respect to the Time 1 state self-esteem subscales, we found that both extraversion and emotional stability emerged as unique positive predictors of Time 1 social self-esteem, after controlling for the other B i g Five traits. Moreover, nondisplay of imperfection was negatively associated with social self-esteem, beyond the effects of the B i g Five traits and the other perfectionism dimensions. When we looked at appearance self-esteem, we found that all but one of the B i g Five traits uniquely predicted Time 1 appearance self-esteem. Extraversion, agreeableness, conscientiousness, and emotional stability were all positively associated with beliefs that one is attractive, after controlling for the other B i g Five traits. However, the previously significant negative associations of perfectionism with appearance self-esteem, were rendered nonsignificant after controlling for the B i g Five personality traits. It is worth noting that, in the previous analysis, there were multiple unique perfectionism predictors for appearance self-esteem. A similar pattern was observed when the B i g Five personality traits were entered into the regression: many of these traits emerged as significant predictors of appearance self-esteem. It suggests that confidence regarding one's appearance may be multiply determined by a number of different, but related, components of personality. 56 Associations among Perfectionism, Belongingness, and Time 2 Anorexia Nervosa Symptom Variables Restricting the Sample A t Time 1, the sample consisted of 149 participants. However, due to ethical concerns about the harmfulness of the social exclusion manipulation to high-risk participants, participants were screened on levels of depression and anxiety, as measured by the B D I and B A I . Individuals deemed to be at least moderately high on these scales were debriefed and provided with partial credit at the end of session 1 and did not return for session 2. Eighteen participants were screened out after the first session: 14 had high B A I scores, 2 had high B D I scores, and 2 had elevated scores on both the B A I and B D I . Thus, only 131 participants (46 accepted, 43 control, and 42 excluded) participated in session 2 and were, therefore, available for analyses involving Time 2 variables. Within this sample, participants varied on how believable they felt the social feedback (i.e., experimental manipulation) to be, and we elected to restrict the sample to only those participants who responded with a rating greater than 2 to the question "How believable was the feedback?" (l=Not at all believable, 10=Very believable). The remaining sample contained a total of 122 participants (45 accepted, 43 control, 34 excluded). In making this decision, we considered the possibility that the nonbelievers that we intended to exclude differed from the believers in a way that would make the meaning of our results unclear. Thus, we attempted to address a three key concerns: 1) D i d the nonbelievers differ from the believers on level of Time 2 rejected affect (i.e., the manipulation check)?, 2) D id those participants who did not believe the feedback differ in some meaningful way from those who did?, and 3) Given that there were a larger proportion of nonbelievers in the excluded group 57 than in the accepted group, did the experimental groups differ in some meaningful way once the sample was restricted? With respect to the first issue, we wished to explore whether the manipulation was actually less effective for participants who reported a low level of belief in the experimental feedback. We used participants in the excluded condition to determine whether there was a significant difference in Time 2 rejected affect between those participants who believed the feedback and those who did not. The results of the linear regression indicated that the nonbelievers reported significantly less rejected affect ( M = 4.50, SD = 4.34) than did the believers ( M = 9.02, SD = 5.97; F( l ,40) = 4.066,p = .50). Thus, it seems that the manipulation was not as effective at inducing feelings of social exclusion in participants who reportedly did not believe the feedback. To address the second concern, we examined the group of nonbelievers that we intended to eliminate from the sample to see i f they differed from the rest of the sample on any characteristics of interest. The results of the simple linear regression analyses indicated that those participants who claimed to have believed the feedback did not significantly differ from those who did not believe the feedback on age ( F ( l , 129) = 2.555, p = .112), number of years in Canada ( F ( l , 129) = .760, p = .385), or years of university education ( F ( l , 128) = 3.082, p = .082). The chi-square analysis confirmed that the nonbelievers did not significantly differ from the believers on ethnicity 0^(2) = 3.883,p = .143). Further, simple regression analyses showed that the believers were also indistinguishable from the nonbelievers with respect to the Time 2 outcome variables (Amount Eaten: F ( l , 129) = 2.778,/? = .098.; P O M S - S F Depression (square root): F(\, 129) = .251,p = .617.; M A C - R Self-Control: F ( l , 129) = .067,p = 797; M A C - R Rigid Weight Regulation: F ( l , 129) = .027,/? = 870; M A C - R Weight and Approval: F(\, 129) = 58 1.613,/?= .206; SSES Social: F ( l , 129) = .793,p = .375; SSES Appearance: F ( l , 129) = 2.500, /? = . 116). Thus, other than a significant difference on the degree to which they believed the social feedback (F(l, 129) = 61.669,p< .001) and their level of rejected affect, the nonbelievers were statistically identical to the larger sample of believers. To address the third concern, we attempted to determine i f the experimental groups differed significantly from one another on a range of variables once the sample was restricted to only those participants who believed the feedback (N= 122). The results of the simple linear regression analyses indicated that the three experimental groups (accepted, control and excluded) did not differ on age (F(2, 119) = 1.831, p = .165), number of years in Canada (F(2, 118) = .076, p = .927), or years of university education (F(2, 118)= 1.347,/? = .264), and the chi-square findings indicated that the groups were not significantly different in ethnic composition 0^(4) = 1.161,/? = .885). Finally, regression analyses also demonstrated that the three groups were not significantly different on level of perfectionism (SOP: F{2, 119) = .010,/? = .990; SPP: F(2, 119) = .382,/? = .684; PSP: F(2, 119) = .102,/? = .903; N D P : F{2, 119) = .079,p = .924; N D C : F(2, 119) = .132,/? = .876). Thus, taking participants' self-reported belief in the validity of the feedback at face value, and given that a) the manipulation was less effective for participants who did not believe the feedback, b) that nonbelievers were otherwise indistinguishable from believers on demographic and outcome variables, and c) that the experimental groups remained statistically similar on demographic and predictor variables, we restricted analyses involving Time 2 variables to only those participants who expressed a moderate to strong belief in the experimentally manipulated feedback. Therefore, all analyses involving Time 2 outcome variables were conducted on the sample of 122 participants (45 accepted, 43 control, 34 excluded). 59 Dummy Cod ing For the purposes of the regression analyses, social feedback condition was dummy coded, as outlined by A i k e n and West (1991). The control group was the designated reference group, thus facilitating comparisons between the control group and the accepted group, and between the control group and the excluded group. In order to compare the accepted group with the excluded group, we compared BA to BE by calculating relevant t-scores (Cohen, Cohen, West, & Aiken, 2003). However, for each Time 2 regression analysis we reported all three comparisons in the same table for ease of interpretation. Manipula t ion Check The experimental manipulation in this study included 3 levels of social feedback (accepted, excluded, and no feedback control). Our manipulation check consisted of 9 adjectives embedded in the P O M S - S F . Within our sample of 122 participants, these items had an internal consistency of a = .83. A multiple regression analysis (with experimental condition dummy coded) revealed a significant main effect for social feedback on feelings of rejection at Time 2 (F(2, 119) = 9.541, p < .001). Participants in the accepted group reported lower levels of rejected affect than those in either the control group (|3 = -233, p = .018) or the excluded group (P = -.448,p < .001), and participants in the excluded group reported significantly higher levels of rejected affect than participants in the control group (P = .199,/? = .041).4 In addition to the Time 2 manipulation check, we also explored whether there were group differences in feelings of rejection at Time 1 as a means to confirm that randomization was effective in equating the groups on this variable. The results o f the multiple regression analysis confirmed that there was no significant main effect of experimental condition on feelings of 60 rejection at Time 1 (F(2, 119) = .051, p = .951), suggesting that the group differences observed at Time 2 were indeed due to the experimental manipulation. Main Effects of Social Feedback To examine the effect of social feedback condition (accepted, excluded, or no feedback control) on anorexia nervosa symptoms at Time 2, we conducted a series of hierarchical multiple regression analyses. We entered the Time 1 equivalent of the Time 2 outcome variable into the regression in Step 1, followed by the social feedback condition (dummy coded) in Step 2. We were, therefore, predicting residual change scores, which are more reliable than simple change scores (Russell, 1990). The regression results are presented in Table 5. Following a multistage Bonferroni correction, there was a significant main effect of social feedback condition on the Time 2 P O M S - S F depressed mood (square root), after controlling for the Time 1 levels of P O M S - S F depressed mood (square root). Participants who received the social exclusion feedback at Time 2 had a greater increase in depressed mood compared to participants who received the social acceptance feedback. There were no other significant findings. „ Main Effects of Perfectionism We were also interested in the main effect of perfectionism on Time 2 A N symptoms, after controlling for Time 1 symptom levels. We executed a series of hierarchical regression analyses in which we entered the Time 1 equivalent of the Time 2 outcome variable into the regression in Step 1, and one of the five perfectionism dimensions (self-oriented perfectionism, socially prescribed perfectionism, perfectionistic self-promotion, nondisplay of imperfection, or nondisclosure of imperfection) in Step 2. A multistage Bonferroni correction was applied and the results are presented in Tables 6-12. 61 The findings revealed several significant main effects. Both perfectionistic self-promotion and nondisclosure of imperfection significantly predicted Time 2 M A C - R self-control scores, after the variance associated with the Time 1 level of M A C - R self-control was removed (Table 8). Thus, perfectionistic self-promotion and nondisclosure of imperfection predict changes in beliefs about self-control over time: as a person's level of PSP or N D C increases, over time the strength of her belief that she must exert self-control to feel good about herself also increases. Similar results were found for the weight and approval subscale of the M A C - R . Both perfectionistic self-promotion and nondisclosure of imperfection significantly predicted Time 2 M A C - R weight and approval scores, after the variance associated with Time 1 M A C - R weight and approval was removed (Table 10). The form o f the relationship indicated that as a woman's level of perfectionism (PSP or N D C ) increases, over time the strength of her belief that she must be thin to be liked and accepted by others also increases. Perfectionistic self-promotion and nondisclosure of imperfection were also significant predictors of changes to state social self-esteem such that as a person's level of perfectionism (PSP or N D C ) increases, over time her confidence in how she is regarded by others or compares to others decreases (Table 11). Similarly, nondisclosure of imperfection significantly predicted Time 2 SSES appearance self-esteem, after the variance associated with the Time 1 level of SSES appearance self-esteem was removed (Table 12). Thus, as a woman's reluctance to admit or discuss her shortcomings increases, over time, her satisfaction with her appearance decreases. Together these findings suggest that perfectionistic self-promotion and nondisclosure of imperfection play an important role in shaping changes in anorectic cognitions and state self-esteem over time. 62 Moderation Effects: Testing the Diathesis-Stress Model In order to ascertain whether the relationship between perfectionism and A N symptoms is moderated by social feedback condition, we conducted a series o f hierarchical multiple regression analyses. On the first step, we entered the Time 1 equivalent of the Time 2 outcome variable into the regression. In Step 2, one of the trait perfectionism or perfectionistic self-presentation variables was entered, followed by entry of the social feedback condition (dummy coded) in Step 3. The final step consisted of the perfectionism x social feedback condition interaction term. The criterion variables were the Time 2 measures of A N symptoms. Given the statistical difficulty of detecting moderator effects (McClel land & Judd, 1993), the family-wise Type I error rate was controlled at the . 10 level for all moderation analyses and a Bonferroni correction was not applied. Eating Behavior The results revealed that socially prescribed perfectionism interacted with social feedback to predict unique variance in the amount of food eaten at Time 2 (Table 6). The interaction indicated that the effect of socially prescribed perfectionism on the quantity of food consumed at Time 2 depends on a person's sense of belongingness and is different in the excluded group compared to the control group. To further explore the significant interaction, we conducted a series of simple slope regression analyses by substituting specific values for social feedback (based on the dummy coded values of 0 or 1) and graphed the interaction. A l l reported simple slope analyses were conducted in this fashion. The results of the simple slope analyses are depicted in Table 13. They indicated that the slope of the regression line of quantity of food eaten at Time 2 on socially prescribed perfectionism was significantly different from zero for the excluded group, but that the slope was not significantly different from zero for the control group. 63 More specifically, when women were given socially excluding feedback, those with higher levels of socially prescribed perfectionism showed dietary restriction. In contrast, when women did not received any social feedback, higher levels of socially prescribed perfectionism did not predict changes in the amount of food consumed (Figure 2). Although this finding is inconsistent with research showing that social exclusion leads to binge behavior (Baumeister et al., 2005; Patron, 1992), that work did not consider personality vulnerabilities and how those predispositions could alter response to social exclusion. In that regard, these findings are more appropriately compared with work showing that women with anorectic tendencies (predisposition) subsequently ate less during a taste test when they had been socially rejected (Rezek & Leary, 1991). Depressed M o o d With respect to depressed mood, the regression results displayed in Table 7 indicated that nondisplay of imperfection interacted with social feedback to predict unique variance in Time 2 P O M S - S F depressed mood (square root). The effect of nondisplay of imperfection on Time 2 depressed mood depends on a person's sense of belongingness and is different in the accepted group compared to both the excluded and control groups. Follow-up simple slope analyses indicated that the slope of the regression line of depressed mood at Time 2 on nondisplay of imperfection was significantly different from zero for the excluded group, but that the slope was not significantly different from zero for either the accepted or control group. More specifically, when women were given socially excluding feedback, higher levels of nondisplay of imperfection predicted higher levels of depressed mood. In contrast, when women did not received any social feedback or when they were given socially accepting feedback, higher levels of nondisplay of imperfection did not predict changes in depressed mood (Figure 3). 64 Although there were significant group differences between the accepted and control groups, the slopes o f the regression lines were not significantly different from zero (Table 13). Therefore, to better understand the nature of the group differences, we calculated 95% confidence intervals for the predicted values of Time 2 depressed mood for each group (accepted and control) at low and high observed values of nondisplay o f imperfection (Cohen et al., 2003). At low levels of nondisplay of imperfection, there were no significant differences in level of depressed mood for participants in the two groups (Control: Predicted 7=.38, 95%> CI = -1.17 to 1.92; Accepted: Predicted 7 = 7 1 , 95% CI = -.83 to 2.25). In contrast, at high levels of nondisplay of imperfection, women who were given socially accepting feedback had lower levels of Time 2 depressed mood than women in the no feedback control group (Control: Predicted 7=1.48, 95% CI = .10 to 2.86; Accepted: Predicted 7 = 0 8 , 95% CI = -1.29 to 1.46). Together, these findings suggest that social acceptance ameliorates the normally unfavorable influence of perfectionism on depressed mood, whereas an experience of social exclusion exacerbates the negative impact of perfectionism on depressed mood beyond that experienced by members o f the control group. Anorectic Cognitions Self-Control. There were no significant interactions in the prediction o f Time 2 M A C - R self-control (Table 8). Rigid Weight Regulation. Our analyses for Time 2 M A C - R rigid weight regulation suggested that the relationship between perfectionistic self-promotion and strict thoughts regarding weight regulation is affected by social feedback condition and is different in the excluded group compared to both the accepted group and the control group (Table 9). Upon further examination, the simple slope analyses demonstrated that the slope of the regression line 65 of Time 2 M A C - R rigid weight regulation on perfectionistic self-promotion was significantly different from zero for the excluded group, but that the slope was not significantly different from zero for either the accepted or control group (Table 13). Therefore, for women who received socially excluding feedback, higher levels of perfectionistic self-promotion predicted more rigid thinking about one's weight. In contrast, higher levels of perfectionistic self-promotion did not predict changes in rigid beliefs about weight regulation among women who received socially accepting feedback or who did not receive any social feedback (Figure 4). Weight and A p p r o v a l . Two significant interactions emerged as predictors of unique variance in Time 2 M A C - R weight and approval subscale (Table 10). The results indicated that the effect of socially prescribed perfectionism on Time 2 beliefs about weight and approval depends the type of social feedback one receives and is different for those in the accepted group compared to those in the excluded group. Follow-up simple slope analyses indicated that the slope of the regression line of Time 2 M A C - R weight and approval on socially prescribed perfectionism was significantly different from zero for the accepted group, but not for the excluded group (Table 13). Thus, for women who were given feedback indicating social acceptance, higher levels of socially prescribed perfectionism predicted a stronger belief that one needs to be thin to obtain social approval. However, for women who received socially excluding feedback, higher levels of socially prescribed perfectionism did not predict any change in beliefs about the need to be thin and attractive in order to be liked (Figure 5). The results also demonstrated that the effect of nondisclosure of imperfection on Time 2 beliefs about weight and approval depends the type of social feedback one receives and is different for those in the accepted group compared to those in the control group. Simple slope analyses suggested that the slope of the regression line of Time 2 M A C - R weight and approval 66 on nondisclosure of imperfection was significantly different from zero for the accepted group, but not for the control group (Table 13). Thus, for women who were given feedback indicating social acceptance, higher levels nondisclosure of imperfection predicted a stronger belief that one needs to be thin to obtain social approval. However, for women who did not receive any social feedback, higher levels of nondisclosure of imperfection did not predict any change in beliefs about the need to be thin and attractive in order to be liked (Figure 6). These results suggest that receiving feedback that one is socially accepted sparks an increase anorectic cognitions about weight and approval with increasing levels of perfectionism. However, for women who received no feedback or for women who were given socially excluding feedback, higher levels of perfectionism predicted no change in the levels of these types of thoughts. State Self-Esteem Social Self-Esteem. Wi th respect to Time 2 SSES social self-esteem, our findings indicated that self-oriented perfectionism interacted with social feedback condition to predict unique variance in Time 2 state social self-esteem, such that the effect of self-oriented perfectionism on social self-esteem was different in the accepted group compared to the control group (Table 11). Further, the simple slope analyses revealed that the slope of the regression line of Time 2 SSES social self-esteem on self-oriented perfectionism was significantly different from zero for the control group, but that the slope was not significantly different from zero for the accepted group. (Table 13). More specifically, for women who did not receive any feedback, increases in level of self-oriented perfectionism predicted increases in her state social self-esteem (i.e., belief that others view her favorably). In contrast, for women who were given socially accepting feedback, higher levels of self-oriented perfectionism did not predict change in social self-esteem (Figure 7). 