INFORMATION TO USERS

This manuscript has been reproduced from the microfrlm master. UMI films the text direct^ from the origmal or copy submitted. Thus, some thesis and disseitation copes are in typewriter ù o t, while others may be from any type o f computer printer.

The quality of this reproduction is dependent npon the qnniity of the copy submitted. Broken or indistmct prmt, colored or poor quality illustrations and photogr^hs, print bleedthrough, substandard margins, and improper alignment can advosety affect reproduction.

In the unlik^ event that the author did not send UMI a complete manuscript and there are missmg pages, these will be noted. Also, if unauthorized copyright material had to be removed, a note will indicate the deletion.

Oversize materials (e g., maps, drawings, charts) are reproduced by sectioning the origituü, b%inning at the upper left-hand comer and continuing from left to right in equal sections with small overiaps. Each original is also photographed in one exposure and is included in reduced form at the back o f the book.

Photographs included in the original manuscript have been reproduced xerographicaOy in this copy. Higher quality 6” x 9” black and white photographic prints are available for any photographs or illustrations appearing in this copy for an additional charge. Contact UMI directly to order. UMI A Bell & Howell Information Company 300 North Zeeb Road, Ann Arbor MI 48106-1346 USA 313/761-4700 800/521-0600

STATE SELF-ESTEEM REACTIONS TO REJECTION: AN APPLICATION OF THE SOCIOMETER MODEL TO WOMEN WITH AND WITHOUT EATING DISORDER BEHAVIOR

DISSERTATION

Presented in Partial Fulfillment of the Requirements for

the Degree Doctor of Philosophy in the Graduate

School o f The Ohio State U niversity

By Deborah L. Downs, M.A.

*****

The Ohio S tate U niversity 1997

D issertation Committee: Approved by Professor Richard K. Russell, Adviser Professor Nancy Betz Advisor Professor W. Bruce Walsh Department of Psychology ÜMX Nuaber: 9801680

UMI Mkrofonn 9801680 Copyright 1997, by UMI Company. All rights reserved.

This microform edition is protected against unauthorized copying under Title 17, United States Code.

UMI 300 North Zeeb Road Ann Arbor, MI 48103 ABSTRACT

People are commonly concerned about the degree to which they are accepted or rejected by others, although the extent of this concern may be greater for some groups than others. The Sociometer Theory of self­ esteem proposes that the increase in negative emotions, and decrease in state self-esteem that follows social rejection, is an adaptive response that promotes group living. Previous research has shown that subjects report greater levels of negative affect, and lower levels of state self-esteem when they believe, or imagine that, they have been socially excluded. The purpose of the present study was twofold. First, although previous research has found support for the presence of a sociometer that "monitors" social inclusion, the present study was the first to examine the sociometer hypothesis in live, face-to-face interactions. Second, individuals who scored high and low on an eating disorder measure were included to examine the hypothesis that the sociometer may not function normally for individuals with eating disorders. It was hypothesized that people who reported higher levels of eating disorder characteristics would report more negative reactions to social rejection. The results of the study found support for the ii hypothesis that social rejection leads people to feel more negatively about themselves» experience fewer positive feelings, and think more negative thoughts. Support for the hypothesis that subjects with high eating disorder scores would respond significantly more negatively to rejection was mixed. Although rejected subjects with high eating disorder scores reported more negative thoughts, changes in state self-esteem and affect were unrelated to eating disorder scores. Future studies of people's reactions to social rejection in face-to-face interactions should aim to develop a rejection condition that is subtle and prevents ceiling e ffe c ts .

m ACKNOWLEDGMENTS

I wish to thank my adviser. Rich Russell, for his encouragement, support and professional expertise without which this project would not have been possible. Also, I would like to express appreciation to my d isse rta tio n committee members, Nancy Betz and Bruce Walsh, fo r th e ir time and valuable contributions. Special thanks go to Mark Leary for his encouragement and theoretical contributions, particularly In relation to the continued development of the sociometer model of self-esteem. And Andy Clifford, you have a special place In the process of this dissertation; thank you for your time and efforts throughout! I am also grateful to John Downing for his statistical consultation and support, and to Lelsa and Mary for their understanding presence before, during and beyond the dissertation process. Finally, I would like to extend a special thank you to my research assistants Casey, Gretchen, Jameson, Kimberly, Melissa, Stacy and Tina whose enthusiasm, dedication, and humor made data collection such an enjoyable experience.

IV VITA

May 12, 1966...... Born - Wooster, Ohio 1988 ...... B.A. Psychology and Spanish, Denison University 1993 M.A. Psychology, Wake Forest University 1993-199 4...... Academic Advisor, The Ohio S tate U niversity 1994-199 6...... Teaching Associate, The Ohio State University 1995-199 6...... Student Personnel Assistant, Counselor The Ohio S tate U niversity 1996...... Adjunct Faculty, Otterbein College 1996 - present ...... Predoctoral Intern The U niversity of South Florida Tampa, Florida

PUBLICATIONS Research Publication

Leary,LccLrYy M.PI» R., K* 9 Tambor, IttWDOi 9 E. C» S w# .. # Terdal, 6 liQ ttl 9 S. w# K.. # & 0 1 Downs, UOWnS 9 D. U# L. L # (1995 ) )• The functionfunction, of self-esteem feelings: » Testing the sociometer hypothesis. " *lQM.r.nâ.1-ftf-P.ersonaJit.Y-âDd .Sflcial. PsYcholoqiY. 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.), Efficacv. aoencv. and self-esteem. New York: Plenum. H eft, H. & Downs. D. L. (1995). W. B. Walsh, K. H. Craik, and R. Price (Eds.) (1992). Persfln-EpvlrQDBftnt-EiychQ]o,9y;■Hg.dg.ls-and. ^ P erspectives. H illsd ale. NJ: Lawrence Erlbaum A ssociates, Publishers. X1+ 269 ppT + indices. ISBN 0-8058-0344-0 Leary, M. R.. Nezlek, J. B., Downs, D., Radford-Davenport, J., Martin, J., & McMullen, A. (1994). Self-Presentation in everyday interactions: Effects of target fam iliarity and gender composition. Journal of Personalitv and Social Psvcholoov. SI, 664-673. V FIELDS OF STUDY Major Field: Psychology , ^ ^ , Studies in counseling and social psychology

VI TABLE OF CONTENTS Eagê A bstract...... ii Acknowledgements ...... iv V ita...... V List of Tables ...... ix

Chapters : 1. INTRODUCTION...... 1 1.1 Social inclusion as a fundamental motive ...... 3 1.1.1. Attachment theory ...... 3 1.1.2. Social exclusion theory ...... 5 1.1.3. The need to belong ...... 6 1.2 Social rejection and self-esteem ...... 11 1.2.1. Self-esteem ...... 11 1.2.2. Etiology of self-esteem...... 13 1.2.3. Sociological perspectives...... 13 1.2.4. Clinical perspectives ...... 16 1.2.5. The self-esteem motive...... 19 1.2.6. The sociometer...... 19 1.2.7. Social interactions and self esteem...... 20 1.2.8. Personal exclusion and state self esteem...... 21 1.2.9. Trait self-esteem and exclusion ...... 22 1.3 A dysfunctional sociometer ...... 23 1.4 Eating disorders and the sociometer ...... 23 1.4.1. The clinical picture ...... 24 1.4.2. Social components of anorexia and bulimia nervosa ...... 27 1.4.3. Purpose of study ...... 28 1.4.4. Hypotheses...... 29

2. METHOD...... 31 2.1. Participants ...... 31 2.2. Instruments ...... 33 2.3. Procedure ...... 35

3. RESULTS...... 40 3.1. Manipulation check ...... 40 3.2. Reactions to feedback ...... 41

vn 4. DISCUSSION...... 47

4.1. Social feedback and self-esteem ...... 48 4.2. Eating disorder scores and self-esteem ...... 51 4.3. Affective reactions ...... 52 4.4. Cognitive reactions...... 55 4.5. Limitations and suggestions for future research ...... 57 4.7. Implications for counseling ...... 60

LIST OF REFERENCES...... 62

APPENDICES...... 67 A. Self-esteem Inventory ...... 67 B. Eating disorder Inventory ...... 68 C. State self-esteem scale ...... 69 D. A ffective rea ctio n s...... 70 E. Cognitive rea ctio n s...... 71 F. Informed consent form ...... 72 G. Demographic filler Items...... 73 H. Manipulation check ...... 74

vlli LIST OF TABLES

Table Page 3.1 Means and standard deviations ...... 43 3.2 Grand means and standard deviations by condition ...... 43 3.3 Grand means and standard deviations by s ta tu s ...... 43 3.4 ANOVA tables for self-esteem difference scores, affective sums, and thought listing positivity scores...... 44