67 Appearance Self-Esteem. Our analyses for Time 2 SSES appearance self-esteem suggested that the relationship between self-oriented perfectionism and appearance satisfaction is affected by belongingness and is different in the excluded group compared to both the accepted group and the control group (Table 12). Upon further examination, the simple slope analyses demonstrated that the slope of the regression line of Time 2 SSES appearance self-esteem on self-oriented perfectionism was significantly different from zero for the excluded group, but that the slope was not significantly different from zero for either the accepted or control groups (Table 13). Therefore, for women who received socially excluding feedback, higher levels of self-oriented perfectionism predicted lower levels of state appearance self-esteem. In contrast, higher levels of self-oriented perfectionism did not predict changes in appearance satisfaction among women who received socially accepting feedback or who did not receive any social feedback (Figure 8). Mediated Modera t ion In addition to testing the moderational models, we were also interested in examining whether the observed interaction effects were mediated by rejected affect. Our study was unique in that we used an experimental design and measured all of the components of the model in the temporally appropriate sequence. That is, we measured perfectionism at Time 1 to establish that it predated the onset of the moderator and mediator. A t the outset of session 2, we manipulated the participants' sense of belongingness (putative moderator), and then went on to measure rejected affect (putative mediator) and, subsequently, the dependent variables. Thus, positive findings for mediated moderation in our study would represent a strong test of the model: one that steps beyond a purely correlational research design, one in which causal inferences are possible. That is, because our moderator was experimentally manipulated and is independent of 68 perfectionism, we can make causal inferences about the overall effect on A N symptoms as well as the effect on our mediator, rejected affect (Muller, Judd, & Yzerbyt, 2005). In order to test the mediated moderational model, we used the method of Muller and colleagues (2005). Based on their work, three criteria must be met to establish support for mediated moderation (see Appendix B for the relevant equations): 1) There is an overall direct moderation effect (i.e., fa ^ 0); 2) Moderation of the residual direct effect of the treatment should be reduced compared to the moderation of the overall treatment effect (i.e., | fa1 < I fa1); and 3) One or both of the indirect paths from the treatment to the outcome must be moderated (i.e., both fa and fa are significant and/or both fa and fa are significant). Criterion 1 was already established for the significant findings in the preceding moderational analyses: the relationship between perfectionism and A N symptoms was moderated by social feedback condition. In this set of analyses, we explored whether the moderation effect on amount of food eaten and anorectic cognitions was mediated by rejected affect. The results of the multiple regression analyses indicated that only one of the significant moderation effects was mediated by rejected affect (Figure 9). The relationship between perfectionistic self-promotion and strict thoughts regarding weight regulation was affected by social feedback condition and was different in the excluded group compared to the accepted group (Criterion 1: fa = -.157, p = .046). Further, there was support for the indirect path via rejected affect (Criterion 3: fa = -.253,p = .054 and fa = . 2 1 8 , / ? = .013, fa = .447,/? = .002 and fa = .008,/? = .900). Finally, the moderation of the residual effect o f perfectionism was reduced (Criterion 2: fa = -.103,/? = .194). In fact, the moderation of the residual perfectionism effect was not only reduced in magnitude, but was, in fact, rendered nonsignificant, indicating "ful l" mediated moderation. 69 These results indicated that social feedback condition (accepted versus excluded) affects the magnitude of the effect of perfectionism on rejected affect. Follow-up simple slope analyses indicate that the slope of the regression line of Time 2 rejected affect on perfectionistic self-promotion was significantly different from zero for the excluded group ((3 = .296, t = 2.113,/? =.037), but that the slope was not significantly different from zero for the accepted group (P = -.063, t = -.537,p = .592). Therefore, for women who received socially excluding feedback, higher levels of perfectionistic self-promotion predicted higher levels of rejected affect. In contrast, higher levels of perfectionistic self-promotion did not predict changes in rejected affect among women who received socially accepting feedback (Figure 10). Furthermore, Time 2 rejected affect, in turn, is positively correlated with M A C - R rigid weight regulation (r = .35, p < 001). Thus, for women who were given socially excluding feedback, higher levels of perfectionistic self-promotion predicted an increase in rejected affect and a subsequent increase in rigid thinking about one's weight. In contrast, for women who received socially accepting feedback, higher levels o f perfectionistic self-promotion did not predict any change in level of rejected affect and thus, no change in subsequent thoughts about rigid weight regulation. 70 DISCUSSION We conducted an experimental study to test a diathesis-stress model of eating disorder development in which multidimensional perfectionism interacted with sense of belongingness to predict changes in a broad range of state symptoms of anorexia nervosa. In addition, we tested whether the diathesis-stress model was mediated by rejected affect. Finally, in the interest of replicating past work and providing more information on the unique role of perfectionism in the prediction of eating problems, we examined the concurrent relationships between perfectionism and eating disorder symptoms after controlling for other dimensions of perfectionism and for the B i g Five personality traits. We w i l l explore here our understanding and interpretation of the findings. Concurrent Relationships Trait Perfectionism The findings for the concurrent relationship between trait perfectionism and A N symptoms highlight the importance of socially prescribed perfectionism to anorectic eating disturbance. Consistent with current knowledge (e.g., Chang & Sanna, 2001; Flett et al., 2003), we found that socially prescribed perfectionism was significantly associated with depressed mood. Some have argued that depressed states have evolved as an adaptive mechanism to help organisms in unfavourable situations (Nesse, 2000). For example, depression may serve as a distress call to rally social support (Frijda, 1994), may inhibit risky activities in the absence of secure attachment bonds (Gilbert, 1992), or may signal submission in hierarchical conflict (Price, Sloman, Gardner, Gilbert, & Rohde, 1994). In addition, research suggests that depressed states play a role in helping an individual to disengage from an unreachable goal (Hamburg, 1974; Klinger, 1975) and that depressed mood is associated with more adaptive reasoning about social 71 risks (Badcock & Al len , 2003) and more systematic and realistic cognitions (Schwartz & Clore, 1996). Given that socially prescribed perfectionism is characterized by a drive to meet the (mis)perceived unrealistic expectations of others and is associated with a diminished tendency to carefully weigh alternatives to a problem in a planful and goal-directed manner (Flett, Hewitt, Blankstein, Solnik, & V a n Brunschot, 1992; Hewitt & Flett, 1993b), mi ld depressed mood may exert a much needed adaptive pull on such perfectionists. Unfortunately, we also know that socially prescribed perfectionism is associated with sociotropy (Bhar & Kyrios , 1999; Flett et al., 1997; Hewitt & Flett, 1993a), which appears to inhibit the adaptive risk-averse biases normally associated with depressed states and, thereby, perpetuate rejection experiences (e.g., via excessive reassurance seeking) and produce a depressive spiral (Badcock & Al len , 2003). Thus, socially prescribed perfectionists may be immune to the adaptive effects of mild depressed states and, therefore, more likely to spiral into serious depression. A s predicted, we also found that socially prescribed perfectionism was positively related to concurrent anorectic cognitions about the need for strict weight regulation and the need to be thin to be liked. The latter finding suggests that individuals high on socially prescribed perfectionism may see appearance as a means by which to achieve social acceptance and confirms past work indicating that socially prescribed perfectionists are focused on social approval (Bhar & Kyrios , 1999; Flett, et al, 1996; Hewitt & Flett, 1991b; Sherry et al., 2003). Finally, we also found that socially prescribed perfectionism predicted concurrent appearance and socially based self-esteem, suggesting that extrinsic motivation (i.e., drive to be perfect for others) is tied to less satisfaction with one's appearance and decreased confidence in how one is viewed by others. This fits with recent research on exercise motivation and self-esteem that indicates that although exercise is normally associated with a boost to self-esteem (Rodgers & 72 Gauvin, 1998), extrinsic motives for exercise (e.g., social recognition) are associated with lower self-esteem (Maltby & Day, 2001). Moreover, the results are consistent with existing literature on the importance o f socially prescribed perfectionism in eating difficulties (Cockell et al., 2002; Hewitt et al., 1995) and, more specifically, with the known relationship of socially prescribed perfectionism to lower levels of state self-esteem (Flett et a l , 1991; Flett et al., 1996; Hewitt et al., 2003). Unexpectedly, self-oriented perfectionism, or the self-driven need to be perfect, was not predictive of A N symptoms in the cross-sectional sample. Not only was this inconsistent with our predictions, it failed to replicate past research that reveals an association between self-oriented perfectionism and anorectic thoughts and behaviors (Bastiani, et al., 1995; Cockell et al., 2002), depressed mood (Flett et al, 1995; Flett et al., 2003), and self-esteem (Klibert et al. 2005). It is difficult to explain this discrepancy. It may stem from the unique characteristics of the sample used in this study. We measured subclinical A N symptoms in normal individuals rather than examining women with a clinical disorder. The strongest evidence for a relationship between self-oriented perfectionism and A N has been found in studies using carefully diagnosed clinical samples of A N patients (e.g., Bastiani, et al., 1995; Cockel l et al., 2002). Some have suggested that perfectionism in A N patients could be a behavioral expression of an underlying heritable biological vulnerability, such as a serotonergic disturbance (Kaye, Gwirtsman, George, & Ebert, 1991). Moreover, evidence indicates that the different intrapersonal components of perfectionism differ in their heritability and that, in general, intrapersonal components (e.g., concern over making mistakes) may have a stronger genetic basis than do interpersonal components (e.g., belief that parents have unrealistic expectations for oneself; Tozzi et al., 2004). Thus, it may be that self-oriented perfectionism in A N reflects a shared genetic trait, whereas the 73 link between socially prescribed perfectionism and A N is based on environmental rather than genetic factors. Given that A N patients are a select sample, we may be more likely to see potentially rare relationships with a genetic basis (i.e., S O P - A N link) than we would in a general nonclinical sample, which may only reveal relationships resulting from more common environmental factors (i.e., S P P - A N link). Alternatively, the failure to replicate past work may reflect the particular measures used here. Rather than assessing trait symptoms of A N , as is common in the literature, we measured state levels of the variables of interest. It may be that self-oriented perfectionism is related to chronic enduring trait symptoms, but not to state levels of those same variables. In any case, further research is warranted to clarify this deviation from previous reports. Perfectionistic Self-Presentation The results of our concurrent analyses supported a strong relationship between perfectionistic self-presentation and the cognitive, self-evaluative, and affective symptoms of anorexia nervosa. This is in line with existing work on the critical role of the concealing dimensions in depressive symptoms (Cockell, et al, 2002, Hewitt et al., 2003) and low self-esteem (Hewitt et al., 2003). However, the results provide new insight into the relationship between perfectionistic self-presentation and thoughts that are typical of anorexia nervosa. Our findings suggest that the need to appear to be perfect has a close relationship to these cognitive symptoms. From an information processing perspective, such thoughts derive from a negative cognitive schema that serves as a mental shortcut and, as such, is prone to perceptual and informational biases (Williamson, Stewart, White, & York-Crowe, 2002). Consistent with this, perfectionism has been associated with both negative cognitive bias (Beevers & Mil ler , 2004) 74 and rumination (Flett, Hewitt, Blankstein, & Gray, 1998). But, why are anorectic cognitions, specifically, so prevalent for individuals high on perfectionistic self-presentation? One possible explanation relates to the identity disturbance common amongst anorectics. Eating disordered women are thought to attempt to construct an adequate social self through heightened attention to their physical appearance (Striegel-Moore et al., 1993). They suffer from an overdeveloped false self (Johnson & Connors, 1987), and in the absence o f a true self they become hypervigilant to their public facade and the opinions of others. Thus, they may be more prone to internalize cultural or familial ideals of thinness, which is known to predict onset of eating pathology (Stice & Agras, 1998). Similarly, perfectionistic self-presentation is associated with public self-consciousness (Hewitt et al., 2003), which increases self-monitoring and processing of appearance related information (Cash, 2002). Along with its attendant need for social approval (Hewitt et al., 2003), perfectionistic self-presentation may increase one's investment in socially valued appearance standards. Therefore, like A N patients, perfectionistic self-presenters are likely more susceptible to the formation of negative appearance schemas via internalization of familial or cultural messages that pertain to body ideals and offer prescriptive messages about how to achieve such ideals. The presence of anorectic thoughts suggests that a cognitive or schema-based approach to treatment, involving reconstruction of distorted cognitions and change in core beliefs, may be fruitful for individuals high on perfectionistic self-presentation. Interestingly, none of the perfectionism dimensions was significantly associated with the amount of food eaten at Time 1. This suggests that, by itself, neither the need to be perfect nor the need to appear to be perfect can tell us how much food a woman w i l l eat at a given point in time. These findings are inconsistent with recent work pointing to a strong link between 75 perfectionism and a lifetime history o f fasting (Forbush et al., 2007), but fit with theoretical models and empirical findings in which perfectionism is predictive o f eating disorder symptoms only in the presence o f ego-involving stress (Heatherton & Baumeister, 1991; McGee et al., 2005; Slade, 1982). Consistent with work on A N patients that shows that perfectionism persists following recovery (Bastiani et al., 1995), our findings indicate that perfectionism may be an underlying vulnerability factor that lies dormant until the individual encounters a stressful failure and only then prompts pathological changes in food consumption. Our findings challenge future researchers to tease apart the symptoms of A N rather than collapsing across affective, cognitive, and behavioral symptoms, to use behavioral measures where possible, and to examine factors that may moderate the relationship between perfectionism and A N symptoms. Unique Contributions of Perfectionism Trait Perfectionism After controlling for the variance associated with the other dimensions of perfectionism, both self-oriented perfectionism and socially prescribed perfectionism were uniquely and significantly associated with concurrent appearance self-esteem. However, neither o f these predictors remained significant after controlling for the B i g Five traits. Thus, it would seem that trait perfectionism does not add significantly to the prediction o f concurrent appearance self-esteem beyond what can be gleaned from higher order traits. However, it is important to note that the pattern of results suggests that confidence in one's appearance is determined by a number of different, but related personality factors, suggesting that a parsimonious explanation for appearance self-esteem is elusive even when we move outward to higher order traits. Our results caution against the use of single factors (e.g., neuroticism) to predict confidence in and satisfaction with one's appearance. 