IX CHAPTER 1

INTRODUCTION

Most people would agree that individuals naturally strive to be liked and accepted by others, and to avoid social rejection. The fact that people are motivated to be accepted is evident in the considerable time, energy, and money that people spend to appear attractive to others (Downs, 1993; Leary, 1990). Although people's desires to be socially included may generalize to strangers or casual acquaintances, these desires are particularly salient when Interactions consist of individuals or groups who are important to them (Downs, 1993; Baumeister & Leary, 1995). Despite e ffo rts to avoid exclusion, social rejection is relatively common (MacLeod, Mathews & Tata, 1986), although people are rarely completely shunned or excluded (Leary, 1990). Perhaps the construct of acceptance is better understood as a continuum (Leary, 1990). Although inclusion and exclusion represent two dichotomous interpersonal outcomes, people more often experience a range of inclusionary feedback from others. For example, whereas people sometimes perceive that they have been completely accepted, at other times they may feel ignored, mildly excluded, or utterly rejected (Leary, 1990). Although the experience of being accepted by others may lead to good feelin g s and p o sitiv e self-evaluations (Betz, Wohlgemuth, Serling, Harshbarger & K lein, 1995; Shrauger & Schoeneman, 1979; Tambor & Leary, 1992), the present investigation will focus on the negative reactions people have to rejection. While the experience of acceptance is certainly important, people rarely suffer or seek help for experiencing too much of it. In other words, social feedback in the form of smiles and nods from others feels good, and indicates that the person is likeable and relating in an acceptable manner. However, people who experience chronic or extreme social rejection may suffer a variety of deleterious effects, including loneliness, low self-esteem, and difficulties in social relationships (Bullock, 1992; Hazen & Shaver, 1987; Jones, 1990; Parkhurst & Asher, 1992; Tambor & Leary, 1993). Research aimed at understanding how people react to and cope with rejection may be directly useful to practitioners. For example, research examining the processes underlying the relationship between social rejection and mental health may generate new theories and more effective interventions. Although people who feel rejected may aggress, withdraw, or try to remedy social rejection with compensatory behaviors such as excuses, apologies, and justifications (Alden & Phillips, 1990; Parkhurst & Asher, 1992; Patterson & Stoolmiller, 1991; Schlenker, 1980), the present paper will focus on self-related affective and cognitive reactions to social rejection. Specifically, changes in people's level of self-esteem in response to social rejection and acceptance will be examined. In addition, because research suggests that women with eating disorders may be more sensitive to negative social feedback (deGroot & Rodin, 1994), women with high scores on eating disorder measures w ill be compared women who do not report eating disorder symptoms on a variety of measures. Prior to describing the exact nature of the proposed study, the program of theory and research leading to the investigation will be reviewed. Social Inclusion as a Fundamental Motive Few would disagree that people are motivated to maintain their social relationships and that feeling a part of a group contributes to w ell-being. However, psychologists d iffe r in the extent to which they believe the maintenance of social relationships is a fundamental motive th a t drives behavior. Whereas tra d itio n a l Freudian analysts propose that aggressive and sexual drives guide a large portion of people's behavior, other theorists such as Sullivan or Rogers would place a greater emphasis on the importance of social relationships (Rogers, 1961; Sullivan, 1953). In reality, people's behavior is probably driven by many motives. For the present purposes, it is assumed th a t behavior is motivated to a great extent by a need to be socially accepted. Support for this position comes from theory and research that describes people's negative reactions to the experience of social rejection (Baumeister & Tice, 1990; Leary, 1990). Attachment Theorv Theory and research has examined the impact of rejection at various ages. Ainsworth's (1989) extension of Bowlby's (1969; 1973) theory of social attachment described the impact of social rejection during infancy and childhood on subsequent behavior. Ainsworth devised a research paradigm known as the "Strange Situation" to study children's behavior in the context of mother and

child interactions. In the Strange Situation, an infant was placed in a novel playroom and observed during the presence, brief absence, and return of the caretaker. Results indicated that the children reacted in predictable ways depending on the quality of the attachment between the infant and her mother. Proponents of attachment theory argue that children who receive affection and consistent care form a "secure" attachment with their

caretaker. Research indicates that children who form secure attachments later display higher levels of social competence, self­ esteem and empathy (see Crain, 1992). However, in contrast to the development of a secure attachment, children who are raised by rejecting or unreliable caretakers form avoidant or anxious-ambivalent attachment styles, respectively (Ainsworth, 1989). Children with avoidant attachments ignore the presence or departure of a caretaker in novel situ a tio n s. However, children with anxious-ambivalent attachments are initially clingy and cry loudly at the departure of the caretaker, yet appear angry at their return. Because children who were judged to have different attachment styles displayed unique and predictable behavior patterns, Ainsworth concluded that the quality of the parent child relationship had an impact on the child's social behavior (Ainsworth, 1989). Subsequent research on attachment theory proposed that the quality of a child's relationship with his or her primary caretaker(s) has an impact on later adult relationships. For example, people who report avoidant or anxious-ambivalent attachment styles as children have difficulties as adults in developing and maintaining close relatio n sh ip s (Hazan & Shaver, 1987). Attachment theory provides initial, although indirect, support for the contention that children are able to process information about their relationships early on, and that the quality of those relationships are important to social development. More recently, proponents of social exclusion theory have extended Bowlby's and Ainsworth's conception of innate attachment processes to account for people's reactions to social exclusion. Social Exclusion Theorv

Not only do people want to be liked and accepted, they appear to react strongly to threats to their inclusionary status. When people perceive that they have been rejected, they feel a range of aversive emotions including anxiety (Baumeister & Tice, 1990), loneliness, depression, jealousy, and low self-esteem (Jones, 1990; Leary, 1990; Tambor & Leary, 1992). Proponents of social exclusion theory suggest that the anxiety that people feel in response to rejecting feedback is adaptive. With respect to evolutionary history, individuals who lived in groups were more likely to survive due to the greater availability of crucial resources such as food, protection, shelter, and the opportunity to reproduce. Humans who felt anxious after leaving a group were more likely to return and to survive, compared to those who did not. Over time, exclusionary anxiety, and other negative emotions associated with threats to social membership, evolved and functioned to promote

group living. The effects of an innate desire to live in groups is evident in modern day reactions to re je c tio n . Social exclusion theory suggests that the anxiety rejected people experience functions as an "interrupt mechanism" th a t s h ifts a tten tio n from the in teractio n i t s e l f to the anxious person's behavior (Baumeister & Tice, 1990). In other words, feeling negative emotions in response to exclusionary information

encourages a timely cessation of an individual's offensive behavior in order to salvage his or her inclusionary status. Although most people desire at least a minimum level of social approval, people vary in the extent to which they are motivated to maintain social relations (Leary, 1983). But regardless of whether an individual prefers to maintain few or numerous acquaintances, theory and research suggest that most people react negatively to even mild forms of perceived re je c tio n . Whereas social exclusion theory aims to provide an explanation of the negative affect that follows social exclusion, Baumeister and Leary (1995) propose that the need to be accepted can be expressed as a social motive. The Need to Belong The fact that people want so strongly to be accepted suggests that thoughts, feelings, and social behaviors are likely to be highly

influenced by the desire to maintain close interpersonal relationships. Although many prior theorists including Homey (1945), Sullivan (1953), and Bowlby (1969, 1973) have described the importance of positive and enduring social connections, little has been done to investigate the assumption that people are driven by a need for intimate and accepting social relationships (Baumeister & Leary, 1995). To rectify the lack of attention paid to people's acceptance needs, Baumeister and Leary (1995) reviewed a wide range of research to support what they termed the "belongingness hypothesis." The belongingness hypothesis states that people are driven by a fundamental need to maintain significant interpersonal relationships. Similar to attachment and social exclusion theory, the desire to belong to groups is assumed to be innate and to have developed via evolutionary processes. Because the desire to have social connections is posited to be a fundamental need, the authors reviewed research that illustrated how belongingness needs might affect typical human processes. For example, the need to maintain social relations appears to have an impact on the content and structure of emotions, cognitions, and behavior. To examine the validity of their basic hypothesis, Baumeister and Leary (1995) developed a series of logical predictions derived from belongingness theory. For example, they proposed that a fundamental desire to belong to groups should influence basic human processes, including how people think about and store information (Baumeister & Leary, 1995). In fact, Baumeister and Leary (1995) reviewed findings that described how people process and store information about close relationships differently from information regarding casual acquaintances (Sedikides, Oslen & Reis, 1993). These findings not only support the view that belongingness needs may influence the way people process information, but also may provide cues for understanding the nature of dysfunctional thinking patterns characteristic of certain psychological disorders. Practitioners such as Beck (1976), and Ellis (1962) have long focused on the importance of changing maladaptive cognitions in therapy. Traditionally, cognitive approaches have not focused on the origin of clients' maladaptive cognitions and whether or not the issues a t hand were rela te d to negative social experiences. However, clinical theorists who assume that problematic interpersonal experiences underlie and support many of people's dysfunctional thought patterns may be able to more effectively guide the therapeutic process. For example, by incorporating a greater awareness of the importance of interpersonal functioning on cognition, clinicians may be more efficient in the diagnosis and treatment of client difficulties. If the content and process of memory is, in fact, organized according to social relationships (Sedikedes et al., 1993), then negative interpersonal experiences during childhood may lead to the development of a perceptual system that is highly sensitive to negative interpersonal cues. For example, a child whose desires to be accepted were consistently thwarted may have developed a cognitive system highly attentive to cues connoting rejection. The proposition that people may perceive information based on learned categories of information is similar to the concept of cognitive schema. Schema are hypothetical cognitive structures that describe how people store and use information in memory (Fiske & Taylor, 1991).

8 Schema that are frequently "primed" by external cues are used more than others. For example, in most cultures every person has a self­ schema which represents a h ierarch ically organized co llectio n of information about the self. Individuals who store negative information about the self based on negative social feedback during childhood and adolescence are likely to be more attentive to negative social cues than others. A learned tendency to focus on negative external cues driven by a heightened need to belong seems consistent with many of the characteristics typical of individuals with eating disorders. For example, women with eating disorders have been shown to be more attentive to information connoting rejection (deGroot & Rodin, 1989). Also, eating disorder symptoms coincide with characteristics that are interpersonal in nature such as an external locus of control, low assertiveness and low self-esteem (Williams et al, 1993). Although self-esteem may be considered a construct that is more intrapersonal than interpersonal, the inclusion of self-esteem as a variable strongly connected to social processes will be described below. In addition to social cognitions, people's emotions are also strongly influenced by the degree to which they believe they are accepted and liked by others (Baumeister & Leary, 1995). Positive feelings of e la tio n and joy accompany social inclusion, and negative feelings of anxiety, depression, jealousy, guilt and loneliness follow perceptions of reje ctio n (Leary, 1990; Leary, Tambor, Terdal & Downs, 1994). Research findings have shown that people who lack adequate social support experience higher levels of physical and emotional distress (Swindle & Moos, 1992). The positive relationship between social acceptance and physical and mental health has been shown to be particularly true for women (Reis, Wheeler, Kernis, Spiegel, & Nezlek, 1985). Theory and research indicating that women show a greater susceptibility to negative feelings following social rejection may be important to understanding disorders that occur at higher rates among women. As mentioned earlier, eating disorders in particular may be rooted in a hypervigilance of social rejection cues directed toward the self. A hypervigilence to rejection may have developed as a reaction to prior interpersonal difficulties. For example, women with eating disorders report having more trouble separating from their mothers compared to normal populations (Armstrong & Roth, 1989; Sours, 1974).

The desire to belong also influences a large portion of people's behavior (Baumeister & Leary, 1995). For example, people spend considerable time presenting themselves in ways they believe will be most a ttra c tiv e to others (Goffman, 1959; Leary & Kowalski, 1990) and not jeopardize their social relationships (Baumeister & Leary, 1995). People evaluate others as socially acceptable according to at least three basic criteria. Socially desirable individuals are physically attractive, contribute resources to the group, and follow established group norms (Baumeister & Tice, 1990). Individuals who feel they do not belong may put forth extra effort to meet acceptance

10 criteria. For example, current cultural trends In the United States equate thinness to beauty, and many American women go to great lengths to lose weight and thereby be perceived as attractive. It may be that as dieters begin to win attention and praise for their weight loss, they begin to equate self-esteem with body image. Unfortunately, the perception that self worth is based on weight may signal or precede a clinical eating disorder. Eating disorders often begin with what was originally intended to be a time bound period of restricted eating (American Psychological Association (APA) Diagnostic and S ta tis tic a l Manual (DSM), 1994; Nagel & Jones, 1992). Of course, the link between self-esteem and social acceptance in not limited to women with eating disorders. Often using male and female college students as subjects, recent research on self-esteem has focused on the Interaction between social acceptance and self-feelings. Social Rejection and Self-esteem

Self-esteem Self-esteem has been a topic of Interest to both practitioners and researchers In social and counseling psychology (Betz, et al., 1995; Brown, 1993; Kernis, 1995; Mecca & Smelser, 1989). Historically, self-esteem has been conceptualized as a generalized attitude toward the self which, like all attitudes, has both cognitive and affective components (Brewer & Crano, 1994; Rosenberg, 1965). Self-esteem has been proposed to Include Individuals' evaluations of self worth as well as positive or negative feelings about those judgments.

11 More contemporary descriptions of self-esteem state that the processes underlying people's global self-evaluations are primarily affective, not cognitive (Betz et ai., 1995; Brown, 1993). Specifically, the affective component of self-esteem concerns the positive and negative feelings associated with high and low self­ esteem, respectively (Leary & Downs, 1995).

Although researchers and theorists interested in the self have begun to address the constructs' social aspects, self-esteem continues to be treated as an isolated construct independent of a particular theory, and unrelated to sociocultural influences (Betz et al., 1995; Kernis, 1995; Oyserman & Narkus, 1993). More recent conceptualizations have expanded the concept of self-esteem from a single, global entity to a multidimensional construct (Kernis, 1995). Early conceptualizations described self­ esteem as a general, stable and unidimensional construct (Rosenberg, 1965). Since then, different dimensions or subcategories of self­ esteem have been formulated (Fleming & Courtney, 1984; Heatherton & Pol ivy, 1991; Pelham & Swann, 1989). For example, recent classifications of global self-esteem include baseline (stable) versus barometric (Rosenberg, 1989), explicit and implicit (Epstein & Morling, 1995), tru e versus contingent (Deci & Ryan, 1995) and stab le or unstable components (Greenier, Kernis & Maschull, 1995). Recently, self-esteem has been divided into state and trait self-esteem. State self-esteem describes the transient self-related thoughts and emotions that vary from situation to situation (Heatherton & Polivy, 1991), whereas trait self-esteem is more general

12 and enduring (Rosenberg, 1965). Although these theories differ somewhat in how they conceptualize self-esteem, it seems clear that models of self-esteem have matured into more complex accounts of people's self-feelings. In conjunction with the multiple efforts to reach a consensus on what self-esteem really is (Brown, 1993), theories also have emerged that address the origins and development of self-esteem. EtWogy of_&elfrg;tsem Some people seem prone to dislike themselves, whereas others are more accepting of their strengths and weaknesses. The origin of people's enduring positive or negative self-feelings is frequently attrib u te d to so cial learning experiences (Sanford & Donovan, 1984). Just as early social interactions are important to how children later respond to others (Ainsworth, 1989; Hazen & Shaver, 1987), positive social feedback and acceptance are an important part of the development of self-esteem (Rogers, 1961). Theories of self-concept development also use past social experiences as an explanatory construct. Other people are frequently a source of information about the composition of the self including traits and abilities, as well as evaluations of those qualities (Cooley, 1902; Fazio e t a l . , 1981; Head, 1934; Tice, 1992). Because both self-esteem and self-concept can be subsumed under the general construct of the self, theories related to the development of the self concept can be argued to relate to the development of self-feelings. Sociological Perspectives. People often tie the formation of their self descriptions to what they have learned about themselves from

13 other people (Schrauger & Schoeneman, 1979; Tice, 1992). Although the

content of people's self descriptions, such as "I am a car salesman", differentiates the self concept from the affective self-evaluations that define self-esteem, "I really like who I am", both can be assumed to be shaped by social factors (Cooley, 1902; Head, 1934). This is not meant to imply that people's self-descriptions and self-feelings are not influenced by intrapersonal factors; however, the present author considers the influence of social factors on the development of the self as primary. Symbolic interactionists represent a class of sociologists who propose that people develop their sense of self primarily from the social feedback. Cooley (1902) was the first to describe this phenomena as "the looking glass self" and proposed th a t people compose their self concepts in part by imagining how they appear to others, which he labeled re fle c te d appraisals (Cooley, 1902; Shrauger &

Schoeneman, 1979). Cooley also proposed th a t feedback generated during interactions with partners who were subjectively important had a greater influence on individuals' self-perceptions. For example, a father telling his daughter that she was bright and creative would have a greater impact on her subsequent self descriptions than the same message from a casual acquaintance. Mead (1934) further developed the idea that people come to know who they are through reflected appraisals, and suggested that over time people imagine a "generalized other" when speculating on how they appear to others. Unlike Cooley who suggested that people imagine the reactions of specific interactants, the generalized other represents a

14 more h o lis tic c u ltu ra l environment based on social norms (Shrauger & Schoeneman, 1979). Research on propositions derived from a symbolic interactionist perspective has provided mixed results. The contents of people's self descriptions appear to correlate significantly with how they believe th a t others perceive them. However, research f a ils to support a relationship between people's beliefs about the content of others perceptions, and how others actually perceive them (Brown, 1993; Fazio E ffrein & Falender, 1981; Shrauger & Schoeneman, 1979). The lack of consistent evidence to support that people accurately perceive others' feedback may be because the relationship is nonlinear. If so, correlational statistics based on linear assumptions would fail to capture significant nonlinear relationships among variables (Brown, 1993). Also, the lack of positive findings between targets' and perceivers' social judgements may be due to a neglect of variables that mediate interpersonal perception such as interaction expectancies (Fazio et al., 1981). Overall, the process of incorporating information about the self from social feedback appears to be a complex, reciprocally interactive process worthy of further research. Research findings related to the sociological perspective on self development have distinguished the role of social feedback as a crucial component in the development of self­ descriptions and evaluations. In addition to the sociological perspective, theory and research derived from clinical perspectives have also acknowledged that positive social relationships are related to high self-esteem.

15 C linical Perspectives and S elf Esteem

Historically, Adler (1929), Homey (1937), and Sullivan (1953) have provided clinical perspectives that have emphasized the role of social and cultural factors on the self and psychopathology. Adler was one of the first to break away from traditional psychoanalytic theory that attributed behavior to sexual and aggressive drives (Carson & Butcher, 1992). Rather, Adler believed that personality and behavior were primarily motivated by social forces and the desire to belong to groups. Therefore, according to Adler's perspective, people's evaluative self-feelings might be based on criteria that would enhance social acceptance. Homey proposed that neurotic symptoms varied according to the social norms of a particular culture. From this perspective, because American culture currently values thinness as a symbol of femininity, the increase in eating disorders among American women is not surprising. Similarly, because Homey described neurotic symptoms as a result of learned social constraints, it can be argued that women's lower levels of self-esteem, as well as the rise of eating disorders within industrialized cultures, can be attributed in part to sociological constraints and cultural pressures. Contemporary theorists have also identified traditional social assumptions that have labeled women as d e fic ie n t as contributing to women's interpersonal and psychological difficulties (Worrell & Remer, 1992). Sullivan (1953) placed the development of the self in an interpersonal context. According to Sullivan, people learn to think of themselves as good or bad based on the quality of their early

16 interactions with caregivers. For example, a negative relationship with a caregiver would result in anxiety and a negative self-view.