76 Perfectionistic Self-Presentation Perfectionistic self-promotion emerged as a unique predictor of anorectic thoughts, and nondisplay of imperfection uniquely predicted social self-esteem, beyond the variance accounted for by the other perfectionism dimensions. Moreover, these effects remained significant after controlling for the influence o f the B i g Five personality traits. Together, these findings suggest that a desire to actively assert one's strengths is an important predictor of concurrent anorectic cognitions, whereas the need to avoid overt displays of imperfection is closely tied to concerns about how one is viewed by others, beyond the influence o f other personality traits. Paunonen (1998) argued that broader personality factors, such as the B i g Five, are likely to offer predictive superiority over the more numerous lower order traits because broad factors combine the variance common to those lower order measures. However, i f there is evidence of incremental validity for the lower order trait measures, then one must question whether the common variance underlying the broader factors is sufficient to optimally account for variations in human behavior. That is, incremental validity for the traits (e.g., perfectionism) suggests that there are predictive losses when the specific variance of the traits is discarded and the traits are aggregated into their superordinate factors. A critical conclusion from our findings is that the dimensions of perfectionistic self-presentation offer predictive information beyond that of the B i g Five factors and are, in fact, necessary to optimally predict A N symptoms. Moreover, the self-presentation dimensions are differentially predictive of A N symptoms. Thus, the field would benefit by the use of multidimensional measures of perfectionistic self-presentation and, in light of these findings, it would be inappropriate to use only B i g Five measures of personality to predict A N symptoms in future work. This is consistent with research in the depression literature showing that perfectionistic self-presentation offers predictive validity beyond that of the B i g 77 <J Five factors in predicting depression (Hewitt et al., 2003), and it extends these findings to the A N literature. Overall, the findings from the Time 1 analyses confirmed the role of the perfectionistic self-presentation dimensions and socially prescribed perfectionism in the prediction of A N symptoms. Moreover, the results extend existing research in two ways. First, we demonstrated that perfectionistic self-presentation (specifically PSP and N D P ) predicts cognitive and self-evaluative symptoms of A N beyond what can be explained by the B i g Five factors. This validates the use of perfectionism to predict A N symptoms, rather than relying solely on higher order personality traits. Further, it affirms the importance of considering multiple dimensions of perfectionism, as they are predictive of different aspects of the anorectic experience and, therefore, may suggest different targets for intervention in treatment. Second, our results extend existing work to state symptoms of A N . Prior research has generally focused on predicting non-state symptoms of eating disturbance (e.g., Cockell et al., 2002; Hewitt et al., 1995; M c V e y et al., 2002) rather than examining more immediate symptoms that could aid in understanding the episodic development or course of A N . Although the course of A N is quite variable, the illness is often marked by brief periods of partial remission interspersed between episodes of relapse ( D S M - I V - T R ; American Psychiatric Association, 2003; Kordy et al., 2002). Over the past two decades there has been a shift in patterns of hospitalization in eating disorder patients from a long-term inpatient treatment model to stabilization of acute episodes (Wiseman, Sunday, Klapper, Harris, & Halmi , 2001), but little work has focused on the precipitants of these episodes. Our study takes a microscopic view of A N symptoms and examines how such symptoms may change in the moment and what combination o f personality and/or events may cause that shift. This is consistent with Heatherton and Pol ivy 's (1992) spiral model, which 78 identifies negative self-awareness and low self-esteem as proximal causes of dietary restriction, and it suggests additional avenues for early intervention that focus on curtailing maladaptive shifts in state symptoms before they become entrenched patterns (Currin & Schmidt, 2005). Predicting Change in Anorexia Nervosa Symptoms Main Effects Social Feedback Condition Consistent with theory and research regarding the effect of social exclusion on negative mood (Baumesiter & Leary, 1995; Buckley et al., 2004; Will iams et al., 2000), we found that women who received socially excluding feedback experienced a greater increase in depressed mood than did women who received the social acceptance feedback. This suggests that social exclusion causes immediate and negative changes in a person's mood compared to being socially accepted. Our findings are consistent with work showing that participants who are told that no one wants to work with them or are ignored over the internet experience more negative affect (Buckley et al., 2004; Wil l iams et al., 2000), but the design of the current study did not allow us to discern how long these mood effects last. Although we expected to find a significant main effect of social feedback condition on anorectic cognitions, self-esteem and amount of food eaten, we found none of these. One possible explanation for this may be the strength of the manipulation. The mean difference in rejected affect between our accepted and rejected groups, while significant, was small (M A c cepted = 4.31, A/excluded = 9.03, possible range = 0 to 36), suggesting that our manipulation was relatively weak. Considering that the null findings are inconsistent with other studies showing that social exclusion leads to lower self-esteem (Williams et al., 2000) and larger amounts of food consumed (Baumeister et al., 2005), and that these studies used different types of 79 manipulations, it may be that the paucity of main effects for social feedback condition in our study is due in part to having a weak manipulation. However, this same issue may allow us to place relatively more emphasis on significant effects because they have occurred in the context of a relatively mild social exclusion manipulation. Perfectionism Trait Perfectionism. Contrary to expectations, there were no significant main effects for trait perfectionism on A N symptoms over time. Although this is inconsistent with past findings of a longitudinal relationship between trait perfectionism and depressive symptoms (Flett et al., 1995; Hewitt et al., 1996) and with recent work showing that perfectionism (as measured by the EDI-P) is strongly associated with fasting (Forbush et al., 2007), there are no studies that have explored the role of trait perfectionism in predicting state changes in anorectic symptoms, particularly over such short time period. However, our results suggest that the need to be perfect is not able to predict fluctuations in state A N symptoms over time. Perfectionistic Self-Presentation. Our main effect results underscored the importance of perfectionistic self-presentation in predicting change in state A N symptoms. We found that both perfectionistic self-promotion and nondisclosure of imperfection emerged as significant predictors of increasing belief in the need to exercise self-control to feel good about oneself and increasing belief in the need to be thin to be liked. Together with the concurrent findings, these findings offer new evidence of a strong role for perfectionistic self-presentation in the development of anorectic thoughts. In his functional analysis of anorexia, Slade (1982) discussed the strong need for perfectionists to assert control in some aspect of life. He suggested that the domain in which a person chooses to exert control is shaped by environmental influences such as peer modelling or 80 cultural norms. Consistent with this, Keel and Klump (2003) presented historical examples of medieval women who engaged in self-starvation in a cultural climate in which ascetic saints openly fasted in pursuit of a religious ideal. Three hundred years later, we see few examples of spiritual starvation, but many examples of starvation in pursuit of the thin ideal, with its emphasis on self-control and weight regulation. Thus, it is perhaps not surprising that the thoughts of modern day perfectionistic self-presenters who desire control would turn to ascetic beliefs about denying hunger cues and experiencing guilt and self-devaluation in response to indulgences. In addition to the relationship between perfectionistic self-presentation and anorectic thoughts, we also found that both perfectionistic self-promotion and nondisclosure of imperfection predicted decreasing confidence about how one is regarded by others. These results are consistent with past findings demonstrating a link between PSP and N D C and low self-esteem (Hewitt et al., 1995, Hewitt et al., 2003) and extend this work to a longitudinal context. Research has demonstrated that individuals with a desire to present a flawless facade are fearful of negative evaluation and have a strong drive for social approval (Hewitt et al., 2003). Moreover, they tend to be socially anxious and to focus on outwardly observable aspects of themselves, such as appearance (Hewitt et a l , 2003; Saboonchi et al., 1997). Unfortunately, however, their interpersonal strategy does not appear to be very successful in winning friends. There is evidence that others can detect the disingenuous front they construct (Hewitt et al., 2003) and may respond negatively (Hewitt et al., 2006). Therefore, the fact that perfectionistic self-presenters experience decreasing confidence in their own popularity over time could reflect hypervigilance to imagined signs of rejection, or it could be that such people actually are disliked by others and the decrements in social self-esteem are a natural response to ostracism. 81 Related to this, Hewitt, Flett, Sherry, and Caelian (2006) have explicated a model for suicidal behavior that suggests that, as a consequence of their interpersonal hostility, socially prescribed perfectionists are likely to experience objective social disconnection and then to engage in suicidal behaviors in response to these conditions. Thus, we suggest that, like socially prescribed perfectionism, perfectionistic self-presentation may have dire interpersonal consequences and, consistent with theory on stress generation and stress enhancement (Hewitt & Flett, 2002), perfectionists w i l l then react negatively to that social strain. Diathesis-Stress Model Eating Behavior Trait Perfectionism. Consistent with work by Rezek and Leary (1991), in the context of a personality predisposition, social exclusion produced dietary restriction. When women were given socially excluding feedback, higher levels of socially prescribed perfectionism were associated with greater dietary restriction, whereas when they were not given any feedback, higher levels of socially prescribed perfectionism did not produce any change in the amount of food consumed. This is a very interesting finding because our results show that perfectionism does not in and of itself predict quantity of food eaten, and past work indicates that, by itself, social exclusion causes binge behavior, not restriction (Baumeister et al., 2005). Thus, it would seem that dietary restriction occurs when a personality predisposition (i.e., SPP) rubs up against an experience of social exclusion. Individuals who believe that others have unrealistic standards for them are highly attuned to the wishes and expectations of others, display dependent attitudes (Sherry et al., 2003), and harbor a strong need for social approval and fear of negative evaluation (Bhar & Kyrios, 1999; Flett et al., 1996). We would expect that such persons would be extremely 82 sensitive to signs o f social exclusion, but why might an experience o f social exclusion prompt dietary restriction? There are a number of possibilities. First, we know that individuals high on socially prescribed perfectionism have a perceived lack of control over the standards that they feel compelled to live up to (Hewitt & Flett, 1991b). We also know that anorectic patients typically cite control as a reason for their behavior (Bruch, 1973, 1978). Perhaps exerting rigid control over eating behavior after an experience o f social exclusion is a means by which a woman high on socially prescribed perfectionism can compensate for the lack of control that she feels over the rest of her life. Second, dietary restriction may be a deliberate attempt to lose weight and, thereby, increase one's social attractiveness and restore social approval (Gilbert, 1997). However, given the pervasive sense of helplessness experienced by socially prescribed perfectionists (Flett et al., 1998), their focus on emotional rather than behavioral coping (Flett et al., 1994), and evidence that believing that others' acceptance is contingent on performance appears to inhibit active goal pursuit (Campbell & D i Paula, 2002), this seems a less likely explanation. Third, in line with escape from self-awareness theory (Heatherton & Baumeister, 1991), social failure may prompt negative self-awareness that is resolved through a dissociative state of cognitive deconstruction. Just as the repetitive behaviors associated with a binge may facilitate this state of reduced awareness (Schupak-Neuberg & Nemeroff, 1993), we propose that anorectic symptoms such as thoughts of calories, breaking food into small pieces, or slow chewing may divert attention from painful self-awareness. Alternatively, from a biopsychological perspective, dietary restriction may exert a positive effect on mood. Recent work suggests that serotonergic disturbances are evident in both i l l and recovered anorectic patients such that, while i l l , A N patients have lower levels of serotonin metabolites than controls but, once recovered, A N patients have higher levels of 83 serotonin metabolites compared to control women (Frank et al., 2001; Kaye, Strober, & Jimerson, 2004). Higher levels of serotonergic activity are associated with higher levels of anxiety, harm avoidance, perfectionism and, notably, appetite suppression (Kaye, 1997; Kaye et al., 2004). This may in turn lead to dietary restriction, which w i l l reduce the availability of the metabolic precursors of serotonin and, thereby, suppress serotonergic activity. Thus, dietary restriction may be an adaptive response for women with high levels of serotonergic activity (e.g., perfectionistic women) to the extent that it reduces dysphoric mood caused by social exclusion (Kaye et a l , 2003). Perfectionistic Self-Presentation. Perfectionistic self-presentation did not interact with social feedback condition to predict changes in the amount of food eaten. Together with the main effect findings, this suggests that the need to appear to be perfect does not exert a direct effect on a person's eating behavior, whether alone or in the context of an interpersonal stressor such as social exclusion. However, it does not preclude the (untested) possibility that perfectionistic self-presentation may indirectly affect the amount of food a person eats through its effect on depressed mood, anorectic thoughts, or self-esteem. Depressed Mood Trait Perfectionism. There were no significant interactions between trait perfectionism and social feedback condition that predicted changes in depressed mood. This was inconsistent with prior work showing that both self-oriented and socially prescribed perfectionism interact with stress to predict changes in depression over time (Chang & Rand, 2000; Enns et al., 2005; Hewitt et al., 1996). However, past work has operationalized depressive symptoms using the BDI , which includes somatic, cognitive, and behavioral symptoms in addition to mood ratings, and has typically looked for changes in depression over a period of a few months, rather than a 84 single week. It may be that trait perfectionism interacts with stress to predict long-term changes to the somatic, cognitive or behavioral aspects of depression, but not immediate shifts in depressed mood. Perfectionistic Self-Presentation. Our results indicated that women who are focused on avoiding overt displays of their imperfections are susceptible to mood fluctuations following interactions aimed at changing their sense of belongingness. Given the strong relationship between nondisplay of imperfection and depression in student and psychiatric samples (Hewitt et al., 2003), it is not surprising that social exclusion feedback led to an increase in depressed mood in women who were high on nondisplay of imperfection. However, this is the first demonstration of a diathesis-stress model for depressed mood involving nondisplay of imperfection. Moreover, the results present a novel opportunity to consider not only how a highly perfectionistic person might respond to failure (i.e., social exclusion), but also how she might respond to success (i.e., social acceptance). In this case, it appears that women with a desire to conceal their flaws can realize mood benefits from a positive experience of social acceptance. This has important implications for treatment. If a mild, experimentally induced form of social acceptance delivered by a stranger can elevate mood, imagine what the effect might be of a more powerful message of acceptance delivered by a trusted therapist. It suggests a possible means by which to shift mood amongst persons with concerns about having their imperfections seen by others. Anorectic Cognitions Trait Perfectionism. Our moderation results indicated that socially prescribed perfectionism interacted with sense of belongingness to predict thoughts about the importance of a thin appearance to social acceptance. Interestingly, for women who were given feedback indicating social acceptance, higher levels of socially prescribed perfectionism predicted a 85 stronger belief that one needs to be thin to obtain social approval. However, for women who received socially excluding feedback, higher levels of socially prescribed perfectionism did not predict any change in beliefs about the need to be thin and attractive in order to be liked. We did not expect that social acceptance would have this effect on highly perfectionistic women: that being told that one w i l l have many close relationships and w i l l be well-liked in the future would cause an increase in anorectic cognitions for women high on socially prescribed perfectionism. However, we believe that socially prescribed perfectionists may have heard predictions o f future interpersonal success as an expectation o f future success, an expectation that they felt compelled to live up to. This may have reinforced their existing beliefs about the high standards that others have for them and led to anticipation of future stress and failure (Hewitt & Flett, 2002). Given that high levels of socially prescribed perfectionism are associated with elevated levels of anxiety (Hewitt & Flett, 1991b), negative outcome expectancies (Martin, Flett, Hewitt, Krames, & Szantos, 1996), and a sense of personal helplessness and tendency to attribute both positive and negative outcomes to external factors (Flett et al., 1998), it is likely that being told that one w i l l be enormously popular in the future could lead to worry about how to achieve this seemingly unattainable goal. Indeed, there is evidence that believing that social acceptance is contingent on high achievement is associated with increased goal rumination (Campbell & D i Paula, 2002). In the current cultural context, it is not surprising that these goal-oriented worries would centre on thoughts about the need to be thin to be liked. A s well , such thoughts may take a complex and seemingly uncontrollable interpersonal situation (i.e., how can I fulfill the