Most measures of self-esteem have been designed based on personality theory or clinical perspectives, despite the relevance of self-esteem to counseling psychology (Betz et ai., 1995). To rectify this, Betz and her colleagues (1995) devised a measure of self-esteem based on Rogers concept of unconditional positive regard. In a series of four studies, a 20-item measure of self-regard was developed and provided evidence for the measure's internal reliability and convergent, construct and discriminant validity. Central to Roger's person-centered approach to psychotherapy is the contention that individuals will value themselves if people who are close to them provide unconditional positive regard. Unconditional positive regard (UPR) is the noncontingent acceptance of a person as a valuable and unique individual (Rogers, 1961). However, this does not imply that an individual's behavior is always acceptable. In fact, part of a therapist's role is to provide UPR so that the client will develop an unconditional self-acceptance, and thereby select behaviors conducive to se If-growth. Because of the emphasis that Rogers placed on the role of early experiences of UPR, Betz and her colleagues (1995) designed a study to measure the relationship of unconditional self-regard (USR) to subjects' perceptions of the amount of UPR received from three "important people" during childhood (Betz et a l., 1995). Results showed that for women, the ability to demonstrate USR was significantly related to their perception that they were

17 unconditionally accepted by others. This relationship was

particularly true for female subjects' levels of self acceptance, and the degree to which they reported receiving UPR from their mothers. For men, levels o f USR were not re la te d to the amount of UPR received from important others.

In sum, the process of self-esteem formation appears to be related strongly to the processing and storage of information gleaned from important social relationships. Proposed conditions for the development of high self-esteem include secure attachments (Ainsworth, 1989; Bowlby, 1969, 1973), imagined p o sitiv e evaluations from sp e cific or generalized others, (Cooley, 1902; Mead, 1934), and the reception of UPR from significant others (Betz et a l., 1995; Rogers, 1961).

Each of these conditions assumes that high self-esteem is more desirable than low esteem and that individuals learn to feel positively or negatively about themselves from their social interactions. Theories of self-esteem assume that people strive to maintain high levels of esteem for a variety of reasons. For example, people are motivated to have positive self-feelings because high esteem is associated with good feelings, facilitates achievement strivings, and suggests that an individual has power and influence (see Leary & Downs, 1995). However, others have postulated that self-esteem represents one of several needs which include desires for pleasure and absence of pain, the maintenance of an accurate view of reality, and the retention of social relations (Epstein & Morling, 1995). Leary

18 and Downs (1995) propose that self-esteem is a functional system that serves to monitor the quality of people's social relationships. The Self-Esteem Motive. Few theories of self-esteem have considered why people might benefit from experiencing positive or negative emotions about themselves. In fact, most theories assume that people strive to protect and increase their self-esteem because of the desirable attributes that accompany high levels of self-regard (Leary & Downs, 1995). Previously, the motive to avoid low self-esteem and to achieve positive self-feelings has been explained by the desire to protect the self from negative feelings (Narkus, 1980), increase goal achievement (Bandura, 1977), and to win dominance over others (Tedeschi & Norman, 1985). However, it is possible that the many hypothesized self-esteem motives may be more parsimoniously accounted for under a single theoretical umbrella (Leary & Downs, 1995). The Sociometer. The fact that human beings are so powerfully motivated to have good feelings about themselves suggests that desires to increase self-esteem are adaptive. In the search for a single, overarching theory of self-esteem, Leary and Downs (1995) introduced an interpersonal, functional model grounded in evolutionary theory. The sociometer hypothesis proposes that state self-esteem acts as a social barometer that monitors people's inclusionary status. Specifically, low state self-esteem corresponds to the negative feelings that people experience when they are rejected or excluded in a particular situation. Conversely, high state self esteem is related to the positive feelings that follow perceptions of acceptance. As

19 described in social exclusion theory, the anxiety or negative feelings that follow rejection function as a signal to the individual that his or her inclusionary status may be in jeopardy. People are motivated to maintain an adequate level of state self-esteem because high esteem represents the fulfillment of belongingness or acceptance needs. The sociometer hypothesis identifies state self-esteem as a functional system that is based on interpersonal factors, unlike traditional models that assume self­ esteem is a predominantly intrapsychic variable. The motive to avoid low state self-esteem represents an adaptive tendency to want to belong to groups and to maintain important social relationships (Leary & Downs, 1995). Research that provides evidence to support the sociometer hypothesis continues to build. Although extant studies designed to examine predictions based on the sociometer hypothesis are relatively new, results have been consistent and supportive. The following represents some of the research that has been seminal to the continuing development of the sociometer hypothesis. Social Interactions and Self-Esteem Events that represent threats to self esteem are also related to social rejection. Terdal and Leary (1990) asked subjects to rate the degree to which others would accept or re je c t them i f they had performed each of a list of socially desirable or undesirable behaviors. For example, items such as "I cheated on an exam" or "I donated blood" were given ratings based on a continuum that ranged from being totally rejected by others to totally accepted. Subjects

20 also responded to the same items by rating how they would fee l (proud- ashamed; valuable-worthless) if they had performed each behavior. Results showed a canonical correlation of .70 between subject's judgments of others' reactions and their self-esteem ratings. As predicted by the sociometer hypothesis, actions that were perceived to lead to social exclusion corresponded with variations in subjects' reported se lf-fe e lin g s. However, people's b e lie fs about the consequences of exclusionary behavior do not necessarily correspond to

how they would actually respond to real rejection. Personal exclusion and s ta te self-esteem Perceptions of rejection correspond to decreases in state self­ esteem. For example, people who are asked to imagine being rejected (Spivey, 1990), to re c a ll the la s t time they were rejected (Tambor & Leary, 1993) or to perceive th a t they have been rejected (Downs, 1993; Terdal & Leary, 1991), consistently report lower levels of state self­ esteem. Downs (1993) measured state self-esteem several weeks prior to, and immediately after subjects received positive, negative, or no feedback from an "interaction partner" who was ostensibly in another room. Subjects received written feedback from the partner after the subject had selected and answered questions over a microphone from a mildly disclosing list of topics. The feedback that subjects received was, in fact, created by the experimenter and was designed to be highly accepting, moderately accepting, or mildly negative. Results showed that self-esteem difference scores between time one and time two did not significantly differ for subjects who

21 received positive or no feedback. However, subjects who received

rejecting feedback reported lower self esteem scores after receiving the negative feedback. Because subjects in this study were thoroughly debriefed about the contrived nature of the feedback, the duration of self-esteem change following naturally occurring rejection is unclear.

However, i t is lik e ly th a t i f people frequently perceive s e lf directed rejection, the cumulative impact of state effects would transfer into more chronic, trait levels of low self-esteem. Trait Self-Esteem and Exclusion

Trait self-esteem is characterized by chronic expectations of rejection. Therefore, trait self-esteem should reflect people's beliefs about the extent to which they are typically rejected or excluded. Substantial support for the relationship between trait self­ esteem and rejection can be gleaned from the literature. First, trait self-esteem may moderate how people process social information, particularly information related to social acceptance or rejection (Leary, 1990). For example, because trait esteem represents people's positive and negative self-descriptions (Smelser, 1989), these perceptions may affect how they believe that others perceive them (Downs, 1993). Because trait self-esteem is formed over time and is related to people's social experience, the degree of parental acceptance perceived by an individual should be an important source of his or her adult, global self-esteem (Pelham & Swann, 1989). In fact, reports of parental rejection appear to contribute significantly to low self-

22 esteem (Coopersmlth, 1967), particularly for women with eating disorders (de Groot & Rodin, 1994).

A Dysfunctional Sociometer As discussed earlier, a normally functioning sociometer preconsciously alerts individuals that their inclusionary status is at ris k and stim ulates corrective behavior. However, the sociometer may temporarily malfunction, or may not function normally, for some people. For example, a person's sociometer may be "improperly calibrated" and either hypersensitive or hyposensitive to exclusionary

cues (Leary & Downs, 1995). It may be that some people who struggle with chronic feelings of low self worth may have a sociometer that interprets benign

interpersonal feedback as rejecting or disapproving. In fact, people with psychological disorders characterized by low self-esteem may have malfunctioning sociometers that are "calibrated" differently compared to normal populations. Individuals with eating disorders may represent one group with a miscalibrated sociometer. One indication of the relevance of the sociometer system to individuals with eating disorders is that most, if not all, clearly report significantly lower levels of self-esteem compared to other groups (de Groot & Rodin, 1994). Eating Disorders and the Sociometer Individuals, particularly women, who struggle with eating disorders represent a large and growing population in the United States. Although statistics vary, estimates of the prevalence of anorexia nervosa (AN) range from .2 to .7 percent of American women

23 whereas bulimia nervosa (BN) represents about 1 to 3 percent (deGroot

& Rodin, 1994; DSM IV, 1994). Also, as many as 15 percent of female middle-class college students are estimated to have eating disorders (Carson & Butcher, 1992), and the number of women who manifest subclinical levels of eating disorders is estimated to be much higher

(deGroot & Rodin, 1994).

Thg Ç11n,1ÇâJ-E.i£fatcs According the the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV) eating disorders are divided into Anorexia Nervosa

(AN) and Bulimia Nervosa (BN). Also, AN is divided into two subtypes; A restricting type and a binge-eating/purging type. The binge- eating/purging type, sometimes labeled "bulimarexia" (Dunn & Ondercin,

1981), is distinguished from the restricting type and classic BN by the presence of binging and purging in addition to periods of restricted eating. Individuals with these disorders share a common desire to be thin, but AN is characterized by excessive weight loss primarily through dieting and exercise. Individuals with BN are not necessarily underweight and eating is characterized by a pattern of binging followed by efforts to prevent weight gain such as self­ induced vomiting, an excessive use of laxatives, diuretics, enemas, or by other means (APA, 1994). The etiology and full diagnostic criteria used to differentiate AN from BN will not be presented here (see APA DSM IV, 1994). Research and theory suggest that the development of eating disorders involves a complex interaction of biological, psychological and cultural variables; however, the etiology and course of disordered

24 eating remains unclear (Dunn & Ondercin, 1981; Lask & Bryant Waugh,

1992). One way to c la rify the nature of a disorder is to compare the behavior In a standardized situation of individuals who have the disorder to those who do not. With this in mind, the following review Is designed to support the application of the sociometer hypothesis to women with eating disorders. Eating disorders are clearly more predominant In women than In men (Carson & Butcher, 1992; OSM IV, 1994). Although this difference might be attributed to genetic factors (Lask & Bryant-Waugh, 1992), gender differences In socialization (Worell & Remer, 1992) and a cultural emphasis on thinness In women provide evidence for causal factors that are highly Interpersonal (Crandall, 1988). Eating disorder diagnoses are most common In Industrialized societies that value thinness, and often occur In middle to upper-class families (Carson & Butcher, 1992; OSM IV, 1994). The recent Increase In eating disorders may be strongly related to current social acceptance criteria that equate thinness and beauty with acceptance. Therefore, girls and women with low self-esteem who also believe they do not have the skills or resources to be acceptable In other ways may select an acceptance strategy, such as losing weight, that can be "controlled". In other words, the development of an eating disorder may represent a socially reinforced attempt to fu lfill particularly strong acceptance or belongingness needs.

In addition to other psychological traits, eating disorders are characterized by low self-esteem (Lask & Bryant-Waugh, 1992; Nagel & Jones, 1992; Shlsslak, Pazda & Crago, 1990; Tiggemann, M inefield,

25 Winefield & Goldney, 1994). According to the sociometer hypothesis, individuals Mho perceive that they have been recently excluded will report lower levels of trait self-esteem compared to others (Downs, 1993). The research reviewed herein has provided extensive support for the relationship between eating disorders and low trait self­ esteem, however measures of more transient but present-focused measures of state self-esteem have been neglected. It may be that women with eating disorders are hypervigilant to

interpersonal cues that imply social rejection. De Groot and Rodin (1994, p. 299) have suggested that eating disorders symptoms are related to an "underlying disturbance in the self" which accounts for the feelings of ineffectiveness and perfectionism common to women with eating disorders. They also note that women with eating disorders are more attentive to others' expectations and to disapproving responses. However, the relatio n sh ip of eating disorders to these c h a ra c te ristic s may be explained by a malfunctioning interpersonal monitoring system such as the sociometer. The symptoms associated with eating disorders mentioned earlier contain features that suggest the individual is concerned with others' social acceptance of her. Characteristics such as feelings of ineffectiveness suggest that the individual perceives that others perceive that she is not contributing to the group and thereby at risk for rejection. Women with eating disorders often have an external locus (Williams e t a l ., 1993) and are highly a tte n tiv e to external social cues (deGroot & Rodin, 19994) which supports the p o ssib ility th a t th is

26 disorder is a function of maladaptive perceptual processes about the self. In other words, women with eating disorders may be

hypervigilant to any cues that remotely imply social rejection, and thereby report consistently lower levels of state self-esteem. Perceptions of rejection that recur over time may transform to more stable low self-esteem. Social Componentsjof Anorex.ta-and Bulimia Nervosa Individuals who struggle with eating disorders may represent one of the psychological disorders that is tied to the interpersonal processes that influence self-esteem. Although individuals with eating disorders commonly report lower levels of self-esteem compared to individuals who are obese, moderate dieters, or normal controls (de Groot & Rodin, 1994; Williams e t a l . , 1993) few th eo ries have considered why this is so. Traditional explanations regarding the etiology of eating disorders have considered biological, social, cultural, cognitive, and personality factors as well as maladaptive family dynamics (Crandall, 1988; DSM IV, 1994; Friedlander & Siegel, 1990). Further, despite an accumulation of research on traits and behaviors characteristic of individuals with eating disorders, theory regarding the etiology and maintenance of eating disorders is lacking. Although it is not the purpose of the present paper to devise a comprehensive theory of the origins of eating disorders, this investigation will attempt to provide evidence that eating disorders develop in part from a disturbance in the social cognitive processing related specifically to information connoting social rejection.

27 Specifically, the thought processes and affective components characteristic of individuals who display disordered eating are proposed to generate as a function of the feedback they receive from important early relationships (Friedlander & Siegel, 1990). In conjunction with having learned maladaptive cognitions related to the self within the family, sociocultural factors in the United States are assumed to support and maintain eating disorder symptoms (DSM IV, 1994). Also, the obsession with weight and body image characteristic of women with eating disorders is assumed to represent an underlying desire for social acceptance as evidenced by maladaptive esteem- related affect and cognitions. Purpose gf. Study The present study will compare changes in the state self-esteem of women who report characteristics of eating disorders to women who do not. For both groups, current levels of self-esteem will be measured prior to, and after, experimental conditions of acceptance and rejection. This comparison will be unique in two ways. First, previous research that has examined the sociometer hypothesis has only used indirect means (such as reactions to fictional social situations) to confirm and calibrate the changes in self-esteem that occur following different levels of positive and negative feedback. The present experiment will be the first to examine the sociometer hypothesis in direct, face-to-face interactions. Second, the present study will examine the emotional and cognitive concommitants of eating disorders from both a clinical and social psychological perspective.

28 Specifically, the degree to which women with high EDI scores perceive social feedback differently from others will be examined.

Hmo ttig sss Leary and Downs (1995) proposed that state self-esteem varies as a function of negative interpersonal feedback, which they termed the sociometer hypothesis. In light of this proposal, the first hypothesis for the present study predicts a main effect for acceptance rejection. Hypothesis I. Rejected subjects will report greater changes in state self-esteem scores compared to subjects who were not rejected. Hypothesis 1(a). Rejected participants with high eating disorder scores will reoort significantly greater discrepancies between s ta te s e lf-e s te em scores compared to other orouos. Women with high EDS scores are expected to experience significantly greater decreases in self-esteem following rejection because research suggests that women with eating disorders are more attentive to negative social feedback (deGroot & Rodin, 1994). Hypotheses II. Subjects in the rejection condition will report more negative affect compared to accented subjects.

Hypothesis 11(a). Women With high EDI scores will report significantly more negative affect in the rejection condition compared t o women. With-JoitiPI-scores .who.were, rejected. Hypothesis III. Rejected participants will reoort more negative thoMohts compared to.accepted participants. Hypothesis III (a). Women with high eating disorder scores will reoort more negative t houghts compared to women with low eating

29 disorder scores. It is possible that women with eating disorders hold negative self-schemas, or maintain biased information processing systems, that influence how they think and feel about themselves in relation to others. This is consistent with the assertion that women with eating disorders are characterized by dysfunctional sociometers that may distort social perceptions and self-related cognitions.

30 CHAPTER 2

METHOD

Participants

Seventy six female undergraduate students at a large, midwestern university served as subjects. Subjects ranged in age from seventeen to fourty seven, although ninety one percent of participants were twenty two years old or younger. The mean age for subjects was 19.74 and the mode was 18. The racial and ethnic distribution of subjects included ten African American, two Native American, ten Asian, fourty- nine Caucasian, two Hispanic and one individual who described herself as a composit of multiple races and ethnicities.

The subjects were recruited from introductory psychology courses and earned research credit for participation. All subjects participated in a screening process which involved completing the Rosenberg (1965) Self-Esteem scale and Garner & Olmstead's (1983) Eating Disorders Inventory (EDI). To obtain a general index of subjects' eating disorder behavior, eating disorder scores were determined by summing the subscale items to arrive at a total EDI score, rather than comparing scores on the different subscales. Lower scores indicated fewer eating disorder characteristics. Cutoff scores were determined by dividing the frequency distribution into thirds.

31 This procedure resulted In a “low EDI" cutoff score of 224 and below, and a "high EDI” c u to ff score of 260 and above. Subjects whose scored in the upper and lower thirds of the EDI were invited by phone to participate. The mean score for the "low EDI" group was 192 with a standard deviation of 22.27 and the mean score for the "high EDI" group was 287 with a standard deviation of 18.55. Hart and Ollendick (1985) used a similar scoring procedure in which 12 bulimic and nonbulimic subjects' responses were ranked from 1 to 6 and summed across subscales. Hart and Ollendick (1985) found mean scores of 259.33 (standard deviation * 6.71) and 193.42 (standard deviation - 4.91) which were comparable to high and low EDI groups in the present study. The cuttoff scores of 260 (high EDI) and 224 (low EDI) used in the present study are similar to the mean scores of the bulimic and nonbulimic scores in the study mentioned above. However, although the mean scores indicate that the cutoff scores in the present study are comparable to other studies, the differences in the standard deviations between the two studies appears to be significant. The lower standard deviations found by Hart and Ollendick (1985) may be related to the fact that the "bulimic" subjects in the Hart and Ollendick (1985) study met eating disorder diagnostic criteria which was likely to produce a lower degree of variability among EDI scores. The subjects in the present study were recruited based on their EDI scores, but were not diagnosed according to their eating disorder status. Trait self-esteem scores for the entire sample produced a mean of 37.22 with a standard deviation of 5.95 and were used in later

32 analyses. Two female experimenters and four female confederates ran an equal number of subjects in each condition.

Instruments Rosenberg's fl965i Self-Esteem Inventory. Rosenberg's trait self-esteem measure consists of 10 items that measure relatively stable global self-feelings, with 4-point scales ranging from strongly disagree to strongly agree. Typical items Include "On the whole, I am satisfied with myself" and "All in all, I am inclined to think I am a failure." Fleming and Courtney (1984) reported a Cronbach Alpha of .88 and provided evidence for convergent validity by reporting negative correlations between high self-esteem and concepts such as anxiety

(-.64) and depression (-.54). Also, support for discriminant validity was provided by a lack of relationships found between self-esteem and age, gender, birth order, and marital status (Fleming & Courtney, 1984). A copy of this scale may be found in Appendix A.