unrealistic expectations of others and gain social approval?) and narrow the focus to an intrapersonal arena (i.e., body shape/weight) that is directly under personal control. Consistent with this, Lawrence (1979) suggested that A N patients substitute 'self-control' for 'effective self-86 control', which involves interactions with others. They resolve their sense of interpersonal ineffectiveness by directing control inwards. This finding underscores the important role of the social acceptance condition in this study. It might have been tempting to assume that, with their desperate need for social approval and intense fear of negative evaluation (Flett et al., 1996, Hewitt & Flett, 1991b), socially prescribed perfectionists would benefit from feedback predicting future belonging. However, our results clearly show that high levels of perfectionism can foster deleterious outcomes even in the face of seemingly positive experiences. Perfectionistic Self-Presentation. We also found a significant interaction involving nondisclosure of imperfection that predicted anorectic thoughts about attractiveness and social approval. A s with socially prescribed perfectionism, we found that for a woman high on nondisclosure of imperfection, being told that she wi l l be well regarded by others in the future prompted an upswing in thoughts about the need to be thin in order to be accepted. In this case, individuals with a strong desire to avoid verbal admissions o f their imperfections who are told that they can look forward to a lifetime of close relationships may worry that such relationships wi l l only present additional situations in which they must monitor their disclosures. We believe that in this situation, persons high on nondisclosure of imperfection may begin to consider ways to achieve social popularity that do not involve personal disclosures, such as being physically attractive (Gilbert, 1997). This is consistent with evidence that nondisclosure o f imperfection is positively associated with self-monitoring and self-concealment (Hewitt et a l , 2003), and with a tendency to judge the self by external standards (Geller et al., 2000). The results of our study also indicated that perfectionistic self-promotion interacted with social feedback condition to predict change in thoughts about strict weight regulation. Among 87 women who received socially excluding feedback, higher levels of perfectionistic self-promotion predicted an increase in strict thinking about one's weight. However, higher levels of perfectionistic self-promotion did not predict changes in rigid beliefs about weight regulation among women who received socially accepting feedback or who did not receive any social feedback. When confronted with information suggesting future social exclusion, a perfectionistic self-promoter may redouble her efforts to erect a flawless facade and, thereby, secure social approval or admiration. One way to do so is to alter one's appearance to fit with cultural beauty norms. Thus, she may experience an increase in prescriptive thoughts such as "When I eat desserts, I get fat. Therefore I must never eat desserts so I won't be fat.", that suggest a means (i.e., dietary restriction) by which she can achieve her desired appearance. Indeed, this is consistent with research indicating that perfectionistic self-promotion, or the desire to assert one's strengths, is a proactive, rigid, and narcissistic interpersonal style characterized by high fear of negative evaluation (Hewitt et al., 2003). Moreover, we know from other studies that perfectionistic self-promotion is highly correlated with thoughts about having cosmetic surgery and that this is particularly true amongst gym-goers who may already be very appearance oriented (Sherry et al., 2004). Thus it is plausible that social exclusion w i l l catapult perfectionistic self-promoters into action-oriented thoughts geared towards securing a physical appearance that can elicit positive attention or admiration from others. State Self-Esteem Tra i t Perfectionism. The results indicated that self-oriented perfectionism interacted with social feedback condition to predict state changes in both social self-esteem and appearance self-esteem. However, the nature of these interactions was different. With respect to social self-esteem, amongst women who did not receive any feedback, increases in levels of self-oriented 88 perfectionism predicted increases in her state social self-esteem (i.e., belief that others view her favorably). In contrast, for women who were given socially accepting feedback, higher levels of self-oriented perfectionism did not predict change in social self-esteem. With respect to state appearance self-esteem, amongst women who received socially excluding feedback, higher levels of self-oriented perfectionism predicted lower levels of state appearance self-esteem. In contrast, higher levels of self-oriented perfectionism did not predict changes in appearance satisfaction among women who received socially accepting feedback or who did not receive any social feedback. H o w can we make sense of these findings? Recent research has demonstrated that self-oriented perfectionism is positively correlated with self-esteem (Klibert et al., 2005). However, work by Campbell and D i Paula (2002) indicated that the self-oriented perfectionism construct can be broken into two components, the belief that it is important to be perfect and the belief that one actively strives for perfection, and that these two components are differentially related to self-esteem. Perfectionistic striving is related to higher global self-esteem, whereas the importance of being perfect predicts lower global self-esteem. Similarly, our results also caution against simplistic explanations of the relationship between self-oriented perfectionism and self-esteem. That is, i f self-oriented perfectionism was positively correlated with higher social or appearance self-esteem in a straightforward manner, we might have expected that when their sense of belongingness was enhanced, self-oriented perfectionists would feel more confident about how others view them or feel more attractive. Instead we saw no change in state social or appearance self-esteem. Researchers have noted that we must look beyond simple considerations of the level of self-esteem and also consider the stability o f self-esteem (Kernis, 2005). Level of self-esteem 89 refers to a person's representations of her feelings of self-worth, whereas stability of self-esteem refers to the magnitude of short-term fluctuations that people experience in their contextually based, immediate feelings of self-worth (Kernis, 2005). Our findings imply that, in the absence of external stressors, self-oriented perfectionists may generally enjoy higher levels of self-esteem, but that their self-worth is immune to positive events or successes, while remaining susceptible to failures. This peculiar type of self-esteem instability may render them vulnerable to sudden downward shifts in state self-esteem in response to failure, thus creating the conditions of low self-worth that are so common in eating disorders (Baumeister, Campbell, Krueger, & Vohs, 2003). Perfectionistic Self-Presentation. There were no significant interactions of perfectionistic self-presentation with social feedback condition to predict state self-esteem, suggesting that although the desire to appear to be perfect is related to state self-esteem, a person's sense of belongingness does not affect this relationship. Overall the results from the moderation analyses indicate that both trait perfectionism and perfectionistic self-presentation function as diatheses that predict changes to state A N symptoms in the presence of varying levels of belongingness. In particular, the results highlighted the importance of using multidimensional measures of perfectionism because the different dimensions of perfectionism are predictive of different aspects of the anorectic experience. For example, both perfectionistic self-promotion and nondisclosure of imperfection have a close relationship to anorectic cognitions, whereas self-oriented perfectionism is predictive of changes in state self-esteem, and only socially prescribed perfectionism predicted a change in eating behavior following social exclusion. In addition, the diathesis-stress results suggested that a 90 seemingly positive experience of social acceptance may have no effect or may even cause an adverse reaction in highly perfectionistic individuals. These findings have important implications for prevention and treatment. In order to effectively target AN symptoms, it would be useful to know what dimensions of perfectionism are elevated in a particular individual and to use that information as a guide to select treatment targets (e.g., mood, thoughts, eating behavior, self-esteem). Further, in light of perfectionists' sometimes maladaptive responses to social acceptance, it would be prudent to debrief with the patient following supportive interventions. For example, the therapist might inquire about how it was to receive accepting feedback in session and whether the patient has any thoughts or concerns about that feedback or about what the therapist might be thinking or expecting of her. This follow up will highlight any maladaptive reactions to therapist acceptance and provide opportunities to discuss and clarify concerns. The Role of Rejected Affect The results of the mediated moderation analyses indicated that rejected affect mediated the diathesis-stress model for anorectic cognitions about rigid weight control. For women who were given socially excluding feedback, higher levels of perfectionistic self-promotion predicted an increase in rejected affect and a subsequent increase in rigid thinking about one's weight. In contrast, for women who received socially accepting feedback, higher levels of perfectionistic self-promotion did not predict any change in level of rejected affect and thus, no change in subsequent thoughts about rigid weight regulation. Therefore, we believe that increasing thoughts about using strict dietary strategies to control weight emerge in perfectionistic self-promoters when they have been socially excluded either as a means by which to distract from distressing rejected affect, or as an active solution-focused approach to resolving the problem. 91 From a cognitive perspective, rejected affect may activate negative self-schemas regarding weight and appearance. In their cognitive information-processing model of body image, Will iamson and colleagues (2002) suggested that because negative self-schemas (such as negative body image) are typically associated with unpleasant mood states, the activation of a negative mood can reciprocally activate the body self-schema. That is, negative mood interacts with the self-schema to increase the likelihood of cognitive biases. In a similar way, rejected affect arising from high levels of perfectionistic self-promotion in the context of social exclusion, may activate associated negative self-schemas regarding weight and appearance that are comprised of thoughts and beliefs about the need for rigid weight control. Our findings reflect not only a statistically sound model of episodic A N symptom development, but one that is backed by a strong experimental research design in which the components of the model were assessed in the appropriate temporal sequence. Thus, we feel confident in the validity of our conclusions. To our knowledge, this is the first time a mediated moderational model of A N symptoms has been tested. It provides insight into the mechanism of action for the observed changes to state A N symptoms and stands in contrast to Heatherton and Baumeister's (1991) assertion that failure in the context of perfectionistic standards wi l l lead to a deconstructed state characterized not by negative affect, but by emotional numbing. It also offers a counterpoint for work demonstrating that mood does not mediate the relationship between social exclusion and outcome (Baumeister et a l , 2005; Baumeister et al., 2002, Williams et al., 2000), while offering general support for recent work revealing a mediational role for negative affect (Buckley et al., 2004). However, we must also acknowledge that the presence of null findings in our mediated moderation analyses suggests that rejected affect does not always mediate the effects of perfectionism on A N symptoms in the context of social exclusion. 92 Conclusions about Perfectionism and Anorectic Symptoms The findings of this work have important implications for our understanding of perfectionism, particularly as it relates to A N symptoms. In the early days of perfectionism research, the construct was generally defined as a unidimensional construct rooted in the rigid pursuit of unrealistic self-imposed standards (e.g., Burns, 1980). However, consistent with clinical observations (Hamachek, 1978; Pacht, 1984) and growing empirical support (e.g., Hewitt & Flett, 1991b, Frost et al., 1990) the field of perfectionism shifted to a multidimensional conceptualization of perfectionism that made room for interpersonal as well as intrapersonal components. Although there is now over a decade of research supporting the validity and utility of a multidimensional model of perfectionism (see Flett & Hewitt, 2002), the eating disorders field has been slow to warm to this new understanding, perhaps because the commonly used Eating Disorders Inventory (Garner et al., 1983) carries its own convenient, i f clouded, unidimensional perfectionism subscale. Unfortunately, research on the relationship between unidimensional perfectionism and eating pathology has stalled in its ability to predict specific eating disorder symptoms or to explain the nature of those relationships. This, in turn, has affected our ability to design treatments (e.g., cognitive behavioral, interpersonal, family systems) that w i l l target those issues directly. The results o f this study provide a new understanding o f the relationship between the different dimensions of perfectionism and particular aspects of the anorectic experience. Broadly, our findings indicate that trait perfectionism and perfectionistic self-presentation play different roles in the development of A N symptoms. It appears that the perfectionistic self-presentation, but not trait perfectionism, is able to predict changes in anorectic thoughts and evaluations of how one is viewed by others over a short period of time. In particular, 93 perfectionistic self-promotion (i.e., desire to actively assert one's perfection) and nondisclosure of imperfection (i.e., avoidance of verbal admissions of imperfection) appear to be tightly linked to changes in anorectic thoughts and self-esteem, suggesting that these dimensions have a special relationship to the cognitive aspects of the A N experience. We found a somewhat different picture for the relationship between perfectionism and A N in the context o f social exclusion. Dimensions o f both trait perfectionism and perfectionistic self-presentation interacted with stress to predict change in A N symptoms. Again, perfectionistic self-presentation predicted changes in the cognitive symptoms of anorexia as well as to the affective signs. In contrast, trait perfectionism was predictive of changes in eating behavior and self-esteem following an experience of social exclusion. Thus, our findings suggest that the trait and self-presentational dimensions are related to different aspects of the anorectic experience, and are likely to lead to a rise in symptoms under different conditions. In brief, trait perfectionism is likely to lead to changes in behavioral and self-evaluative A N symptoms under conditions of stress, whereas perfectionistic self-presentation is likely to lead to changes in cognitive and affective A N symptoms during stress, and to a worsening of cognitive and self-evaluative difficulties under normal conditions. These findings offer new insight into the pattern of A N development and suggest that individuals may take different paths en route to developing a full blown disorder. It supports the importance of dividing both perfectionism and anorectic symptoms into their underlying components, rather than considering each as a uniform block. A n d it offers new targets for early intervention or relapse prevention in subclinical populations. 94 Strengths and Limitations of the Current Study It is important to consider these results in light of this study's strengths and limitations. There are a number of strengths in this work. First, in addition to self-report measures, we conducted a behavioral eating assessment and collected informant ratings of perfectionism. Second, we used state measures for all variables to reflect immediate symptom fluctuations that are most relevant to an episodic model of A N symptom development. Third, we included a social acceptance condition to explore what effect a seemingly positive experience would have on highly perfectionistic individuals. Fourth, we used a longitudinal experimental design that permits us to draw causal conclusions from our results. Finally, we explored a putative mechanism to explain how perfectionism leads to A N symptoms in the context of social exclusion and social acceptance. There are also several limitations of the present work that should be acknowledged. First, this study uses a university student sample rather than a clinical one. Although eating disorders are conceptualized by many as a continuum o f symptom severity (Butler et al., 1990; Stice et al., 1998) and an estimated 64% of female university students exhibit some degree of eating disordered behavior (Mintz & Betz, 1988), it may be that perfectionism plays a slightly different role in A N patients. For instance, the strongest evidence for a link between self-oriented perfectionism and A N symptoms has emerged from research on anorectic patients (e.g., Bastiani, et al., 1995; Cockel l et al., 2002). Thus, we may have been more likely to see significant effects involving self-oriented perfectionism in a clinical sample. It's possible that the relationship between self-oriented perfectionism and A N has not been apparent in nonclinical samples because self-oriented perfectionism predicts severe psychopathology, but not subclinical symptoms. If true, this would have enormous clinical value in helping us to predict which 95 individuals w i l l become seriously i l l . However, it is a possibility that has yet to be tested directly and additional work to determine i f these findings generalize to a clinical sample is warranted. In addition, we made no attempt to tease apart the A N subtypes and this too could affect the perfectionism-AN relationship. For example, some work has indicated that women with restricting subtype A N are more perfectionistic than women with the binge-purge subtype (Steiger, Puentes-Neuman, & Leung, 1991). Second, due to ethical concerns, we limited our sample to individuals with only mild to moderate levels of anxiety or depression. In so doing, we likely excluded a number of participants with high perfectionism scores, thus, eliminating the very people we were interested in studying. On the other hand, the means and ranges for our perfectionism subscales were comparable to other university samples, which suggests that our screening efforts did not affect these basic sample characteristics. Third, we restricted our sample to those participants who believed the social feedback manipulation. Although we were unable to find any meaningful differences between participants who did not believe the feedback and those who did, the possibility remains that the two groups differed on a variable that we did not measure, but one that could affect the interpretation of our findings. Fourth, our social feedback manipulation was relatively weak. Although this strengthens the validity of the significant findings, it raises questions about what effects might be seen with a stronger manipulation or one that involves immediate social exclusion (or acceptance) with a live person rather than predictions of future social exclusion (or acceptance) based on personality feedback. For example, perhaps emotional numbing occurs in the context of a stronger manipulation when aversive self-awareness is more likely to be overwhelming. Finally, although we attempted to study a variety of anorectic symptoms, there are a number of important signs, such as body image, excessive exercise, and anger suppression, that we did not explore. Consistent with recent 96 work by Forbush and colleagues (2007), breaking the anorectic experience down into discrete characteristics may help to clarify the complex relationship between perfectionism and A N . Future research should address these issues directly. 