Garner & Olmstead's ( 19831 Eatj no Disorder inventory (EDI). The EDI measures cognitive and behavioral dimensions that can be used to differentiate subgroups of women with eating disorders (ie: Anorexia and Bulimia Nervosa). The instrument consists of eight subscales including Drive for Thinness, Bulimia, and Ineffectiveness. Research indicates that EDI subscales have good internal reliability scores ranging from .83 to .92 in combined samples of women with anorexia, bulimia, or bulimarexia (Garner & Dlmstead, 1983). Test-retest reliability for non-patient samples range from .67 to .95 for one week, .65 to .92 fo r three weeks and .41 to .75 fo r one

33 year after taking the EDI for the first time. Also, the EDI demonstrates criterion-related validity because it has been shown to discriminate between eating disorder and nonclinical samples (Garner & Dlmstead, 1983). Concurrent validity is provided by research showing that EDI scores correlate significantly with expert ratings from clinician's familiar with each patient. Because the EDI demonstrates good reliability and validity, it is a frequently used measure in eating disorder research. A copy of the EDI can be found in Appendix B.

Heatherinoton & Polivv's (19911 State Self-Esteem Scale fSSESl. The SSES is a 20 item inventory that measures people's self-esteem in the present time and situation. Although questions are similar to items found in measures of trait self-esteem, the directions ask respondents to answer according to how they feel righ t now. The SSES has been found to correlate with performance, social, and appearance self-esteem. Typical items include "I feel confident about my abilities" and "I am worried about what other people think of me". Through a series of five studies, Heatherington and Pol ivy (1991) report that the scale demonstrates high internal consistency (coefficient alpha * 0.92) and a robust factor structure that includes performance, social, and appearance related self-esteem. The SSES can also reliably discriminant between mood and state self-esteem, and shows construct validity. A copy of the SSES is located in Appendix C. Affective Reactions. To assess emotional responses to the manipulation, subjects were asked to rate themselves using 7 point

34 scales on a se ries of 15 bipolar ad jectiv es. Examples of bipolar

items included “tense versus relaxed" and “secure versus insecure". To limit response bias, items that began with positive adjectives were alternated with negatively worded items. Downs (1993) reported a Cronbach's alpha of .91 for the same 15 affective adjectives using 7- point bipolar scales. A copy of the affective reactions scale can be found in Appendix 0. Cognitive Reactions. Cognitive reactions were assessed by using a thought listing questionnaire that asks subjects to list thoughts within 10 1 X 15 centimeter rectangles without concern for spelling or grammar. In stru ctio n s on the thought lis tin g questionnaire encouraged subjects to record any thoughts they were having during the interaction. Later, two experimenters rated an equal number of statements as positive, neutral or negative. To categorize each written statement as positive, neutral, or negative, positivity scores were assigned a number one, two, or three respectively. Ratings were not determined according to the specific content of each statement, so that topics related, or unrelated, to the study itself were scored. To ensure a high level of inter rater reliability, each rater rated the same 95 responses which revealed an inter-rater reliability of 89 percent. Items that were not rated identically were discussed and resolved, and the remaining responses were rated independently. A copy of the thought listing questionnaire can be found in Appendix E. Procedure

Each session consisted of one experimenter, one subject and two confederates. Based on pre screening EDI scores, subjects were

35 recruited by phone for the study which was titled "Social Interactions”. When subjects arrived at the laboratory, they were asked to complete an informed consent form (see Appendix F). Next, subjects were directed to a quiet room to complete a packet of questionnaires ostensibly for "a different study". Packets included various fille r items, including requests for demographic information, ratings of the quality of university life, as well as the informed consent form for the interaction study (Appendix G). Embedded in the questionnaire packet was Heatherton & Polivy's (1991) measure of state self-esteem which was included to gauge subjects' self-feelings prior to the experimental manipulation. Subjects were told that the study was concerned with how people form impressions of others, and that they would interact with two other subjects. Next, the experimenter introduced the subject to the confederate subjects who had supposedly been filling out forms in separate rooms. All participants were led to a room equipped with video recorders which was designated for the experiment.

It is important to mention that, to ensure the quality and uniformity of each trial, confederates received several hours of training. Training included practicing, and refining, the roles of "experimenter" and "subject" to maximize the degree to which confederates' actions remained consistent across trials. Also, training included the development of a short script that was memorized by the "experimenter" who provided the directions throughout each trial. After the confederates role-played trials with each other, they received more practice by running a pilot study with individuals

36 who had been prescreened, but who had not fit the selection criteria of scoring either high or low on the EDI. The pilot study provided a means for confederates to run several trials prior to the study proper.

Finally, experimental sessions were randomly videotaped and reviewed by the researcher for homogeneity. Inconsistent procedures were discussed with the confederates, and corrected. The resultant procedures performed by the confederates across trials included the

follow ing. Once seated in the room, each participant was asked to draw a slip of paper from a container and to read the selection out-loud to indicate who would be the "asker" and who would be the two "respondents". The drawing itself was predetermined so that the subject and one confederate always served as the respondents, and the other confederate served as the asker. After presenting information regarding the role of each participant, the experimenter provided a list of questions to the asker and left the room. The asker then alternated asking questions from the subject and the confederate for a period of five minutes. Questions consisted of inquiries about everyday events rather than more disclosing subject matter, to minimize the potential for the questions to imqpact subjects' reactions. The experimental manipulation was implemented by adjusting the asker's reactions to each respondent to correspond with the experimental condition. For example, in the rejection condition, the asker's reactions to the subject denoted disinterest and moderate disapproval (i.e., avoiding

37 eye contact, not smiling or nodding). In contrast, the asker responded more favorably (i.e., smiling, verbalizing agreement) toward the other confederate "respondent". During the acceptance condition, the asker responded positively to the subject's answers, and less favorably to the confederate.

To help convince subjects of the authenticity of the interaction, the "asker” became progressively more positive or negative as each "respondent" answered questions about herself over the five minute period. The confederate "respondent" was included to provide subjects with a frame of reference for the asker's behavior. In other words, the difference in the asker's behavior toward the confederate "respondent" was designed to prevent subjects from attributing the asker's positive or negative feedback to the asker's mood state or personality. To minimize the degree to which the respondent confederate influenced the impact of the manipulation, the respondent- confederate's interactions with subjects were limited to brief, neutral statements. Also, the confederate "respondent" reacted moderately positively or neutrally to the "asker's" positive or negative reactions toward her respectively. Following the five minute period, the experimenter returned and announced that all participants would be directed to separate locations to complete a packet of questionnaires. Packets began with the SESS, followed by the affective rating and thought listing questionnaires.

38 Subjects were thoroughly debriefed and thanked for their participation. In addition, all subjects were provided a list of mental health providers and resources. An analysis of variance (ANOVA) was used to examine whether subjects perceived that the "asker" was accepting or rejecting.

Further, a series of analyses of covariance (ANCOVAs) was used to test the hypotheses stated earlier, using trait self-esteem as the covariate.

39 CHAPTER 3

RESULTS

Manipulation Check A manipulation check was used to assess the efficacy of the experimental manipulation. Subjects rated seven items on 5-point scales with ratings ranging from "not at all" to "extremely". Questions rated the positivity of the asker and "other subject's" reactions to the participant during the 5-minute interaction. Negatively worded items were reverse scored. Three of the seven questions were designed to assess the degree to which the "asker" was perceived as accepting or rejecting. The remaining items asked similar questions about the respondent-confederate, and were included to minimize possibility that subjects would deduce the purpose of the experiment prematurely (see Appendix H). An analysis of variance (ANOVA) of subjects' perceptions of the askers' tendency to be accepting or rejecting indicated that subjects were convinced of the authenticity of the experimental manipulation and reported experiencing either favorable or unfavorable feedback, £(1, 73) « 24.88, p < .0001. The means scores and standard deviations for the accepted and rejected groups were 64.89 (sd-8.62)

40 and 53.94 (sd-10.27) resp ectiv ely with higher numbers indicating greater negative perceptions of the asker. Reactions to Feedback

State Self Esteem. State self-esteem ratings were summed, and negative items reverse scored. An examination of the interitem reliability for the 20 state self-esteem items completed both prior to, and after, the experimental manipulation revealed standardized Cronbach's alphas of .91 and .95 respectively. These results indicate that the alpha coefficients for each administration were sufficiently high to warrant summing the responses of each state self-esteem administration to obtain an overall measure of subjects' current self- feelings both before and after the manipulation. A 2 (EDI score: high, low) x 2 (Condition: acceptance, rejection) repeated measures analysis of covariance (ANCOVA) was conducted using tra it self-esteem as the covariate and state self-esteem change scores as the dependent variable. Trait self-esteem was used as a covariate because it has been shown to correlate with (parental) acceptance and rejection (Coopersmith, 1967). Results revealed a significant main effect of condition, E(l,71) = 6.16, p < .02, but no main effect of eating disorder status, p > .54 (see Table 3.2). Table 3.1 contains the condition means and standard deviations for the state self-esteem difference scores, along with the other dependent measures. Inspection of the mean change scores indicated that subjects who were accepted reported higher levels of state self-esteem (M * 2.47) whereas subjects who received rejecting feedback rated themselves less positively (H * -2.14). No significant 41 interactions were found. Specifically, the results were not qualified by subjects' high or low scores on the eating disorder scale.