97 Table 1 Means, Standard Deviations, and Alpha Reliability of the Time 1 and Time 2 Variables Variable M SD a Time 1 Variables TI Self-Oriented Perfectionism (self-report) 71.06 14.76 0.90 TI Socially Prescribed Perfectionism (self-report) 53.57 12.86 0.82 TI Perfectionistic Self-Promotion (self-report) 40.37 11.49 0.87 TI Nondisplay of Imperfection (self-report) 43.77 12.08 0.89 TI Nondisclosure of Imperfection (self-report) 21.10 7.49 0.79 TI Self-Oriented Perfectionism (informant-report) 72.86 16.47 0.92 TI Socially Prescribed Perfectionism (informant-report) 55.27 12.55 0.82 TI Perfectionistic Self-Promotion (informant-report) 41.35 11.68 0.89 TI Nondisplay of Imperfection (informant-report) 42.39 11.00 0.87 TI Nondisclosure of Imperfection (informant-report) 24.09 7.67 0.80 TI B F I Extraversion 3.32 0.56 0.86 TI B F I Agreeableness 3.84 0.56 0.67 TI BFI Conscientiousness 3.60 0.68 0.78 TI BFI Emotional Stability 2.69 0.74 0.75 TI B F I Openness to Experience 3.53 0.67 0.81 TI Total Amount Eaten (grams) 55.01 38.40 ~ TI P O M S - S F Depressed M o o d 2.66 3.53 0.94 98 Variable M SD a TI M A C - R Self-Control 24.85 7.43 0.86 TI M A C - R Rig id Weight Regulation 17.05 5.36 0.72 TI M A C - R Weight and Approval 17.48 5.17 0.71 TI SSES Social 25.48 5.58 0.81 TI SSES Appearance 20.27 4.61 0.83 Time 2 Variables T2 Total Amount Eaten (grams) 66.40 32.35 ~ T2 P O M S - S F Depressed M o o d 3.04 4.38 0.90 T2 M A C - R Self-Control 23.16 7.41 0.87 T2 M A C - R Rig id Weight Regulation 15.48 5.02 0.73 T2 M A C - R Weight and Approval 16.29 4.80 0.76 T2 SSES Social 26.74 4.78 0.81 T2 SSES Appearance 21.37 4.09 0.81 Note. The following labels were used: B F I (Big Five Inventory), P O M S - S F (Profile of Mood States - Short Form), M A C - R (Mizes Anorectic Cognitions Scale - Revised), and SSES (State Self-Esteem Scale). The Time 1 sample consisted of 149 participants, whereas the Time 2 sample consisted of only 131 participants (18 were screened out after session 1). 99 Table 2 Zero-Order Correlations of the Time 1 Personality and Anorexia Nervosa Symptom Variables Variable SOP SPP PSP N D P N D C SOP 1 ~ — ~ — SPP 33** 1 — — ~ PSP .51** .56** 1 — — N D P .26** ^ j * * .68** 1 ~ N D C 29** .61** .65** .56** 1 BFI Ext .02 -.18* -.19* -.42** -.24* BFI Agr -.09 -.28** _ 29** -.18* _ 32** BFI Cons .38** -.21* .02 -.16 -.11 BFI Emo _ 29** -.36** _ 39** -.45** -.31** BFI Open .08 -.17* -.19* -.23* -.06 T l Food Eaten .09 .04 -.06 -.04 -.04 T l P O M S - S F Dep .16 31 **-j- .22* .25*1 .28**1 T l M A C - R Self-Control .17* .17* .36**1 .31**t .23* T l M A C - R Rig id Weight Regulation .22* 4 0 * * | .54**f .46**1 47**| T l M A C - R Weight and Approval .18* .36**f .52**f 37**! .43**1 T l SSES Social -.20* . 4 9 * * | -.52**f -.63**f -.50**1 T l SSES Appear .00 .43**1 -.36**1 -.45**f -.39**f 100 Variable B F I Ext B F I Agr B F I Cons B F I Emo B F I Oper SOP ~ ~ — SPP — — — ~ — PSP — — — — N D P — — — — N D C — — — — BFI Ext 1 — — ~ BFI Agr .14 1 ~ — — BFI Cons .14 .16* 1 — BFI Emo .28** 3g** .10 1 ~ B F I Open .26** .05 -.01 .12 1 T l Food Eaten .18* .01 .03 -.06 .08 T l P O M S - S F Dep -.20* _ 31** -.07 -.55** -.06 T l M A C - R Self-Control -.06 -.11 -.11 -.28** -.07 T l M A C - R Rig id Weight Regulation -.19* -.15 -.15 -.37** -.19* T l M A C - R Weight and Approval -.21* -.36** -.36** -.38** -.17* T l SSES Social .23* .23* .52** .05 T l SSES Appear 33** 41 ** .30** .42** .17* 101 Variable TI Food Eaten TI P O M S - S F Depressed Mood TI M A C - R Self-Control TI M A C - R Rig id Weight Regulation TI M A C - R Weight and Approval SOP SPP PSP N D P N D C BFI Ext BFI Agr BFI Cons BFI Emo BFI Open TI Food Eaten TI P O M S - S F Dep TI M A C - R Self-Control TI M A C - R Rig id Weight Regulation TI M A C - R Weight and Approval TI SSES Social TI SSES Appear 1 .08 -.10 -.04 -.03 .03 .16 1 .17* .31 ** .35** . 41 ** -.34** .61 ** .46** -.34** _ 49** .53 ** -.52** -.48** 1 45** .51** 102 Variable T I SSES TI SSES Social Appear SOP SPP PSP N D P N D C BFI Ext BFI Agr BFI Cons B F I Emo BFI Open TI Amount Eaten TI P O M S - S F Dep TI M A C - R Self-Control TI M A C - R Rig id Weight Regulation TI M A C - R Weight and Approval TI SSES Social 1 TI SSES Appear .55** 1 Note. The following labels were used: T I (Time 1), SOP (Self-Oriented Perfectionism), SPP (Socially Prescribed Perfectionism), PSP (Perfectionistic Self-Promotion), N D P (Nondisplay of 103 Imperfection), N D C (Nondisclosure of Imperfection), B F I Ext (Big Five Inventory, Extraversion subscale), B F I Agr (Big Five Inventory, Agreeableness subscale), B F I Cons (Big Five Inventory, Conscientiousness subscale), B F I Emo (Big Five Inventory, Emotional Stability subscale), BFI Open (Big Five Inventory, Opennes to Experience subscale), P O M S - S F Dep (Profile of Mood States - Short Form, Depressed M o o d subscale), M A C - R (Mizes Anorexic Cognitions Scale -Revised), SSES (State Self-Esteem Scale), and Appear (Appearance). * p < .05; ** p < .001; | Significant after multistage Bonferroni correction. 104 Table 3 Summary of Regression Analysis for Perfectionism Dimensions Predicting Unique Variance in Time 1 Anorexia Nervosa Symptoms Predictor P p Total R R2 F df Criterion Variable: T I Total Amount of Food Eaten (grams) .172 .030 .876 5, 143 .499 SOP .155 .112 SPP .101 .361 PSP -.189 .179 N D P .009 .936 N D C -.026 .829 Criterion Variable: T I Square Root of Profile of M o o d States, Depression Subscale .343 .118 3.823 5,143 .003* SOP .066 .474 SPP .197 .064 PSP -.081 .544 N D P .106 .344 N D C .136 .234 105 Predictor p p T o t a l s R2 F df Criterion Variable: T l Mizes Anorexic Cognitions Scale, Self-Control Subscale .372 .138 4.580 5,143 .001* SOP -.004 .965 SPP -.068 .514 PSP .311 .019* N D P .130 .242 N D C -.002 .983 Criterion Variable: T l Mizes Anorexic Cognitions Scale, R ig id Weight Regulation Subscale .577 .333 14.247 5, 143 <.001* SOP -.066 .416 SPP .057 .534 PSP .375 .002* N D P .099 .312 N D C .159 .109 106 Predictor P p T o t a l s R2 F df Criterion Variable: T I Mizes Anorexic Cognitions Scale, Weight and Approval Subscale .545 .297 12.103 5,143 <.001* SOP -.118 .155 SPP .069 .463 PSP .483 <001* N D P -.039 .697 N D C .131 .199 Criterion Variable: T I State Self-Esteem Scale, Social Subscale .666 .444 22.836 5,143 <.001* SOP .044 .553 SPP -.160 .057 PSP -.065 .540 N D P -.444 <.001* N D C -.125 .168 Predictor p p Total R R2 F df p Criterion Variable: T l State Self-Esteem Scale, Appearance Subscale .546 .298 12.161 5,143 <.001* SOP .229 .006* SPP -.258 .007* PSP -.079 .507 N D P -.271 .007* N D C -.098 .335 Note. The following labels were used: SOP (self-oriented perfectionism), SPP (socially prescribed perfectionism), PSP (perfectionistic self-promotion), N D P (nondisplay of imperfection), N D C (nondisclosure of imperfection). * Significant after multistage Bonferroni correction. 108 Table 4 Summary of Hierarchical Regression Analysis for Perfectionism Predicting Unique Variance in Time 1 Anorexia Nervosa Symptoms, Controlling for Big Five Personality Traits Predictor (3 p Total R AR2 AF df p Criterion Variable: T I Total Amount of Food Eaten (grams) S t ep l .218 .047 1.422 5,143 .220 Extraversion .199 .025* Agreeableness .031 .726 Conscientiousness .013 .872 Emotional Stability -.135 .143 Openness .039 .649 P 2 SOP .110 .337 SPP .089 .447 PSP -.221 .127 N D P .080 .544 N D C -.003 .978 .263 .022 .649 5, 138 .662 109 Predictor B p Total R AR2 AF df Criterion Variable: T l Square Root of Profile of M o o d States, Depression Subscale Step 1 .563 .317 13.249 5,143 <.001* Extraversion -.046 .537 Agreeableness -.111 .145 Conscientiousness .010 .884 Emotional Stability -.499 <.001* Openness .015 .836 Step 2 SOP -.029 .763 SPP .135 .169 PSP -.103 .396 N D P -.069 .536 N D C .128 .215 .585 .343 1.096 5,138 .366 Criterion Variable: T l Mizes Anorexic Cognitions Scale, Self-Control Subscale s t e p i .300 .090 2.829 5, 143 .018* Extraversion .019 .823 Agreeableness -.017 .845 Conscientiousness .100 .223 Emotional Stability -.284 .002* Openness -.043 .606 110 Predictor P Total R AR2 AF df p Step 2 .418 .084 2.818 5,138 .019* SOP -.108 .319 SPP -.053 .630 PSP .294 .032* N D P .118 .342 N D C .014 .900 Criterion Variable: T I Mizes Anorexic Cognitions Scale, Rig id Weight Regulation Subscale Step 1 .403 .162 5.538 5,143 <001* Extraversion -.064 .442 Agreeableness -.010 .901 Conscientiousness .018 .817 Emotional Stability -.336 <.001* Openness -.130 .103 Step 2 .608 .208 9.093* 5,138 <.001* SOP -.117 .214 SPP .047 .621 PSP .372 .002* N D P .012 .910 N D C .198 .050 I l l Predictor B p Total R AR2 A F df p Criterion Variable: T l Mizes Anorexic Cognitions Scale, Weight and Approval Subscale S t e p l .452 .204 7.340 5,143 <.001* Extraversion -.092 .255 Agreeableness -.277 .001* Conscientiousness .060 .432 Emotional Stability -.200 .018* Openness -.109 .162 ;p 2 .597 SOP -.148 .122 SPP .052 .592 PSP .438 <.001* N D P -.084 .443 N D C .105 .303 .152 6.539 5,138 <.001 ! Criterion Variable: T l State Self-Esteem Scale, Social Subscale S t ep l .579 .336 14.455 5,143 <.001* Extraversion .233 .002* Agreeableness .015 .842 Conscientiousness .124 .078 Emotional Stability .439 <.001* Openness -.059 .407 112 Predictor P P Total R AR2 AF df P Step 2 .726 .191 11.141 5,138 <.001* SOP .084 .306 SPP -.135 .107 PSP -.123 .232 N D P -.301 .002* N D C -.097 .269 Criterion Variable: T l State Self-Esteem Scale, Appearance Subscale Step 1 .583 .340 14.752 5, 143 <.001* Extraversion .175 .018* Agreeableness .249 .001* Conscientiousness .213 .003* Emotional Stability .244 .002* Openness .087 .219 Step 2 .644 .075 3.514 5,138 .005* SOP .169 .065 SPP -.169 .070 PSP -.050 .665 N D P -.160 .126 N D C -.055 .569 Note. The following labels were used: SOP (self-oriented perfectionism), SPP (socially prescribed perfectionism), PSP (perfectionistic self-promotion), N D P (nondisplay of imperfection), N D C (nondisclosure of imperfection). * Significant after multistage Bonferroni correction. 114 Table 5 Summary of Hierarchical Regression Analyses for Social Feedback Condition Predicting Time 2 Anorexia Nervosa Symptoms, Controlling for Time 1 Anorexia Nervosa Symptoms Predictor P p Total R AR2 AF df p Criterion Variable: T2 Amount of Food Eaten S t e p l .441 .195 29.043 1,120 <.001 TI Amount of .449 <.001 Food Eaten Step 2 .453 .010 .746 2,118 .477 A v s C .103 .274 E v s C .092 .324 A vs E T .003 .974 Criterion Variable: T2 Square Root of Profile of M o o d States, Depressed Mood Subscale S t ep l - .359 .129 17.697 1,120 <.001 TI P O M S - S F .359 <.001 Depressed M o o d (sqrt) Step 2 A v s C -.235 .011 E v s C .158 .085 A v s E f -.406 <.001* .494 .116 9.038 2,118 <.001* 115 Predictor B p Total R AR2 A F df Step 1 Criterion Variable: T2 Mizes Anorexic Cognitions Scale, Self-Control Subscale .899 .808 503.988 1,120 <.001 T l M A C - R Self- .899 <.001 Control Step 2 -899 .001 .228 2,118 .796 A v s C -.031 .501 E v s C -.016 .723 A v s E t -.014 .782 Criterion Variable: T2 Mizes Anorexic Cognitions Scale, Rig id Weight Regulation Subscale S t ep l .830 .689 265.771 1,120 <.001 T l M A C - R .830 <.001 Rigid Weight Regulation Step 2 -835 .009 1.659 2,118 .195 A vs C -.054 .347 E v s C .053 .359 A v s E t -.111 .072 Criterion Variable: T2 Mizes Anorexic Cognitions Scale, Weight and Approval Subscale - I .793 .629 203.424 1,120 <.001 T l M A C - R .793 <.001 Weight and Approval 116 Predictor p p T o t a l s A i ? 2 A F df p Step 2 .799 .010 1.630 2,118 .200 A vs C .022 .722 E v s C .109 .087 A vs E t -.094 .161 Criterion Variable: T2 State Self-Esteem Scale, Social Subscale Step 1 .752 .566 156.443 1, 120 <.001 TI SSES Social .752 <.001 Step 2 .759 .010 1.433 2,118 .243 A v s C .102 .138 E v s C .000 .999 A v s E 1 .102 .165 Criterion Variable: T2 State Self-Esteem Scale, Appearance Subscale S t e p l .790 .624 199.259 1,120 <.001 TI SSES .790 <.001 Appearance Step 2 .801 .017 2.797 2,118 .065 A v s C .138 .030 E v s C .016 .800 A v s E T .120 .074 117 Note. The following labels were used: A (Accepted condition), E (Excluded condition), C (Control condition), T l (Time 1), T2 (Time 2), P O M S - S F (Profile of M o o d States - Short Form), M A C - R (Mizes Anorexic Cognitions Scale - Revised) and SSES (State Self-Esteem Scale). * Significant after multistage Bonferroni correction. 1 To compare the accepted group with the excluded group, we compared BA to 5 £ b y calculating relevant t-scores. 118 Table 6 Summary of Hierarchical Regression Analyses for Perfectionism, Social Feedback Condition and Perfectionism x Social Feedback Condition Predicting Time 2 Quantity of Food Eaten, After Controlling for Time 1 Quantity of Food Eaten Predictor (3 p Total R AR2 AF df p Analysis 1: Self-Oriented Perfectionism Step 1 .441 .195 29.043 1, 120 <.001 T I Amount of .441 <.001 Food Eaten Step 2 .443 .002 .259 1,119 .612 SOP -.042 .612 Step 3 .455 .010 .740 2,117 .479 A v s C .103 .274 E v s C .092 .328 A vs E T .004 .968 Step 4 .460 .005 .334 2,115 .716 SOP x ( A v s C ) -.082 .470 SOP x (E vs C) -.002 .989 SOP x (A vs E ) f -.084 .487 119 Predictor P p Total R AR2 AF df P Analysis 2: Socially Prescribed Perfectionism Step 1 .441 .195 29.043 1,120 <.001 T l Amount of .441 <.001 Food Eaten Step 2 .447 .005 .747 1, 119 .389 SPP -.071 .389 Step 3 .459 .011 .812 2, 117 .447 A v s C .105 .263 E v s C .099 .291 A vs E* -.002 .987 Step 4 .494 .033 2.515 2, 115 .085* SPP x (A vs C) -.108 .364 SPP x (E vs C) -.231 .027 SPP x (A vs E ) t .208 .136 Analysis 3: Perfectionistic Self-Promotion Step 1 .441 .195 29.043 1, 120 <.001 T l Amount of Food Eaten .441 <.001 Step 2 .449 .007 1.042 1, 119 .310 PSP -.084 .310 120 Predictor B p T o t a l s AR2 AF df p S t e p 3 .460 .010 .718 2,117 .490 A vs C .099 .292 E v s C .093 .321 A v s E 1 -.001 .991 Step 4 .462 .002 .145 2,115 .865 PSP x (A vs C) -.052 .704 PSP x (E vs C) -.065 .596 PSP x (A vs Ef .026 .840 Analysis 4: Nondisplay of Imperfection S t ep l .441 .195 29.043 1,120 <.001 T l Amount of .441 <.001 Food Eaten Step 2 .453 .011 1.617 1,119 .206 N D P -.104 .206 Step 3 .464 .009 .707 2,117 .495 A v s C .098 .295 E vs C .092 .325 A v s E 1 -.001 .994 121 Predictor P p Total R AR2 AF df P Step 4 .466 .002 .179 2, 115 .836 N D P x (A vs C) .043 .746 N D P x (E vs C) -.028 .807 N D P x (A vs E ) T .080 .558 Analysis 5: Nondisclosure of Imperfection Step 1 .441 .195 29.043 1, 120 <.001 TI Amount of .441 <.001 Food Eaten Step 2 .467 .023 3.540 1, 119 .062 N D C -.153 .062 Step 3 .478 .011 .817 2, 117 .444 A v s C .103 .268 E v s C .100 .281 A vs E1" -.005 .963 Step 4 .495 .016 1.226 2, 115 .297 N D C x (A vs C) -.064 .597 N D C x (E vs C) -.163 .124 N D C x (A vs E ) T .155 .257 Note. The following labels were used: SOP (self-oriented perfectionism), SPP (socially prescribed perfectionism), PSP (perfectionistic self-promotion), N D P (nondisplay of 122 imperfection), N D C (nondisclosure of imperfection), A (Accepted condition), E (Excluded condition), C (Control condition), T I (Time 1). * Significant after multistage Bonferroni correction. * To compare the accepted group with the excluded group, we compared BA to Bghy calculating relevant t-scores. 123 Table 7 Summary of Hierarchical Regression Analyses for Perfectionism, Social Feedback Condition and Perfectionism x Social Feedback Condition Predicting Time 2 POMS-SF Depressed Mood, After Controlling for Time 1 POMS-SF Depressed Mood Predictor p p Total/? A i ? 2 AF df p Analysis 1: Self-Oriented Perfectionism S t ep l .359 .129 17.697 1,120 <.001 TI P O M S - S F .359 <.001 Depressed M o o d (sqrt) Step 2 .372 .010 1.350 1,119 .248 SOP -.100 .248 Step 3 .504 .115 9.037 2,117 <.001 A v s C -.237 .011 E v s C .156 .090 A v s E T -.404 <.001 Step 4 .504 .001 .059 2,115 .943 SOP x ( A v s C ) .021 .850 SOP x (E vs C) .037 .734 S O P x ( A v s E ) 1 " -.017 .884 124 Predictor P p Total R A i ? 2 A F df P Analysis 2: Socially Prescribed Perfectionism Step 1 .359 .129 17.697 1, 120 <.001 T l P O M S - S F Depressed M o o d (sqrt) .359 <.001 Step 2 .361 .002 .260 1, 119 .611 SPP .046 .611 Step 3 .495 .115 8.874 2, 117 <.001 A v s C -.235 .012 E v s C .157 .090 A vs E T -.404 <.001 Step 4 .501 .006 .453 2, 115 .637 SPP x (A vs C) -.111 .347 SPP x (E vs C) -.032 .753 S P P x ( A vs E ) 1 -.067 .626 Analysis 3: Perfectionistic Self-Promotion Step 1 .359 .129 17.697 1, 120 <.001 T l P O M S - S F Depressed M o o d (sqrt) .359 <.001 Step 2 .375 .012 1.656 1,119 .201 PSP .114 .201 125 Predictor P P Total R AR2 A F df P Step 3 .503 .112 8.800 2, 117 <.001 A v s C -.229 .014 E v s C .160 .082 A vs E1" -.401 <.001 Step 4 .518 .015 1.187 2, 115 .309 PSP x (A vs C) -.155 .240 PSP x (E vs C) .013 .912 P S P x ( A v s E ) T -.171 .173 Analysis 4: Nondisplay of Imperfection Step 1 .359 .129 17.697 1, 120 <.001 TI P O M S - S F Depressed M o o d (sqrt) .359 <.001 Step 2 .374 .012 .1.597 1, 119 .209 N D P .112 .209 Step 3 .503 .113 8.852 2, 117 <.001 A v s C -.229 .014 E v s C .161 .080 A vs E f -.402 <.001 126 Predictor P Total i? A i ?2 AF df P Step 4 .541 .040 3.239 2, 115 .043 N D P x (A vs C) -.228 .075 N D P x (E vs C) .057 .602 N D P x (A vs E ) 1 -.303 .021 Analysis 5: Nondisclosure of Imperfection Step 1 .359 .129 17.697 1, 120 <.001 T l P O M S - S F Depressed M o o d (sqrt) .359 <.00T Step 2 .383 .018 2.552 1, 119 .113 N D C .141 .113 Step 3 .509 .112 8.862 2, 117 <.001 A v s C -.232 .012 E v s C .155 .090 A vs E1" -.400 <.001 Step 4 .513 .004 .350 2, 115 .706 N D C x (A vs C) -.100 .405 N D C x (E vs C) .043 .683 N D C x (A vs E ) f -.044 .751 Note. The following labels were used: P O M S - S F (Profile of M o o d States - Short Form), SOP (self-oriented perfectionism), SPP (socially prescribed perfectionism), PSP (perfectionistic self-127 promotion), N D P (nondisplay of imperfection), N D C (nondisclosure of imperfection), A (Accepted condition), E (Excluded condition), C (Control condition), T I (Time 1), sq rt (square root). * Significant after multistage Bonferroni correction. T To compare the accepted group with the excluded group, we compared BA to 2?£by calculating relevant t-scores. 128 Table 8 Summary of Hierarchical Regression Analyses for Perfectionism, Social Feedback Condition and Perfectionism x Social Feedback Condition Predicting Time 2 MAC-R Self-Control Subscale, After Controlling for Time 1 MAC-R Self-Control Subscale Predictor P p Total R AR2 AF df P Analysis 1: Self-Oriented Perfectionism Step 1 .899 .808 503.988 1, 120 <.001 T l M A C - R Self-Control .899 <001 Step 2 .901 .004 2.712 1, 119 .102 SOP .066 .102 Step 3 .901 .001 .215 2, 117 .807 A v s C -.030 .513 E v s C -.015 .743 A vs -.014 .775 Step 4 .903 .004 1.138 2, 115 .324 SOP x (A vs C) -.079 .153 SOP x (E vs C) -.014 .793 SOP x (A vs E ) f -.064 .269 Analysis 2: Socially Prescribed Perfectionism Step 1 .899 .808 503.988 1, 120 <001 T l M A C - R Self- .899 <.001 Control 129 Predictor p p Total R AR2 A F df p Step 2 .899 .001 .560 1, 119 .456 SPP .030 .456 Step 3 .900 .001 .241 2, 117 .786 A v s C -.032 .493 E v s C -.019 .682 A vs E T -.011 .817 Step 4 .900 .001 .249 2, 115 .780 SPP x (A vs C) -.033 .573 SPP x (E vs C) -.007 .898 S P P x f A v s E ) T -.042 .539 Analysis 3: Perfectionistic Self-Promotion Step 1 .899 .808 503.988 1, 120 <.001 TI M A C - R Self-Control .899 <.001 Step 2 .905 .011 7.009 1, 119 .009* PSP .110 .009* Step 3 .905 .000 .135 2, 117 .874 A v s C -.023 .614 E v s C -.015 .730 A vs E^ -.006 .898 130 Predictor B p Totals AR2 AF df Step 4 -905 .000 .071 2,115 .931 PSP x (A vs C) -.023 .723 PSP x ( E v s C ) -.006 .917 P S P x ( A v s E ) T -.016 .801 Analysis 4: Nondisplay of Imperfection S t ep l .899 .808 503.988 1,120 <.001 TI M A C - R Self- .899 <.001 Control Step 2 .900 .002 1.161 1,119 .284 N D P .044 .284 Step 3 -900 .001 .188 2,117 .829 A vs C -.028 .543 E v s C -.016 .734 A v s E 1 -.011 .817 Step 4 .900 .000 .008 2,115 .993 N D P x (A vs C) .003 .959 N D P x (E vs C) .007 .904 N D P x ( A v s E ) T -.006 .933 131 Predictor P p Total i? A i ? 2 A F df P Analysis 5: Nondisclosure of Imperfection Step 1 .899 .808 503.988 1, 120 <.001 T l M A C - R Self-Control .899 <001 Step 2 .903 .008 5.167 1, 119 .025* N D C .091 .025* Step 3 .904 .001 .231 2, 117 .794 A v s C -.030 .509 E v s C -.020 .657 A vs E^ -.008 .861 Step 4 .904 .001 .383 2, 115 .682 N D C x (A vs C) .046 .435 N D C x (E vs C) .001 .983 N D C x (A vs E ) T .045 .502 Note. The following labels were used: M A C - R (Mizes Anorexic Cognition - Revised), SOP (self-oriented perfectionism), SPP (socially prescribed perfectionism), PSP (perfectionistic self-promotion), N D P (nondisplay of imperfection), N D C (nondisclosure of imperfection), A (Accepted condition), E (Excluded condition), C (Control condition), T l (Time 1). * Significant after multistage Bonferroni correction. T To compare the accepted group with the excluded group, we compared BA to BEby calculating relevant t-scores. 132 Table 9 Summary of Hierarchical Regression Analyses for Perfectionism, Social Feedback Condition and Perfectionism x Social Feedback Condition Predicting Time 2 MAC-R Rigid Weight Regulation Subscale, After Controlling for Time 1 MAC-R Rigid Weight Regulation Subscale Predictor P p Total R Air AF df Step 1 Analysis 1: Self-Oriented Perfectionism .830 .689 265.771 1, 120 <.001 T l M A C - R Rig id .830 <.001 Weight Regulation Step 2 SOP -.014 .794 Step 3 A v s C -.054 .349 E v s C .053 .362 A v s E f -.111 .074 Step 4 SOP x ( A vs C) .027 .705 SOP x (E vs C) .080 .246 S O P x ( A v s E ) f -.056 .442 .830 .000 .069 1,119 .794 .835 .008 1.638 2,117 .199 .837 .004 .696 2,115 .501 133 Predictor B p Total R AR2 AF df p Analysis 2: Socially Prescribed Perfectionism Step 1 .830 .689 265.771 1, 120 <.001 TI M A C - R Rigid .830 <.001 Weight Regulation Step 2 .833 .005 2.056 1,119 .154 SPP .075 .154 Step 3 .838 .008 1.568 2,117 .213 A v s C -.058 .318 E vs C .046 .422 A v s E T -.107 .082 Step 4 .838 .001 .111 2,115 .895 SPP x ( A v s C ) .008 .911 SPP x ( E v s C ) -.023 .719 SPP x (A vs E ) T .040 .644 Analysis 3: Perfectionistic Self-Promotion Step 1 .830 .689 265.771 1, 120 <.001 TI M A C - R Rig id .830 <.001 Weight Regulation Step 2 .834 .007 2.662 1,119 .105 PSP .096 .105 134 Predictor 6 p Totals AR2 AF df p Step 3 .839 .008 1.597 2,117 .207 A v s C -.053 .352 E v s C .051 .371 A v s E t -.109 .077 Step4 .847 .013 2.640 2,115 .076 PSP x ( A v s C ) .018 .821 PSP x ( E v s C ) .146 .049 P S P x ( A v s E ) 1 " -.157 .046 Analysis 4: Nondisplay of Imperfection Step 1 .830 .689 265.771 1, 120 <.001 T l M A C - R Rig id .830 <.001 Weight Regulation Step 2 .832 .003 1.254 1,119 .265 N D P .062 .265 Step 3 .837 .008 1.621 2,117 .202 A v s C -.053 .356 E vs C .053 .360 A v s E f -.110 .075 135 Predictor P p Total R AR2 AF df P Step 4 .838 .001 .217 2, 115 .805 N D P x (A vs C) .014 .869 N D P x (E vs C) .044 .533 N D P x (A vs E) f -.043 .599 Analysis 5: Nondisclosure of Imperfection Step 1 .830 .689 265.771 1, 120 <.001 T l M A C - R Rigid Weight Regulation .830 <.001 Step 2 .831 .001 .567 1, 119 .453 N D C .041 .453 Step 3 .836 .008 1.617 2, 117 .203 A v s C -.055 .339 E v s C .051 .380 A vs -.110 .076 Step 4 .837 .002 .438 2, 115 .647 N D C x (A vs C) .029 .707 N D C x (E vs C) .062 .353 N D C x f A vs E ) f -.054 .528 Note. The following labels were used: M A C - R (Mizes Anorexic Cognition - Revised), SOP (self-oriented perfectionism), SPP (socially prescribed perfectionism), PSP (perfectionistic self-136 promotion), N D P (nondisplay of imperfection), N D C (nondisclosure of imperfection), A (Accepted condition), E (Excluded condition), C (Control condition), T I (Time 1). * Significant after multistage Bonferroni correction. * To compare the accepted group with the excluded group, we compared BA to B f b y calculating relevant t-scores. 137 Table 10 Summary of Hierarchical Regression Analyses for Perfectionism, Social Feedback Condition and Perfectionism x Social Feedback Condition Predicting Time 2 MAC-R Weight and Approval Subscale, After Controlling for Time 1 MAC-R Weight and Approval Subscale Predictor B p Total R AR2 AF df Step 1 Analysis 1: Self-Oriented Perfectionism .793 .629 203.424 1, 120 <.001 TI M A C - R .793 <.001 Weight and Approval Step 2 SOP .056 .322 Step 3 A v s C .023 .721 E v s C .109 .085 A v s E T -.095 .159 Step 4 SOP x ( A vs C) .005 .953 SOP x ( E v s C ) .011 .889 SOP x (A vs E ) t -.006 .939 .795 .003 .988 1,119 .322 .801 .010 1.649 2,117 .197 .801 .000 .010 2,115 .990 138 Predictor P p Total i? A i ? 2 A F df P Analysis 2: Socially Prescribed Perfectionism Step 1 .793 .629 203.424 1, 120 <.001 TI M A C - R Weight and Approval .793 <.001 Step 2 .796 .005 1.475 1, 119 .227 SPP .070 .227 Step 3 .802 .009 1.457 2, 117 .237 A v s C .021 .742 E v s C .103 .106 A vs E^ -.090 .183 Step 4 .810 .014 2.378 2, 115 .097 SPP x (A vs C) .125 .123 SPP x (E vs C) -.047 .500 SPP x (A vs E ) 1 .189 .045 Analysis 3: Perfectionistic Self-Promotion Step 1 .793 .629 203.424 1, 120 <.001 TI M A C - R Weight and Approval .793 <.001 Step 2 .813 .032 11.228 1, 119 .001* PSP .207 .001* 139 Predictor P P Total R AR2 AF df P Step 3 .818 .009 1.554 2, 117 .216 A v s C .031 .606 E v s C .104 .086 A vs E1" -.081 .211 Step 4 .822 .006 1.145 2, 115 .322 PSP x (A vs C) .132 .134 PSP x (E vs C) .061 .441 P S P x ( A vs E ) 1 .059 .484 Analysis 4: Nondisplay of Imperfection Step 1 .793 .629 199.560 1, 120 <.001 TI M A C - R Weight and Approval .793 <.001 Step 2 .796 .004 .535 1, 119 .262 N D P .067 .262 Step 3 .802 .010 1.871 2, 117 .206 A v s C .025 .689 E v s C .108 .087 A vs E* -.091 .175 140 Predictor 3 p Total R AR2 A F df p Step 4 .802 .001 .078 2, 115 .854 N D P x ( A v s C ) .029 .743 N D P x (E vs C) .043 .576 N D P x (A vs E) T -.027 .766 Analysis 5: Nondisclosure of Imperfection Step 1 .793 .629 203.424 1, 120 <.001 T l M A C - R Weight and Approval Step 2 .793 <.00l .814 .034 11.817 1, 119 .001* N D C .196 .001* Step 3 .819 .008 1.413 2, 117 .248 A v s C .022 .718 E v s C . .098 .109 A vs E1" -.083 .198 Step 4 .833 .023 4.396 2, 115 .014 N D C x (A vs C) .225 .004 N D C x (E vs C) .063 .351 N D C x ( A vs E ) t .141 .108 Note. The following labels were used: M A C - R (Mizes Anorexic Cognition - Revised), SOP (self-oriented perfectionism), SPP (socially prescribed perfectionism), PSP (perfectionistic self-141 promotion), N D P (nondisplay of imperfection), N D C (nondisclosure of imperfection), A (Accepted condition), E (Excluded condition), C (Control condition), T l (Time 1). * Significant after multistage Bonferroni correction. ^ To compare the accepted group with the excluded group, we compared BA to BEhy calculating relevant t-scores. 142 Table 11 Summary of Hierarchical Regression Analyses for Perfectionism, Social Feedback Condition and Perfectionism x Social Feedback Condition Predicting Time SSES Social Subscale, After Controlling for Time 1 SSES Social Subscale Predictor P p T o t a l s AR2 AF df P Analysis 1: Self-Oriented Perfectionism Step 1 .752 .566 156.443 1, 120 <.001 T l SSES Social .752 <.001 Step 2 .753 .001 .178 1, 119 .673 SOP .026 .673 Step 3 .760 .010 1.430 2, 117 .243 A v s C .102 .139 E v s C .000 .999 A vs E1" .102 .166 Step 4 .767 .012 1.656 2, 115 .195 SOP x (A vs C) -.143 .083 SOP x (E vs C) -.100 .217 S O P x ( A vs E ) T -.039 .647 Step 1 Analysis 2: Socially Prescribed Perfectionism .752 .566 156.443 1, 120 <.001 T l SSES Social .752 <.001 1 143 Predictor P p Total A i ? 2 A F df P Step 2 .760 .012 3.496 1, 119 .064 SPP -.121 .064 Step 3 .767 .009 1.348 2, 117 .264 A v s C .104 .128 E v s C .014 .838 A vs E1" .089 .225 Step 4 •- .771 .007 .941 2, 115 .393 SPP x (A vs C) -.109 .211 SPP x (E vs C) -.079 .296 S P P x ( A vs E ) ! -.001 .994 Analysis 3: Perfectionistic Self-Promotion Step 1 .752 .566 156.443 1, 120 <.001 TI SSES Social .752 <001 Step 2 .771 .028 8.203 1, 119 .005* PSP -.192 .005* Step 3 .776 .008 1.110 2, 117 .333 A v s C .091 .172 E v s C .009 .895 A vs .082 .254 144 Predictor B p T o t a l s A i ? 2 A F df Step 4 P S P x ( A v s C ) -.132 .171 PSP x ( E v s C ) -.142 .102 P S P x ( A v s E ) T .038 .674 .782 .010 1.492 2,115 .229 Step 1 T l SSES Social .752 <.001 Step 2 N D P -.111 .140 Step 3 A v s C .096 .163 E vs C .004 .952 A v s E 1 .091 .216 Step 4 N D P x (A vs C) -.005 .962 N D P x (E vs C) .089 .285 N D P x ( A v s E ) t -.121 .222 Analysis 4: Nondisplay of Imperfection .752 .566 156.443 1, 120 <.001 .758 .008 2.213 1,119 .140 .763 .009 1.209 2,117 .302 .767 .006 .858 2,115 .427 Step 1 T l SSES Social .752 <.001 Analysis 5: Nondisclosure of Imperfection .752 .566 156.443 1,120 <.001 145 Predictor P p Total i? A i ? 2 A F df p Step 2 .765 .019 5.582 1,119 .020* N D C -.155 .020* Step 3 .771 .009 1.291 2,117 .279 A v s C .100 .137 E v s C .012 .853 A vs E1" .087 .230 Step 4 .774 .005 .689 2,115 .504 N D C x (A vs C) -.093 .293 N D C x (E vs C) -.073 .345 N D C x (A vs E ) T .004 .966 Note. The following labels were used: SSES (State Self-Esteem Scale), SOP (self-oriented perfectionism), SPP (socially prescribed perfectionism), PSP (perfectionistic self-promotion), N D P (nondisplay of imperfection), N D C (nondisclosure of imperfection), A (Accepted condition), E (Excluded condition), C (Control condition), T l (Time 1). * Significant after multistage Bonferroni correction. t To compare the accepted group with the excluded group, we compared BA to BEby calculating relevant t-scores. 146 Table 12 Summary of Hierarchical Regression Analyses for Perfectionism, Social Feedback Condition and Perfectionism x Social Feedback Condition Predicting Time 2 SSES Appearance Subscale, After Controlling for Time 1 SSES Appearance Subscale Predictor p p Total R AR2 AF df p Analysis 1: Self-Oriented Perfectionism S tep l .790 .624 199.259 1,120 <.001 TI SSES .790 <.001 Appearance Step 2 .791 .002 .607 1,119 .437 SOP -.044 .437 Step 3 .802 .017 2.797 2,117 .065 A v s C .138 .031 E v s C .015 .808 A v s E 1 .121 .073 Step 4 .812 .016 2.678 2,115 .073 SOP x (A vs C) -.028 .709 SOP x ( E v s C ) -.162 .030 S O P x ( A v s E ) r .139 .078 Analysis 2: Socially Prescribed Perfectionism Step 1 .790 .624 199.259 1, 120 <.001 TI SSES .790 <.001 Appearance 147 Predictor P P Total R AR2 A F df P Step 2 .792 .003 1.302 1,119 .312 SPP .061 .312 Step 3 .803 .017 2.744 2, 117 .068 A v s C .134 .035 E v s C .011 .865 A vs E^ .123 .069 Step 4 .807 .008 1.281 2, 115 .282 SPP x (A vs C) -.119 .140 SPP x (E vs C) -.009 .899 SPP x (A vs E ) f -.107 .255 Analysis 3: Perfectionistic Self-Promotion Step 1 .790 .624 199.259 1, 120 <.001 TI SSES Appearance .790 <.001 Step 2 .794 .007 2.203 1, 119 .140 PSP -.087 .140 Step 3 .805 .017 2.757 2, 117 .068 A v s C .136 .031 E v s C .017 .789 A vs E^ .118 .078 148 Predictor Total R AR2 AF df Step 1 Step 4 PSP x (A vs C) -.002 .983 PSP x (E vs C) -.005 .948 P S P x ( A v s E ) T .006 .958 .805 .000 .002 2,115 .998 Analysis 4: Nondisplay of Imperfection .790 .624 199.259 1,120 <.001 TI SSES .790 <.001 Appearance Step 2 N D P -.001 .981 Step 3 A v s C .138 .031 E v s C .016 .801 A v s E T .120 .076 Step 4 N D P x ( A vs C) -.069 .453 N D P x (E vs C) -.083 .284 N D P x (A vs E ) r .040 .664 .790 .000 .001 1,119 .981 .801 .017 2.773 2,117 .067 .803 .004 .599 2,115 .551 149 Predictor P p Total R AR2 A F df P Analysis 5: Nondisclosure of Imperfection Step 1 .790 .624 199.259 1, 120 <.001 T l SSES Appearance .802 <.001 Step 2 .802 .020 6.632 1, 119 .011* N D C -.141 .011* Step 3 .813 .017 3.022 2, 117 .053 A v s C .142 .022 E v s C .023 .703 A v s E 1 .117 .076 Step 4 .818 .007 1.227 2, 115 .297 N D C x (A vs C) -.102 .204 N D C x (E vs C) -.098 .164 N D C x ( A vs E ) f .029 .751 Note. The following labels were used: SSES (State Self-Esteem Scale), SOP (self-oriented perfectionism), SPP (socially prescribed perfectionism), PSP (perfectionistic self-promotion), N D P (nondisplay o f imperfection), N D C (nondisclosure of imperfection), A (Accepted condition), E (Excluded condition), C (Control condition), T l (Time 1). * Significant after multistage Bonferroni correction. 1 To compare the accepted group with the excluded group, we compared BA to i?£by calculating relevant t-scores. 150 Table 13 Simple Slope Regression Analyses of Significant Perfectionism x Social Feedback Interactions Predicting Time 2 Anorexia Nervosa Symptoms, Controlling for Time Anorexia Nervosa Symptoms Predictor P t for within-set P predictors Criterion: T2 Amount of Food Eaten SPP x (E vs C) SPP for Excluded -1.062 -2.252 .026 SPP for Control .271 .747 .457 Criterion: Square Root T2 Profile of Mood States-Short Form, Depressed Mood Subscale N D P x (A vs C) N D P for Accepted -.013 -.998 .320 N D P for Control .023 1.449 .150 N D P x (A vs E) N D P for Accepted -.013 -.998 .320 N D P for Excluded .035 2.169 .032 Predictor 3 t for within-set predictors Criterion: T2 Mizes Anorexic Cognitions Scale-Revised, Rig id Weight Regulation Subscale PSP x (E vs C) PSP for Excluded .125 2.795 .006 PSP for Control .002 .044 .965 PSP x (A vs E) PSP for Accepted .015 .407 .685 PSP for Excluded .125 2.795 .006 Criterion: T2 Mizes Anorexic Cognitions Scale-Revised, Weight and Approval Subscale N D C x (A vs C) N D C for Accepted .252 4.369 <001 N D C for Control .012 .193 .847 SPP for Accepted SPP for Excluded SPP x (A vs E) .082 -.039 2.272 -.801 .025 .425 152 Predictor B t for within-set p predictors Criterion: T2 State Self-Esteem Scale, Social Subscale SOP x (A vs C) SOP for Accepted -.030 -.851 .396 SOP for Control .053 1.651 .100 Criterion: T2 State Self-Esteem Scale, Appearance Subscale SOP x (E vs C) SOP for Excluded -.064 -2.357 .020 SOP for Control .017 .681 .497 SOP x (A vs E) SOP for Accepted .003 .120 .904 SOP for Excluded -.064 -2.357 .020 Note. The following labels were used: SPP (socially prescribed perfectionism), PSP (perfectionistic self-promotion), N D P (nondisplay of imperfection), N D C (nondisclosure of imperfection), A (Accepted condition), E (Excluded condition), C (Control condition), T2 (Time 2). 153 Figure 1 Proposed Mediated Moderation Model Perfectionism Rejected Affect Anorexia Nervosa Symptoms Social Feedback Condition 154 Figure 2 The Interaction of Socially Prescribed Perfectionism with Social Feedback to Predict Time 2 Amount of Food Eaten ••— Control Excluded Min Max SPPc 155 Figure 3 The Interaction of Nondisplay of Imperfection with Social Feedback to Predict Square Root of Time 2 POMS-SF Depressed Mood 2.5 -i 1.5 l 1 0.5 • Accepted • Control • Excluded Min Max NDPc 156 Figure 4 The Interaction of Perfectionistic Self-Promotion with Social Feedback to Predict Time 2 MAC-R Rigid Weight Regulation PSPc 157 Figure 5 The Interaction of Socially Prescribed Perfectionism with Social Feedback to Predict Time 2 MAC-R Weight and Approval ••— Accepted Excluded SPPc 158 Figure 6 The Interaction of Nondisclosure of Imperfection with Social Feedback to Predict Time 2 MAC-R Weight and Approval 159 Figure 7 The Interaction of Self-Oriented Perfectionism with Social Feedback to Predict Time 2 SSES Social Self-Esteem 1 6 0 Figure 8 The Interaction of Self-Oriented Perfectionism with Social Feedback to Predict Time 2 SSES Appearance Self-Esteem Min Max SOPc 161 Figure 9 Mediated Moderation Model: The Interaction of Perfectionistic Self-Promotion with Social Feedback to Predict Time 2 MAC-R Rigid Weight Regulation, Mediated by Rejected Affect Perfectionistic Self-Promotion Rejected Affect MAC-R Rigid Weight Regulation Social Feedback Condition (A vs E) 162 Figure 10 The Interaction of Perfectionistic Self-Promotion with Social Feedback to Predict Time 2 Rejected Affect PSPc 163 FOOTNOTES 1 The decision to screen participants on levels of anxiety and depression was based on ethical concern over the reaction of vulnerable participants to our social exclusion manipulation. By excluding participants with elevated levels of anxiety or depression, we may have simultaneously excluded the very participants in whom we were most interested, those highest on perfectionism. Therefore, we feel that this study represents a more stringent test of the model. The experimental manipulation was adapted from previous work (Twenge, Baumeister, Tice, & Stucke, 2001). We chose this manipulation because it had wide support in the literature as a strong social exclusion manipulation (e.g., Baumeister, DeWall, Ciarocco, & Twenge, 2005; Baumeister, Twenge, & Nuss, 2002) and was feasible given constraints on lab space and personnel. Although we initially used a manipulation that was identical to that used elsewhere, our pilot test results indicated that participants were disinclined to believe the feedback, particularly those in the social exclusion condition. Therefore we altered the manipulation in line with suggestions from our pilot test participants to enhance believability. We altered the wording of the personality description slightly, added false citations to the personality description, and provided participants with an official looking computer generated report of a personality measure. 3For the bivariate correlations of interest, the multistage Bonferroni correction was applied in a stepwise manner as follows. The correlations of interest were rank ordered from highest to lowest based on significance level. For the first correlation, the significance level was set at a corrected level of (cc/k), where k = the total number of correlations of interest, resulting in a significance level of .05/45 = .001. For the second correlation, the significance level was set at a corrected level of (a/k-1), the significance level for the third correlation was set at (ct/k-2), 164 and so on. If a correlation was found to be nonsignificant at any stage of the process, then all subsequent correlations in the rank ordered list were considered nonsignificant. Conducting a manipulation check on the unrestricted sample of 131 participants who completed session 2, without removing participants with a low self-reported belief in the social feedback, resulted in less differentiated groups in general, and a loss of the significant difference in feelings of rejection between the excluded and control groups. 165 REFERENCES Aiken, L. S., & West, S. G. (1991). Multiple Regression: Testing and interpreting interactions. Newbury Park: Sage Publications. Allen, K. M , Thombs, D. L., Mahoney, C. A., & Daniel, E. L. (1993). Relationships between expectancies and adolescent dieting behaviors. Journal of School Health, 63, 176-181. American Psychiatric Association. (2003). Diagnostic and statistical manual of mental disorders (4th ed., text revision). Washington, DC: Author. Andersson, P., & Perris, C. (2000). Attachment styles and dysfunctional assumptions in adults. Clinical Psychology and Psychotherapy, 7, 47-53. Arnau, R. C , Meagher, M. W., Norris, M. P., & Bramson, R. (2001). Psychometric evaluation of the Beck Depression Inventory-II with primary care medical patients. Health Psychology, 20, 112-119. Ashby, J. S., & Rice, K. G. (2002). Perfectionism, dysfunctional attitudes, and self-esteem: A srtuctural equations analysis. Journal of Counseling and Development, 80, 197-203. Atlas, J. G. (2004). Interpersonal sensitivity, eating disorder symptoms, and eating/thinness expectancies. Current Psychology, 22, 368-378. Badcock, P. B. T., & Allen, N. B. (2003). Adaptive social reasoning in depressed mood and depressive vulnerability. Cognition and Emotion, 17, 647-670. Bardone, A. M., Vohs, K. D., Abramson, L. Y., Heatherton, T. F., & Joiner, T. E. (2000). The confluence of perfectionism, body dissatisfaction, and low self-esteem predicts bulimic symptoms: Clinical implications. Behavior Therapy, 31, 265-280. Bastiani, A.M., Rao, R., Weltzin, T., & Kaye, W. H. (1995). Perfectionism in anorexia nervosa. International Journal of Eating Disorders, 17, 147-152. 