42 CONDITION

ACCEPTANCE REJECTION Hiah EDI Low EDI Hioh EDI Low EDI State Self-Esteem Difference Scores 3.05 -2.71 -1.59

U i i * ' Affective Ratings** 45.5 f îjjM ) n>18 n»17 Thought Listing Positivity** 1.74 2.44 2.24

1 1 » U S > W#' U U > Trait Self-Esteem*** 37.23 (5.95) * Standard deviations are in parentheses. ** Greater scores on affective and cognitive ratings indicate more negative feelings and thoughts respectively. ***Trait self-esteem covariate mean and standard deviation

Table 3.1: Means and standard deviations.

CONDITION Acceptance Rejection SSES 68.40 (13.71)* 73.25 (14.33) Affect 42.10 (12.31) 50.90 (19.24) Thoughts 1.84 (.40) 2.34 (.37)

* Standard deviations are in parentheses. Table 3.2: Grand means and standard deviations by condition.

Eating Disorder Status High Low SSES -.41 (3.42) .73 (9.53) Affect 49.90 (14.80) 45.84 (16.68) Thoughts 2.19 (.39) 1.99 (.38)

* Standard deviations are in parentheses. Table 3.3: Grand means and standard deviations by status.

43 SELF-ESTEEM

SOURCE DF SS F P Condition 1 395.05 6.16 .01 EDI Status I 24.21 .38 .54 Condition * Status 1 .02 2.21 .99 Residual 71 4553.56

AFFECT

SOURCE DF SS F P Condition 1 1417.38 5.20 .03

EDI Status 1 824.64 3.03 .09

Condition *Status 1 304.53 1.12 .29

Residual 70 19084.29

THOUGHT LISTING POSITIVITY

SOURCE DF SS F P Condition 1 4.58 31.14 .0001

EDI Status 1 .73 4.95 .03 Condition *Status 1 3.43 .002 .96 Residual 71 10.45

Table 3.4: ANOVA Tables for Self-esteem Difference Scores, Affective Sums, and Thought Listing Positivity Scores.

44 Affective Reactions

Subjects provided 7 point ratings on 15 bipolar affective adjectives related to self presentation. Negative items were reverse scored, and a Standardized Cronbach's alpha coefficient of .93 warranted summing the affective ratings for a total positivity score. Higher scores reflected greater negative affectivity. A 2 (EDI scores: high, low) x 2 (Condition: acceptance, rejection) analysis of covariance (ANCOVA) was conducted to examine the sums of subjects' affective responses to acceptance and rejection, with trait self­ esteem used as the covariate. The means and standard deviations for the affective ratings can be found in Table 3.1. A significant main effect of condition was obtained £(1,70) = 5.2, p < .03, but a main effect of status was not significant F (1,70) = 3.03, p = .09. Subjects in the rejection condition reported experiencing more negative emotions following the social interaction (M = 50.9) compared to subjects who were accepted (M - 42.1). No significant interactions were found. Cognitive Reactions

Statements were assigned a one, two, or three for positive, neutral, and negative statements respectively so that lower scores were most positive and higher scores most negative. Thought positivity was examined by performing a 2 (high versus low EDI scores) X 2 (acceptance versus rejection) analysis of variance on the averaged sums of the scored statements. Table 3.1 contains the means and standard deviations for the cognitive reactions.

45 Significant main effects were found for both status £ (1, 71) = 4.95, p < .03 and condition £ (1, 71) « 31.14, p < .0001 (see Table 3.4). Subjects thought more positively or negatively based on their eating disorder score, and whether they were accepted or rejected. Subjects with high eating disorder scores reported qualitatively more negative thoughts (M * 2.19) compared to subjects who had low EDI scores (M = 1.99). Also, subjects in the rejection condition reported more negative thoughts (M « 2.34) compared to subjects who were accepted (M = 1.84). No significant interactions were found.

46 CHAPTER 4

DISCUSSION The objectives of the present study were twofold. First, propositions based on the sociometer hypothesis were examined in face- to-face interactions. Whereas previous research has demonstrated indirect support for the contention that state self-esteem changes occur in tandem with perceptions of being accepted or rejected, the present study tested whether these results would generalize to "in- vivo" social interactions. Second, the hypothesis was tested that women who reported more eating-disorder symptoms relative to other participants would respond to rejection with greater changes in state self-esteem, negative affect, and negative cognitions. The interpretation and discussion of the results will begin with a summary of the findings and how they relate to the hypotheses. Following the summary, limitations of the study will be reviewed, as well as suggestions for future research in this area. Because low self-esteem and eating disorders are problems that are frequently presented to clinicians, the implications of the present study for counseling will also be addressed. The present study provides support for the hypothesis that state self-esteem varies with people's perceptions of the degree to which they are accepted by others. As predicted by the sociometer

47 hypothesis, subjects who received rejecting feedback reported more negative self-evaluations, compared to subjects who were accepted.

Also, accepted and rejected subjects reported more and less positive thoughts and feelings, respectively. Contrary to the hypotheses, eating disorder scores did not moderate state self-esteem or affective reactions, although rejected subjects with high eating disorder scores listed significantly more negative thoughts following rejection compared to other groups. So.ç1al J-gg-dback.and. State. These results are consistent with theory and research connecting esteem deflating events such as social exclusion or task failure with negative affect including social anxiety, loneliness, jealousy, and depression (Baumeister & T ice, 1990; Downs, 1993; Leary, 1990; Tambor & Leary, 1990; Terdal & Leary, 1990). As noted above, support was found for the hypothesis that state self-esteem would change after subjects believed that they had been accepted or rejected. Also, the mean self-esteem score of each group fell in the expected direction. Specifically, whereas the mean score for accepted subjects increased following the interaction, rejected subjects reported a decrease in self-esteem. One explanation for these results is that state self-esteem fluctuates according to individuals' immediate perceptions of their social standing (Leary & Downs, 1995; Terdal & Leary, 1990). The self-deprecation that appears to follow rejection, and the negative feelings th a t accompany i t , may serve as a functional "alarm system" that motivates individuals to attend to their behavior to curtail or

48 avoid exclusion (Baumeister & Tice, 1990). Also, according to the sociometer model, the negative feelings that have been attributed to low self-esteem, are a result of perceived social exclusion and not low self-esteem per se (Leary, Schreindorfer & Haupt, 1995). Overall, the present results provide support for the proposal that state self­ esteem changes in accord with people s perceptions of themselves as included or excluded. Support for the reactivity of state self-esteem, affect, and cognitions to interpersonal feedback is not new. Leary and his colleagues (Leary & Downs, 1995; Leary et al, 1995) have demonstrated support for the sociometer hypothesis by showing that changes in self­ esteem in response to rejection are not only consistent, but vary based on the intensity or severity of the rejecting feedback. In other words, state self-esteem does not respond to a specific threshold of rejecting feedback, but increases and decreases at different rates based on the intensity of the rejecting feedback (Terdal & Leary, 1990; Leary, Haupt, & Chokel, 1994). In fact, future research may wish to examine whether individuals with eating disorders show a different sociometer pattern in real or imagined social situations, compared to the sociometer calibrations based on normal populations (Leary, Haupt, & Chokel, 1994).

Other studies have examined the sociometer hypothesis by asking subjects to respond to hypothetical interpersonal vignettes (Terdal & Leary, 1990), lists of behaviors that range from inclusory to exclusory (Tambor & Leary, 1990), or to written feedback generated from anonymous, one-way in te rac tio n s (Downs, 1993). The present study

49 differs from, and extends, previous findings in support of the sociometer hypothesis to subjects in face-to-face social interactions. Because the hypotheses derived from the sociometer model were supported in an actual social encounter, the resulting changes in self-esteem appear to support the contention that predictions based on the sociometer hypothesis generalize to real-life Interactions. In addition to interactions on paper, the exclusionary feedback that people receive during live Interactions seems to have a significant impact on their evaluative self-perceptions. A second explanation for the current findings is self- presentational. Individuals often compensate for social failure by trying to present themselves in strategically more or less positive ways (Baumeister & Jones, 1978). For example, people who believe that they have failed at a social task frequently attempt to present a more modest or agreeable image (Schneider, 1969). One reason that individuals may compensate self-presentationally for social failure is to salvage or enhance the potential to obtain social resources (Jones & Pittman, 1982; Schlenker & Leary, 1982a, 1982b). Subjects who were rejected in the present study may have wanted to appear modest to counteract or neutralize the impact of being rejected on their self- image, thereby enhancing opportunities to acquire social rewards such as influence or friendship. Although subjects may have been reacting self-presentationally to social rejection, this explanation is unlikely because the directions provided throughout the experiment indicated that only a single social interaction would take place. In fact, with only a few minutes left in the trial following the

50 interaction, subjects probably believed that they would not interact again. Therefore, whereas a self presentational explanation for more negative self-descriptions following social failure is plausible, it seems unlikely. Eating disorder scores and self-esteem Contrary to expectations, EDI scores did not moderate subjects' transitory self-evaluations. Subjects with high EDI scores did not report greater changes in state self-esteem compared to subjects with low EDI scores. Greater changes in state self-esteem scores were anticipated from subjects with high EDI scores because their state self-esteem system was hypothesized to be hypersensitive to evaluative interpersonal feedback. Eating disorder scores may not have moderated state self-esteem because the symptoms associated with individuals diagnosed with eating disorders are unrelated to how they process and respond to social feedback. However, because experts generally have agreed that eating disorders have a strong social component (deGroot & Rodin, 1994; Garner & Olmstead, 1984), correlate negatively with self-esteem (task & Bryant-Maugh; Williams et a l., 1993), and fear social disapproval and rejection (Dunn & Ondercin, 1981), this explanation seems unlikely. Rather, the failure of EDI scores to moderate subjects' self-esteem reactions in the present study may be a result of several fa c to rs. First, participants in the present study did not represent a clinical sample. Relative to the EDI scores of the entire sample of prescreening participants, subjects were recruited based on whether

51 their EDI scores were In the upper or lower third of the sample. Although subjects who scored in the upper third of the sample may report a greater number of eating disorder traits or behaviors compared to others, th is does not id en tify high scoring subjects as having a clinical eating disorder. The EDI is designed to be used as a tool in the context of a full assessment, not as a diagnostic instrument (Garner & Olmstead, 1984). Therefore, it may be that the present sample of subjects did not reflect the self-esteem reactions that may have occurred in a sample of individuals diagnosed with clinical levels of eating disordered behavior. A second possibility is that the manipulation check was sufficiently potent to impact both high and low scoring EDI groups equally, preventing more subtle differences to emerge. Although the present manipulation was designed based on previous research that implemented written interpersonal feedback described as moderately rejecting (Downs, 1993), it may be that "moderately" rejecting feedback has a greater impact in face-to-face situations. If so, the accepting and rejecting feedback may have been perceived as relatively obvious by both groups, which led to a similar decrease in state self­ esteem for both groups.