166 Baumeister, R. F. , Campbell, J . D . , Krueger, J. I., & Vohs, K . D . (2003). Does high self-esteem cause better performance, interpersonal success, happiness, or healthier lifestyles?, Psychological Science in the Public Interest, 4, 1-45. Baumeister, R. F. , DeWal l , C . N . , Ciarocco, N . J. , & Twenge, J . M . (2005). Social exclusion impairs self-regulation. Journal of Personality and Social Psychology, 88, 589-604. Baumeister, R. F. , & Leary, M . R. (1995). The need to belong: Desire for interpersonal attachments as a fundamental human motivation, Psychological Bulletin, 117, 497-529. Baumeister, R. F. , Twenge, J. M . , & Nuss, C . K . (2002). Effects of social exclusion on cognitive processes: Anticipated aloneness reduces intelligent thought. Journal of Personality and Social Psychology, 83, 817-827. Beck, A . T., Epstein, N . , Brown, G . (1988). A n inventory for measuring clinical anxiety: Psychometric properties. Journal of Consulting and Clinical Psychology, 56, 893-897. Beck, A . T., Steer, R. A . , & Brown, G . K . (1996). Manual for the Beck Depression Inventory-II. San Antonio, T X : Psychological Corporation. Beevers, C. G . , & Mi l le r , I. W . (2004). Perfectionism, cognitive bias, and hopelessness as prospective predictors of suicidal ideation. Suicide and Life-Threatening Behavior, 34, 126-137. Belangee, S. E . , Sherman, M . F. , & Kern, R. M . (2003). Exploring the relationships between lifestyle, personality attributes and eating disorder symptoms and behaviors in a non-clinical population. Journal of Individual Psychology, 39, 461-474. Bhar, S. S., & Kyrios , M . (1999). Cognitive personality styles associated with depressive and obsessive compulsive phenomena in a nonclinical sample. Behavioural and Cognitive Psychotherapy, 27, 329-343. 167 Birmingham, C . L . , Su, J. , Hlynsky, J. A . , Goldner, E . M , & Gao, M . (2005). The mortality rate from anorexia nervosa. International Journal of Eating Disorders, 38, 143-146. Blankstein, K . , & Dunkley, D . (2002). Evaluative concerns, self-critical, and personal standards of perfectionism: A structural equation model strategy. In G . L . Flett & P. L Hewitt (Eds.), Perfectionism: Theory, research, and treatment (pp. 285-315). Washington: American Psychological Association. Braun, D. L . , Sunday, S. R., & Halmi , K . A . (1994). Psychiatric comorbidity in patients with eating disorders. Psychological Medicine, 24, 859-867. Bruch, H . (1973). Eating Disorders: Obesity, Anorexia Nervosa, and the Person Within. New York: Basic Books, Inc. Bruch, H . (1978). The Golden Cage: The Enigma of Anorexia Nervosa. Cambridge, M . A . : Harvard University Press. Bruch, H . (1988). Conversations with Anorexics. New York: Basic Books, Inc. Buckley, K . E . , Winkel , R. E . , & Leary, M . R. (2004). Reactions to acceptance and rejection: Effects of level and sequence of relational evaluation. Journal of Experimental Social Psychology, 40, 14-28. Bulik, C. M . , Sullivan, P. F. , Tozzi , F. , Furberg, H . , Lichtenstein, P., & Pedersen, N . L . (2006). Prevalence, heritability, and prospective risk factors for anorexia nervosa. Archives of General Psychiatry, 63, 305-312. Burns, D. (1980). The perfectionist's script for self-defeat. Psychology Today, November, 34-52. Butler, N . , Slade, P., & Newton, T. (1990). Attitudes toward anorexia nervosa and bulimic disorders: Expert and lay opinion. British Review of Bulimia and Anorexia Nervosa, 4, 61-69. 168 Button, E . (1983). Personal construct theory and psychological well-being. British Journal of Medical Psychology, 56, 313-321. Calam, R., & Waller, G . (1998). Are eating and psychosocial characteristics in early teenage years useful predictors of eating characteristics in early adulthood? A 7-year longitudinal study. International Journal of Eating Disorders, 24, 3 51 -362. Campbell, J. D . , & D i Paula, A . (2002). Perfectionistic self-beliefs: Their relation to personality and goal pursuit. In G . L . Flett & P. L . Hewitt (Eds.), Perfectionism: Theory, research, and treatment (pp. 181-198). Washington, D C : American Psychological Association. Cash, T. F. (1990). The psychology of physical appearance: Aesthetics, attributes, and images. In T. F. Cash & T. Pruzinsky (Eds.), Body images: Development, deviance, and change (pp. 51-79). N e w York: Guilford. Cash, T. F. (2002). Cognitive-behavioral perspectives on body image. In T. F . Cash & T. Pruzinsky (Eds.), Body image: A handbook of theory, research, & clinical practice (pp. 38-46). New York: Guilford Press. Cassin, S. E . , & von Ranson, K . M . (2005). Personality disorders and eating disorders: A decade in review. Clinical Psychology Review, 25, 895-916. Chang, E . C , & Rand, K . L . (2000). Perfectionism as a predictor of subsequent adjustment: Evidence for a specific diathesis-stress mechanism among college students. Journal of Counseling Psychology, 47, 129-137. Chang, E . C , & Sanna, L . J. (2001). Negative attribution style as a moderator of the link between perfectionism and depressive symptoms: Preliminary evidence for an integrative model. Journal of Counseling Psychology, 48, 490-495. Cockell , S. J., Hewitt, P. L . , Seal, B . , Sherry, S. B . , Goldner, E . M . , Flett, G . L . , et al. (2002). Trait and self-presentational dimensions of perfectionism among women with anorexia nervosa. Cognitive Therapy and Research, 26, 745-758. Cohen, J., Cohen, P., West, S. G . , & Aiken, L . S. (2003). Applied multiple regression/correlation analysis for the behavioral sciences (3 r d ed.). N e w Jersey: Lawrence Erlbaum Associates. Connan, F. , & Treasure, J. (1998). Stress, eating, and neurobiology. In H . W . Hoeck, J. L . Treasure, & M . Katzman (Eds.), Neurobiology in the treatment of eating disorders (pp. 211-236). Chichester: Wiley. Corte, C , & Stein, K . F. (2005). Body-weight self-schema: Determinant of mood and behavior in women with an eating disorder. Journal of Applied Social Psychology, 35, 1698-1718. Curran, S. L . , Andrykowski , M . A . , & Stduts, J. L . (1995). Short Form of the Profile of Mood States: Psychometric information. Psychological Assessment, 7, 80-83. Currin, L . , & Schmidt, U . (2005). A critical analysis o f the utility o f an early intervention approach in the eating disorders. Journal of Mental Health, 14, 611-624. Davis, C , Claridge, G . , & Fox, J . (2000). Not just a pretty face: Physical attractiveness and perfectionism in the risk for eating disorders. International Journal of Eating Disorders, 27, 67-73. Dunkley, D . M . , Blankstein, K . R., & Flett, G . L . (1997). Specific cognitive-personality vulnerability styles in depresssion and the five-factor model of personality. Personality and Individual Differences, 23, 1041-1053. Dunkley, D . M . , Blankstein, K . R., & Halsall, J., Will iams, M . & Winkworth, G . (2000).The relationship between perfectionism and distress: hassles, coping and perceived social support as mediators and moderators. Journal of Counseling Psychology, 47, 437-453. Dunkley, D . M . , Sanislow, C . A . , Gri lo , C . M . , & McGlashan, T. H . (2006). Perfectionism and depressive symptoms three years later: Negative social interactions, avoidant coping, and perceived social support as mediators. Comprehensive Psychiatry, 47, 106-115. Enns, M . W. , & Cox, B . J. (2005). Perfectionism, stressful life events, and the 1-year outcome o f depression. Cognitive Therapy and Research, 29, 541-553. Enns, M . W. , Cox, B . J., & Clara, I. (2002). Adaptive and maladaptive perfectionism: Developmental origins and association with depression proneness. Personality and Individual Differences, 33, 921-935. Enns, M . W. , Cox, B . J., & Clara, I. (2005). Perfectionism and neuroticism: A longitudinal study of specific vulnerability and diathesis-stress models. Cognitive Therapy and Research, 29, 463-478. Enns, M , W. , Cox, B . J. , & Inayataulla, M . (2003). Personality predictors for outcome for adolescents hospitalized for suicidal ideation. Journal of the American Academy of Child and Adolescent Psychiatry, 42, 720-727. Enns, M . W. , Cox, B . J., Sareen, J., & Freeman, P. (2001). Adaptive and maladaptive perfectionism in medical students: A longitudinal investigation. Medical Education, 35, 1034-1042. 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, 23-39. Fairburn, C. G . , Cooper, Z , D o l l , H . A . , & Welch, S. L . (1999). Risk factors for anorexia nervosa: Three integrated case-control comparisons. Archives of General Psychiatry, 56, 468-476. 171 Fallon, A . E . (1990). Culture in the mirror: Sociocultural determinants of body image. In T. F. Cash and T. Pruzinsky (Eds.), Body Images: Development, deviance, and change. N e w York: Guilford. Flett, G . L . , Besser, A . , Davis, R. A . , & Hewitt, P. L . (2003). Dimensions o f perfectionism, unconditional self-acceptance, and depression. Journal of Rational-Emotive and Cognitive-Behavior Therapy, 21, 119-138. Flett, G . L . , & Hewitt, P. L . (2002). Perfectionism and maladjustment: A n overview of theoretical, definitional, and treatment issues. In G . L . Flett & P. L . Hewitt (Eds.), Perfectionism: Theory, research, and treatment (pp. 5-31). Washington, D C : American Psychological Association. Flett, G . L . , Hewitt, P. L . , Blankstein, K . R., & Gray, L . (1998). Psychological distress and the frequency o f perfectionistic thinking. Journal of Personality and Social Psychology, 75, 1363-1381. Flett, G . L . , Hewitt, P. L . , Blankstein, K . R., & Mosher, S. W . (1995). Perfectionism, life events, and depressive symptoms: A test of a diathesis-stress model. Current Psychology, 14, 112-137. Flett, G . L . , Hewitt, P. L . , Blankstein, K . R., & O'Brien, S. (1991). Perfectionism and learned resourcefulness in depression and self-esteem. Personality and Individual Differences, 12, 61-68. Flett, G . L . , Hewitt, P. L . , Blankstein, K . R., & Pickering, D . (1998). Perfectionism in relation tc attributions for success or failure. Current Psychology, 17, 249-262. 172 Flett, G . L . , Hewitt, P. L . , Blankstein, K . R., Solnik, M . , & V a n Brunschot, M . (1996). Perfectionism, social problem-solving ability, and psychological distress. Journal of Rational-Emotive and Cognitive-Behavior Therapy, 14, 245-274. Flett, G . L . , Hewitt, P. L . , & DeRosa, T. (1996). Dimensions o f perfectionism, psychosocial adjustment, and social skills. Personality and Individual Differences, 20, 143-150. Flett, G . L . , Hewitt, P. L . , Garshowitz, M . , & Martin, T. R. (1997). Personality, negative social interactions, and depressive symptoms. Canadian Journal of Behavioral Science, 29,28-37. Flett, G . L . , Hewitt, P. L , Oliver, J. M . , & Macdonald, S. (2002). Perfectionism in children and their parents: A developmental analysis. In G . L . Flett & P. L . Hewitt (Eds.), Perfectionism: Theory, research, and treatment (pp. 89-132). Washington, D C : American Psychological Association. Flett, G . L . , Hewitt, P. L . , Shapiro, B . , & Rayman, J. (2001). Perfectionism, beliefs, and adjustment in dating relationships. Current Psychology, 20, 289-311. Flett, G . L . , Russo, F. A . , & Hewitt, P. L . (1994). Dimensions of perfectionism and constructive thinking as a coping response. Journal of Rational-Emotive and Cognitive-Behavior Therapy, 12, 163-179. Forbush, K . , Heatherton, T. F., & Keel , P. K . (2007). Relationships between perfectionism and specific disordered eating behaviors. International Journal of Eating Disorders, 40, 37-41. Frank, G . K . , Kaye, W . H . , Weltzin, T. E . , Perel, J. , Moss, H . , McConaha, et al. (2001). Altered response to meta-chlorophenylpiperazine in anorexia nervosa: Support for a persistent 173 alteration of serotonin activity after short-term weight restoration. International Journal of Eating Disorders, 30, 57-68. Franco-Paredes, K . , Mancil la-Diaz, J. M . , Vazquez-Arevalo, R., Lopez-Aguilar, X . , & Alvarez-Rayon, G . (2005). Perfectionism and eating disorders: A review of the literature. European Eating Disorders Review, 13, 61-70. Frijda, N . H . (1994). Emotions are functional, most of the time. In P. Ekman & R. Davidson (Eds.), The nature of emotion: Fundamental questions (pp. 97-177). N e w York: Oxford University Press. Frost, R., Marten, P., Lahart, C . , & Rosenblate, R. (1990). The dimensions of perfectionism. Cognitive Therapy and Research, 14, 449-468. Gardner, W . L . , Pickett, C . L . , & Brewer, M . B . (2000). Social exclusion and selective memory: How the need to belong influences memory for social events. Personality and Social Psychology Bulletin, 26, 486-496. Garner, D . M . , & Bemis, K . M . (1982). A cognitive-behavioral approach to anorexia nervosa. Cognitive Therapy and Research, 6, 123-150. Garner, D . M . , Olmstead, M . P., Polivy, J., & Garfinkel, P. E . (1983). Development and validation of a multidimensional eating disorder inventory for anorexia nervosa and bulimia. International Journal of Eating Disorders, 2, 15-34. Geller, J., Cockell , S. J., Hewitt, P. L . , Goldner, E . M . , & Flett, G . L . (2000). Inhibited expression of negative emotions and interpersonal orientation in anorexia nervosa. International Journal of Eating Disorders, 28, 8-19. 174 Gerner, B . , & Wilson, P. H . (2005). The relationship between friendship factors and adolescent girls' body image concern, body dissatisfaction, and restrained eating. International Journal of Eating Disorders, 37, 313-320. Gilbert, P. (1992). Depression: The evolution ofpowerlessness. N e w York: Guilford. Gilbert, P. (1997). The evolution of social attractiveness and its role in shame, humiliation, guilt and therapy. British Journal of Medical Psychology, 70, 113-147. Graham, M . A . , E ich , C , Kephart, B . , & Peterson, D . (2000). Relationship among body image, sex, and popularity of high school students. Perceptual and Motor Skills, 90, 1187-1193. Green, S.B. (1991). H o w many subjects does it take to do a regression analysis? Multivariate Behavioral Research, 26, 449-510. Habke, A . M . , Hewitt, P. L . , & Flett, G . L . (1999). Perfectionism and sexual satisfaction in intimate relationships. Journal of Psychopathology and Behavior Assessment, 21 ^  307-322. Hamachek, D . E . (1978). Psychodynamics of normal and neurotic perfectionism. Psychology, 15, 27-33. Hamburg, D . A . (1974). Coping behavior in life threatening circumstances. Psychotherapy and Psychosomatics, 23, 13-25. Halmi, K . A . , Goldberg, S., Eckert, E . , Casper, R., & Davis, J. (1977). Pretreatment evaluation in anorexia nervosa. In R. A . Vigersky (Ed.), Anorexia Nervosa. N e w York: Raven Press. Halmi, K . A . , Sunday, S. R., Strober, M . , Kaplan, A . , Woodside, D . B . , Fichter, M . et al., (2000). Perfectionism in anorexia nervosa: Variation by clinical subtype, obsessionality, and pathological eating behavior. American Journal of Psychiatry, 157, 1799-1805. Haring, M . , Hewitt, P. L . , & Flett, G . L . (2003). Perfectionism, coping, and quality of intimate 175 relationships. Journal of Marriage and the Family, 65, 143-158. Heatherton, T. F., & Baumeister, R. F. (1991). Binge eating as an escape from self-awareness. Psychological Bulletin, 110, 86-108. Heatherton, T. F. , & Polivy, J. (1991). Development and validation of a scale for measuring state self-esteem. Journal of Personality and Social Psychology, 60, 895-910. Heatherton, T. F. , & Polivy, J. (1992). Chronic dieting and eating disorders: A spiral model. In J. H . Crowther, D . L . Tennenbaum, S. E . , Hobfoll , & M . A . P. Stephens (Eds.), The Etiology of Bulimia Nervosa: The Individual and Familial Context, pp. 135-155. Washington: Hemisphere. Heatherton, T. F. , Nichols, P., Mahamedi, F. , & Keel , P. (1995). Body weight, dieting, and eating disorder symptoms among college students, 1982 to 1992. American Journal of Psychiatry, 152, 1623-1629. Hebebrand, J. , Casper, R., Treasure, J. , & Schweiger, U . (2004). The need to revise the diagnostic criteria for anorexia nervosa. Journal of Neural Transmission, 111, 827-840. Hewitt, P. L . , & Flett, G . L . (1991a). Dimensions o f perfectionism in unipolar depression. Journal of Abnormal 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 . (1993a). Dimensions of perfectionism, daily stress, and depression: A test of the specific vulnerability hypothesis. Journal of Abnormal Psychology, 102, 58-65. Hewitt, P. L . & Flett, G . L . (1993b). Perfectionism and goal orientation in impulsive and suicidal behavior. In W . G . M c C o w n , J. L . Johnson, & M . B . Shure (Eds.) The impulsive client: Theory, research, and treatment, (pp. 247-263). Washington, D C : American Psychological Association. Hewitt, P. L . & Flett, G . L . (2002). Perfectionism and stress processes in psychopathology. In P. L . Hewitt & G. L . Flett (Eds.) Perfectionism: Theory, research, and treatment, (pp. 255-284). Washington, D C : American Psychological Association. Hewitt, P. L . , Flett, G . L . , & Ediger, E . (1995). Perfectionism traits and perfectionistic self-presentation in eating disorder attitudes, characteristics, and symptoms. International Journal of Eating Disorders, 18, 317-326. Hewitt, P. L . , Flett, G . L . , & Ediger, E . (1996). Perfectionism and depression: Longitudinal assessment of a specific vulnerability hypothesis. Journal of Abnormal Psychology, 105, 276-280. Hewitt, P. L . , Flett, G . L . , & M i k a i l , S. F. (1995). Perfectionism and family adjustment in pain patients and their spouses. Journal of Family Psychology, 9, 335-347. Hewitt, P. L . , Flett, G . L . , Sherry, S. B . , & Caelian, C . (2006). Trait perfectionism dimensions and suicide behavior. In T. El l i s (Ed.), Cognition and suicide: Theory, research, and practice, (pp. 215-235 ). Washington, D C : American Psychological Association. Hewitt, P. L . , Flett, G . L . , Sherry, S. B . , Habke, M . , Parkin, M . , Lam, R. W. , McMurtry , B . , Ediger, E . , Fairlie, & Stein, M . (2003). The interpersonal expression o f perfectionism: Perfectionistic self-presentation and psychological distress. Journal of Personality and Social Psychology, 84, 1303-1325. 177 Hewitt, P. L . , Flett, G . L . , Turnbull-Donovan, W. , & M i k a i l , S. F . (1991). The Multidimensional Perfectionism Scale: Reliability, validity, and psychometric properties in a psychiatric sample. Psychological Assessment, 3, 464-468. Hewitt, P. L . , Habke, A . M . , Lee-Baggley, D . , Sherry, S. B . , & Flett, G . L . (2006). Perfectionistic self presentation and cognitive, affective, and physiological reactions to clinical interventions. Manuscript submitted for publication. H i l l , R. W. , Huelsman, T. J. , Furr, R. M . , Kibler, J., Vicente, B . B . , & Kennedy, C. (2004). A new measure of perfectionism: The perfectionism inventory. Journal of Personality Assessment, 82, 80-91. H i l l , R. W. , Mclntire, K . , & Bacharach, V . R. (1997). Perfectionism and the big five factors. Journal of Social Behavior and Personality, 12, 257-270. Hoek, H . W . (2006). Incidence, prevalence and mortality o f anorexia nervosa and other eating disorders. Current Opinions in Psychiatry, 19, 389-394. Humphrey, L . L . (1992). Family relationships. In K . Halmi (Ed), Psychobiology and treatment of anorexia nervosa and bulimia nervosa, (pp. 263-284). American Psychiatric Press: Washington, D C . John, O. P., & Srivastava, S. (1999). The B i g Five trait taxonomy: History, measurement, and theoretical perspectives. In Pervin, L . A . , & John, O. P. (Eds.), Handbook of Personality: Theory and Research (2 n d ed, pp. 102-138). New York, N Y : Guilford Press. Johnson, C. & Connors, M . (1987). The etiology and treatment of bulimia nervosa: A biopsychosocialperspective. N e w York: Basic Books. 178 Joiner, T. E . , Heatherton, T. F. , Rudd, M . D . , & Schmidt, N . B . (1997). Perfectionism, perceived weight status, and bulimic symptoms: Two studies testing a diathesis-stress model. Journal of Abnormal Psychology, 106, 145-153. Joiner, T. E . , & Schmidt, N . B . (1995). Dimensions of perfectionism, life stress, and depressed and anxious symptoms: Prospective support for diathesis-stress but not specific vulnerability among male undergraduates. Journal of Social and Clinical Psychology, 14, 165-183. Juster, H . R., Heimberg, R. G . , Frost, R. O., & Holt, C . S. (1996). Social phobia and perfectionism. Personality and Individual Differences, 21, 403-410. Kaye, W . (1997). Anorexia nervosa, obsessional behavior, and serotonin. Psychopharmacology Bulletin, 33, 335-344. Kaye, W . H . , Barbarich, N . C , Putnam, K . , Gendall, K . A . , Fernstrom, J., Fernstrom, M . , et al. (2003). Anxiolyt ic effects of acute tryptophan depletion in anorexia nervosa. International Journal of Eating Disorders, 33, 257-267. Kaye, W. , Gendall, K . , & Strober, M . (1998). Serotonin neuronal function and selective serotonin reuptake inhibitor treatment in anorexia and bulimia nervosa. Biological Psychiatry, 44, 825-838. Kaye, W . H . , Gwirtsman, H . E . , George, D . T., & Ebert. M . H . (1991). Altered serotonin activity in anorexia nervosa after long-term weight restoration: Does elevated C S F 5 - H I A A correlate with rigid and obsessive behavior? Archives if General Psychiatry, 48, 556-562. Kaye, W. , Strober, M . , & Jimerson, D . (2004). The neurobiology of eating disorders (2 n d Ed.). In D.S. Charney & E.J . Nestler (Eds.) The neurobiology of mental illness (pp. 1112-1147). New York: Oxford Press. 179 Keel , P. K . & Klump, K . L . (2003). Are eating disorders culture-bound syndromes? Implications for conceptualizing their genetic bases. Psychological Bulletin, 129, 747-769. Kernis, M . H . (2005). Measuring self-esteem in context: The importance of stability of self-esteem in psychological functioning. Journal of Personality, 73, 1569-1605. Klibert, J. J., Langhinrichsen-Rohling, J., & Saito, M . (2005). Adaptive and maladaptive aspects of self-oriented versus socially prescribed perfectionism. Journal of College Student Development, 46, 141-156. Klinger, E . (1975). Consequences of commitment to and disengagement from incentives. Psychology Review, 82, 1-25. Kordy, H . , Kramer, B . , Palmer, R. L . , Papezova, H . , Pellet, J. , Matthias, R. , et al. (2002). Remission, recovery, relapse, and recurrence in eating disorders: Conceptualization and illustration of a validation strategy. Journal of Clinical Psychology, 58, 833-846. Laessle, R. G . , K i t t l , S., Fitcher, M . M . , Wittchen, H - U . (1987). Major affective disorder in anorexia and bulimia. British Journal of Psychiatry, 151, 785-789. Larzelere, R. E . , & Mulaik , S. A . (1977). Single-sample tests for many correlations. Psychological Bulletin, 84, 557-569. Lawrence, M . (1979). Anorexia nervosa: The control paradox. Women's Studies International Quarterly, 2, 93-101. Leary, M . R. (1990). Responses to social exclusion: Social anxiety, jealousy, loneliness, depression, and low self-esteem. Journal of Social and Clinical Psychology, 9, 221-229. Leary, M . R., & Downs, D . L . (1995). Interpersonal functions of the self-esteem motive: The self-esteem system as a sociometer. In M . Kernis (Ed.), Efficacy, agency, and self-esteem (pp. 123-144). N e w York: Plenum Press. Leary, M . R., Schreindorfer, L . S., & Haupt, A . L . (1995). The role of low self-esteem in emotional and behavioral problems: Why is low self-esteem so dysfunctional? Journal of Social and Clinical Psychology, 14, 297-314. Leary, M . R., Tambor, E . S., Terdal, S. K . , & Downs, D . L . (1995). Self-esteem as an interpersonal monitor: The sociometer hypothesis. Journal of Personality and Social Psychology, 68, 518-530. Leary, M . R., Tchividjian, L . R., & Kraxbereger, B . E . (1994). Self-presentationcan be hazardous to your health: Impression management and health risk. Health Psychology, 13, 461-470. Lilenfeld, L . R. R., Stein, D , Bul ik , C. M . , Strober, M . , Plotnicov, K . , Pollice, C , Rao, R., Merikangas, K . R., Nagy, L . , & Kaye, W . H . (2000). Personality traits among currently eating disordered, recovered, and never i l l first-degree female relatives of bulimic and control women. Psychological Medicine, 30, 1399-1410. Lilenfeld, L . R. R., Wonderlich, S., Riso, L . P., Crosby, R. & Mitchel l , J. (2006). Eating disorders and personality: A methodological and empirical review. Clinical Psychology Review, 26, 299-320. Maltby, J., & Day, L . (2001). The relationship between exercise motives and psychological well-being. Journal of Psychology, 135, 651-660. Martin, T. R., Flett, G . L . , Hewitt, P. L . , Krames, L . , & Szantos, G . (1996). Personality in depression and health symptoms: A test of a self-regulation model. Journal of Research in Personality, 31, 264-277. Mazzeo, S. E . , Slof-Op't, M . C . T., Landt, I., Jones, I., Mitchel l , K . , Kendler, K . S., Neale, M . C , Aggen, S. H . , Bul ik , C . M . (2006). Associations among postpartum depression, eating 181 disorders, and perfectionism in a population-based sample of women. International Journal of Eating Disorders, 39, 202-211. McClintock, J. M . M . & Evans, I. M . (2001). The underlying psychopathology of eating disorders and social phobia: A structural equation analysis. Eating Behaviors, 2, 247-261. McDowel l , B . D . , Moser, D . J. , Ferneyhough, K . , Bowers, W . A . , Andersen, A . E . , & Paulsen, J. S. (2003). Cognitive impairment in anorexia nervosa is not due to depressed mood. International Journal of Eating Disorders, 33, 351-355. McGee, B . J., Hewitt, P. L . , Sherry, S. B . , Parkin, M . , & Flett, G . L . (2005). Perfectionism, body image discrepancy, and eating disorder symptoms. Body Image, 1, 29-40. McClel land, G . H . , & Judd, C. M . (1993). Statistical difficulties of detecting interactions and moderator effects. Psychological Bulletin, 114, 376-390. M c V e y , G . L . , Pepler, D . , Davis, R., Flett, G . L . , & Abdolel l , M . , (2002). Risk and protective factors associated with disordered eating during early adolescence. Journal of Early Adolescence, 22, 75-95. Meyer, C , & Waller, G . D . (2001). Social convergence of disturbed eating attitudes in young adult women. Journal of Nervous and Mental Disease, 189, 114-119. Minarik, M . L . , & Ahrens, A . H . (1996). Relations of eating behavior and symptoms of depression and anxiety to the dimensions of perfectionism among undergraduate women. Cognitive Therapy and Research, 20, 155-169. Mintz , L . , & Betz, N . (1988). Prevalence and correlates o f eating disordered behaviors among undergraduate women. Journal of Counseling Psychology, 35, 463-471. Mizes, J. S., Christiano, B . . , Madison, J., Post, G . , Seime, R., & Varnado, P. (2000). Development of the Mizes Anorectic Cognitions Qeuestionnaire-Revised: Psychometric properties and factor structure in a large sample of eating disorder patients. International Journal of Eating Disorders, 28, 415-421. Moulton, P., Moulton, M , & Roach, S. (1998). Eating disorders: A means o f seeking approval? Eating Disorders, 6, 319-327. Muller, D . , Judd, C. M . , & Yzerbyt, V . Y . (2005). When moderation is mediated and mediation is moderated. Journal of Personality and Social Psychology, 89, 852-863. Nesse, R. M . (2000). Is depression an adaptation? Archives of General Psychiatry, 57, 14-20. Oates-Johnson, T., & Clark, D . A . (2004). Sociotropy, body dissatisfaction and perceived social disapproval in dieting women: A prospective diathesis-stress study o f dysphoria. Cognitive Therapy and Research, 28, 715-731. Pacht, A . R. (1984). Reflections on perfection. American Psychologist, 39, 386-390. Patton, C. J. (1992). Fear of abandonment and binge eating: A subliminal psychodynamic activation investigation. Journal of Nervous and Mental Disease, 180, 484-490. Paunonen, S. V . (1998). Hierarchical organization of personality and prediction of behavior. Journal of Personality and Social Psychology, 74, 538-556. P D M Task Force (2006). Psychodynamic Diagnostic Manual. Silver Spring, M D : Alliance of Psychoanalytic Organizations. Pearlstein, T. (2002). Eating disorders and comorbidity. Archives ofWomens's Mental Health, 4, 67-78. Pearson, C. A . , & Gleaves, D . H . (2006). The multiple dimensions of perfectionism and their relation with eating disorder features. Personality and Individual Differences, 41, 225-235. 183 Pla, C , & Toro, J. (1999). Anorexia nervosa in a Spanish adolescent sample: A n 8-year longitudinal study. Acta Psychiatrica Scandinavica, 100, 441-446. Pliner, P., & Haddock, G . (1996). Perfectionism in weight-concerned and -unconcerned women: A n experimental approach. International Journal of Eating Disorders, 19, 381-389. Polivy, J., & Herman, C . P. (1992). Undieting: A program tp help people stop dieting. International Journal of Eating Disorders, 11, 261-268. Price, J., Sloman, L . , Gardner, R., Gilbert, P., Rohde, P. (1994). The social competition hypothesis o f depression. British Journal of Psychiatry, 164, 309-315. Rezek, P. J. & Leary, M . R. (1991). Perceived control, drive for thinness, and food consumption: Anorexic tendencies as a displaced reactance. Journal of Personality, 59, 129-142. Rodgers, W . M . , & Gauvin, L . (1998). Heterogeneity o f incentives for physical activity and self-efficacy in highly active and moderately active women exercisers. Journal of Applied Social Psychology, 28, 1016-1029. Rosenberg, M . (1965). Society and the adolescent self-image. Princeton, N J : Princeton University Press. Rothenberg, A . (1990). Adolescence and eating disorder: The obsessive-compulsive syndrome. Psychiatric Clinics of north America, 13, 469-488. Russell, D . W . (1990). The analysis of psychophysiological data: Multivariate approaches. In J. T. Cacioppo & L . G . Tassinary (Eds.), Principles of psychophysiology: Physical, social, and inferential elements (pp. 775-801). N e w York: Cambridge University Press. Saboonchi, F. , & Lundh, L . - G . (1997). Perfectionism, self-consciousness and anxiety. Personality and Individual Differences, 22, 921-928. 184 Saboonchi, F., Lundh, L . - G . , & Ost, L . - G . (1999). Perfectionism and self-consciousness in social phobia and panic disorder with agoraphobia. Behaviour and Research Therapy, 37, 799-808. Saftner, J. L . , & Crowther, J. H . (1998). Variability in self-esteem, moods, shame, and guilt in women who binge. International Journal of Eating Disorders, 23, 391-397. Sassaroli, S., & Ruggiero, G . M . (2005). The role o f stress in the association between low self-esteem, perfectionism, and worry, and eating disorders. International Journal of Eating Disorders, 37, 135-141. Schwartz, N . , & Clore, G . L . (1996). Feelings and phenomenal experiences. In E . T. Higgens & A . W . Kruglanski (Eds.), Social Psychology: Handbook of Basic Principles (pp. 433-465). N e w York: Guilford Publications. Shacham, S. (1983). A shotened version of the Profile of M o o d States. Journal of Personality Assessment, 47, 305-306. Schupak-Neuberg, E . , & Nemeroff, C. J. (1993). Disturbances in identity and self-regulation in bulimia nervosa: Implications for a metaphorical perspective o f "body as se l f . International Journal of Eating Disorders, 13, 335-347. Sherry, S. B . , Hewitt, P. L . , Besser, A . , McGee, B . J., & Flett, G . L . (2004). Self-oriented and socially prescribed perfectionism in the Eating Disorder Inventory perfectionism subscale. International Journal of Eating Disorders, 35, 69-79. Sherry, S. B . , Hewitt, P. L . , Flett, G . L . , & Harvey, M . (2003). Perfectionism dimensions, perfectionistic attitudes, dependent attitudes, dependent attitudes, and depression in psychiatric patients and university students. Journal of Counseling Psychology, 50, 373-386. Sherry, S. B . , Hewitt, P. L . , Lee-Baggley, D . L . , Flett, G . L . , Besser, A . (2004). Perfectionism and thoughts about having cosmetic surgery. Journal of Applied Biobehavioral Research, 9, 244-257. Shisslak, C. M . , Pazda, S., & Crago, M . (1990). Body weight and bulimia as discriminators of psychological characteristics among anorexic, bulimic, and obese women. Journal of Abnormal Psychology, 99, 380-384. Slade, P. D . (1982). Towards a functional analysis o f anorexia nervosa and bulimia nervosa. British Journal of Clinical Psychology, 21, 167-179. Soenens, B . , Vansteenkiste, M . , Luyten, P., Duriez, B . , & Goossens, L . (2005). Maladaptive perfectionistic self-representations: The mediational link between psychological control and adjustment. Personality and Individual Differences, 38, 487-498. Sommer, K . L . , Wil l iams, K . D . , Ciarocco, N . J., & Baumeister, R. F. (2001). When silence speaks louder than words: Explorations into the intrapsychic and interpersonal consequences of social ostracism. Basic and Applied Social Psychology, 23, 225-243. Sorotzkin, B . (1985). The quest for perfection: Avoiding guilt or avoiding shame? Psychotherapy, 22, 564-571. Spitzer, B . L . , Henderson, K . A . , & Zivian, M . T. (1999). Gender differences in population versus media body sizes: A comparison over four decades. Sex Roles, 40, 545-565. Srinivasagam, N . M . , Kaye, W . H . , Plotnicov, K . H . , Greeno, C , Weltzin, T. E . , & Rao, R. (1995). Persistent perfectionism, symmetry, and exactness after long-term recovery from anorexia nervosa. American Journal of Psychiatry, 152, 1630-1634. 186 Steiger, H . , Puentes-Neuman, G. , Leung, F. Y . (1991). Personality and family features of adolescent girls with eating symptoms: Evidence for restricter/binger differences in a nonclinical population. Addictive Behaviors, 16, 303-314. Stein, D . , Kaye, W . H . , Matsunaga, H . , Orbach, I, Har-Even, D , Frank, G , McConaha, C. W. , & Rao, R. (2002). Eating-related concerns, mood, and personality traits in recovered bulimia nervosa subjects: A replication study. International Journal of Eating Disorders, 32, 225-229. Stice, E . , & Agras, W . S. (1998). Predicting onset and cessation of bulimic behaviors during adolescence: A longitudinal grouping analyses. Behavior Therapy, 29, 257-276. Stice, E . , Ki l l en , J., Hayward, C , & Taylor, C. (1998). Support for the continuity hypothesis of bulimic pathology. Journal of Consulting and Clinical Psychology, 66, 784-790. Striegel-Moore, R. H . , Silberstein, L . R., & Rodin, J. (1993). The social self-in bulimia nervosa: Public self-consciousness, social anxiety, and perceived fraudulence. Journal of Abnormal Psychology, 102, 297-303. Strober, M . (1980). Personality and symptomatological features in young, nonchronic anorexia nervosa patients. Journal of Psychosomatic Research, 24, 353-359. Sullivan, P. F., Bul ik , C. M . , Fear, J. F. , & Pickering, A . (1998). Outcome of anorexia nervosa: A case-control study. American Journal of Psychiatry, 155, 939-946. Sutandar-Pinnock, K . , Woodside, D . B . , Carter, J. C , Olmsted, M . P., Kaplan, A . S. (2003). Perfectionism in anorexia nervosa: A 6—24-month follow-up study. International Journal of Eating Disorders, 33, 225-229. Tabachnick, B . G . , & Fidel l , L . S. (1996). Using multivariate statistics (3 r d ed.). N e w York: Harper Collins. Tasca, G . A . , M i k a i l , S. F. , & Hewitt, P. L . (2005). Group psychodynamic interpersonal psychotherapy: Summary of a treatment model and outcomes for depressive symptoms. In M . E . Abelian (Ed.), Focus on Psychotherapy Research (pp. 159-188). Nova Science. Thomsen, S. R., M c C o y , J. K . , Gustafson, R. L . , & Will iams, M . (2002). Motivations for reading beauty and fashion magazines and anorexic risk in college-age women. Media Psychology, 4, 113-135. Tiller, J. M . , Sloane, G . , Schmidt, U . , Troop, N . , Power. M . , & Treasure, J. L . (1997). Social support in patients with anorexia nervosa and bulimia nervosa. International Journal of Eating Disorders, 21, 31-38. Tozzi , F. Aggen, S. H . , Neale, B . M . , Anderson, C. B . , Mazzeo, S. E . , Neale, M . C , et al. (2004). The structure of perfectionism: A twin study. Behavior Genetics, 34, 483-494. Twenge, J. M . , Baumeister, R. F., Tice, D . M . , & Stucke, T. S. (2001). If you can't join them, beat them: Effects of social exclusion on aggressive behavior. Journal of Personality and Social Psychology, 81, 1058-1069. Twenge, J. M . , Catanese, K . R., & Baumeister, R. F. (2002). Social exclusion causes self-defeating behavior. Journal of Personality and Social Psychology, 83, 606-615. Twenge, J. M . , Catanese, K . R., & Baumeister, R. F. (2003). Social exclusion and the deconstructed state: Time perception, meaninglessness, lethargy, lack o f emotion, and self-awareness. Journal of Personality and Social Psychology, 85, 409-423. 1 Tyrka, A . R., Waldron, I., Graber, J. A . , & Brooks-Gunn, J. (2002). Prospective predictors of the onset of anorexic and bulimic syndromes. International Journal of Eating Disorders, 32, 282-290. 188 Vitousek, K . B . , & Hol lon, K . B . (1990). The investigation of schematic content and processing in eating disorders. Cognitive Therapy and Research, 14, 191-214. Wade, T. D . , Bul ik , C . M . , Neale, M . , & Kendler, K . S. (2000). Anorexia and major depression: Shared genetic and environmental risk factors. American Journal of Psychiatry, 157, 469-471. Wagner, S., Halmi , K . A . , & Maguire, T. V . (1987). The sense of personal ineffectiveness in patients with eating disorders: One construct or several? International Journal of Eating Disorders, 6, 495-505. Weinrich, P., Doherty, J. , & Harris, P. (1985). Empirical assessment of identity in anorexia and bulimia nervosa. Journal of Psychiatric Research, 19, 297-302. Williams, K . D . , Cheung, C . K . T., & Choi , W. (2000). CyberOstracism: effects of being ignored over the Internet. Journal of Personality and Social Psychology, 79, 748-762. Will iams, K . D . , & Zadro, L . (2001). Ostracism: On being ignored, excluded, and rejected. In M . R. Leray (Ed.), Interpersonal rejection (pp. 21-53). N e w York: Oxford University Press. Williamson, D . A . , Stewart, T. M . , White, M . A . , York-Crowe, E . (2002). A n information-processing perspective on body image. In T. F. Cash & T. Pruzinsky (Eds.), Body image: A handbook of theory, research, & clinical practice (pp. 47-54). N e w York: Guilford Press. Wiseman, C. V . , Sunday, S. R. , Klapper, F. , Harris, W . A . , & Halmi , K . A . (2001). Changing patterns o f hospitalization in eating disorder patients. International Journal of Eating Disorders, 30, 69-74. 189 APPENDICES Appendix A Personality Description for Socially Accepted Condition Your profile indicates you're the type who has rewarding relationships throughout life. Research suggests that you're likely to have a long and stable marriage and have friendships that wi l l last into your later years (Schmidt & Vohs, 2003). People generally like you and want to be with you. In fact, your scores suggest that you may already have some experience with interpersonal closeness. Y o u are very likely to maintain your relationships over time, and the odds are that you ' l l always have friends and people who care about you. Y o u can look forward to a lifetime of close relationships. Personality Description for Socially Excluded Condition Your profile suggests that you're the type who wi l l end up alone later in life. Y o u likely have some close friends and relationships now, but research suggests that by late-20s you wi l l have lost touch with many of these people and the relationships w i l l have drifted away (Schmidt & Vohs, 2003). Y o u may marry, or even have several marriages, but these are likely to be short-lived and not last as long as the average marriage. A s you age, you w i l l l ikely find it more difficult to sustain relationships (Crosby et a l , 1999). In fact, your scores suggest that you may already have some experience with interpersonal conflict or perhaps even with the loss of a relationship. Relationships don't last, and when you're past the age where people are constantly forming new relationships, the odds are you ' l l end up being alone more and more. 190 Appendix B Relevant Regression Equations for Testing Mediated Moderation (1) Y=Bl0+BnX+{3nMo + pxyXMo + s, (2) Me = /ho + (hiX+ fhiMo + fciXMo + s 2 (3) Y= /%0 + fh\X+ fhiMo + friXMo + /334Me + fasMeMo + si Note. The following labels were used: Y (dependent variable), X (perfectionism dimension), M o (moderator: social feedback condition), and Me (mediator: rejected affect). 

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.831.1-0100538/manifest

Comment

Related Items