Affective .flMfitlons As predicted, subjects who were included or excluded felt better or worse after the interaction, respectively. Subjects in the rejection condition reported experiencing feelings such as anxiety, embarrassment and in secu rity compared to subjects who were accepted. Contrary to expectations, EDI scores did not moderate subjects'

52 feelings, although mean scores were in the expected direction.

Subjects who scored high on the EDI did not report significantly more negative feelings following rejection (p » .09). As noted earlier, social exclusion theorists posit that the negative affect that accompanies social rejection is an adaptive reaction th a t promotes group membership (Baumeister & Tice, 1990; Leary, 1990). However, it is important to clarify that the negative affect examined thus far by social exclusion theorists has been primarily, if not exclusively, related to feelings that are essentially interpersonal such as anxiety, loneliness, and jealousy (Leary, 1990).

In contrast, Watson, Clark and Tellegen (1988) defined negative affect as "a general dimension of subjective distress and unpleasurable engagement that subsumes a variety of aversive mood states, including anger, contempt, disgust, guilt, fear, and nervousness" (p. 1063). Although the measure of negative affect used in the present study was designed to examine feelings related to social inclusion or exclusion, future research might examine the extent to which subjects' affective responses are primarily interpersonal, or are more accurately described by measures of global negativity such as the PANAS scales developed by Watson and his colleagues (Watson, Clark & Tellegen, 1988). Contrary to the hypotheses, the data failed to show that high scores on the EDI coincided with significantly more negative affect, compared to subjects with low EDI scores. Similar to subjects' self­ esteem reactions, it may be that the negative feedback was so obvious

53 that ceiling effects prevented a distinction between the two groups.

Given more subtle acceptance or rejection, women with high EDI scores may have reported more negative affect following rejection compared to other subjects.

A second explanation is that subjects with high EDI scores dfd experience greater levels of negative affect, but were not able to report it. Women with eating disorders tend to be less aware of, and unable to identify, their emotional and psychosomatic experience (deGroot & Rodin, 1994). In fact, the objectives of treatment programs designed to facilitate eating disorder recovery often include helping patients identify and experience their emotions (Scarano & Kalodner Martin). Therefore, subjects with high scores on the EDI may also be less conscious of their affective experience, and unable to accurately report the extent of their negative feelings. Of course, to some degree all subjects reported were able to report feeling different emotions in response to acceptance and rejection, but it is possible that subjects with higher scores on the EDI responded in accord with the muted emotional experience characteristic of women with eating disorders. While women with eating disorders appear to dissociate from affect, they may be more able to access their cognitions. Subjects with high eating disorder scores in the present study reported more negative thoughts compared to women with low scores.

54 Coflnjtl^g. Reactions The thought-listing ratings were, in part, consistent with the state self esteem and affective scores. As predicted, accepted su b je cts' thoughts were more positive during the in teractio n compared to subjects who were rejected.

Participating in a social interaction may have stimulated subjects' memories of prior social situations that were accepting or rejecting. Theory and research on social cognition suggests that people store information in categories via hypothetical cognitive structures called schema (Fiske & Taylor, 1991). When a particular schema is "activated" by cues in the environment, it guides individuals' current perceptions according to the individual's past experience with similar situations. Similarly, the experimental conditions of acceptance or rejection may have stimulated behavior that initiated a biased scanning process. Biased scanning is a cognitive process that explains the relationship between people's self conceptions and their social behavior (Jones, Rhodewalt, Berglas, & Skelton, 1981). In the present case, by "scanning" for similar events in the past, people's immediate self-conceptions formed according to how they thought of themselves, or the interactants, during past inclusionary or exclusionary experiences. Inspection of the content of the cognitive statements suggests that responses that were rated "negative" were directed either toward the subject, or toward one or both of the confederates. It may be that the degree to which subjects attribute their positive or negative social experiences to external, versus internal, factors differs as a

55 function of self-esteem or eating disorder status. Future research would benefit by categorizing statements according to a variety of categories, including to whom the statement is directed. For example, subjects with high EDI scores may have written more hostile comments toward the self compared to other subjects. In fact, individuals with eating disorders have been shown to demonstrate a greater degree of self directed hostility (Williams, et ai., 1993). However, because the present study did not address the target of subjects' thoughts, future research is needed to examine specifically how subjects consciously think about their exclusionary experiences. Unlike the results for state self-esteem and affect, EDI scores did moderate subjects' cognitive responses. Subjects with high EDI scores reported significantly more negative thoughts compared to subjects with low EDI scores. One reason for these findings is that individuals with higher EDI scores may have been more attentive to the negative feedback. This explanation is consistent not only with the current hypotheses, but with theories that suggests girls and women with eating disorders demonstrate a particularly high sensitivity to the approval or disapproval of others (deGroot and Rodin, 1994; Dunn & Ondercin, 1981). It seems that the tendency of subjects with high EDI scores to report their reactions to negative feedback was limited to cognitive variables. Measures of state self-esteem and affective were unable to detect differences between the two groups because both instruments measured affective, not cognitive, information. Although self-esteem

56 may be argued to include cognitive components, most researchers agree that self-esteem is a primarily affective construct (Brown, 1993). A second explanation resides in the structure of the cognitive thought listing task. Unlike the other measures, subjects responses were less restricted, and they were free to record any thoughts that had occurred during the experiment. Because individuals with eating disorders experience less perceived control compared to normal eaters (Scarano & Kalodner-Martin, 1994), the g reater number of negative statements provided by subjects with high EDI scores may have represented an attempt to exercise control following the experience of social rejection. A third explanation is that women with high EDI scores may be generally more negative compared to other women. Research has shown a tendency toward depression among eating disordered populations (task & Bryant-Waugh, 1992), which became m anifest in the present study through a free response measure related to cognition. Of course, it may be that state self-esteem and the particular affective measure used in the present study simply did not capture relevant differences between the two groups. Future research may benefit from including a variety of affective, self-esteem, and cognitive measures to detect differences between the affective and cognitive processes that may exist among individuals with various degree of eating disordered behavior. Limitations and Suggestions for Future Research The limitations of the present study should be considered when interpreting the results. First, although the experimental

57 manipulation was effective for subjects with high and low EDI scores, it may have been so potent that ceiling effects prevented real differences between the two groups to emerge. The present study included only two conditions; acceptance and re je c tio n . However, exclusory feedback is not dichotomous, but is more accurately represented as a continuum (Leary, 1990). It may be that the differences in self-esteem and affect between subjects with high and low EDI scores may have emerged had experimental conditions lower, or higher, on the exclusory continuum been included. Future research may benefit by testing subjects' responses to experimental conditions that represent a range of accepting and rejecting feedback, with care given to the operational definitions of

conditions and experimental design. For example, subjects with high or low EDI scores might be asked to read, and respond to, social scenarios that vary from low to high degrees of rejection. Social interactions that occur "on paper" may increase the control over the experimental manipulation and allow significant differences in affect, cognition, and behavior between groups to emerge. Second, the hypotheses were based on the assumption that subjects with relatively high scores on the EDI would mimic, at least to some extent, the thoughts and feelings of women who have been diagnosed with an eating disorder. However, i t is important to acknowledge th a t subjects with high EDI scores were not recruited from a clinical population. As such, the present findings may not generalize to women who meet the diagnostic criteria required for an eating disorder diagnosis. Future researchers interested in whether the sociometer

58 functions differently for individuals with eating disorders may benefit from obtaining true clinical samples. Therefore, although partial support was found for the hypothesis that individuals with eating disorder characteristics would respond more negatively to social rejection, more research needs to be done with clinical populations. Third, although the state self-esteem, affective, and cognitive measures used in the present study were judged to be sufficient to explore the hypotheses, it is possible that the measures were not sensitive to real differences between subjects with high and low EDI scores. Future research may profit by including several affective and cognitive measures that vary in structure or content. The present study found that rejected subjects with high EDI scores reported more negative reactions on an open-ended thought listing task compared to subjects with low EDI scores, but not on more structured measures that consisted of 7 or 9-point rating scales. Also, affective measures designed to measure more global feelings such as the PANAS scales (Watson, Clark & Tellegen, 1988) may be more apt to elucidate significant effects. Fourth, the results of the present study were based on the use of female college students taking a psychology 100 course at a large university, and may not generalize to other populations. Similarly, external validity of the present results may be limited due to the structured nature of the social interaction, and the experimental setting. In other words, generalizing the present results to other populations and settings should be done with caution.

59 Imp-UQfttjQn? for Counseling The results from this study suggest that the quality of people's social interactions is one factor that impacts moment-to- moment self evaluations. Subjects in the rejection condition reported a significant decrease in the positivity of their thoughts, feelings, and state self-esteem compared to subjects who were not rejected. Therefore, the treatment of low self-esteem may benefit by applying techniques that ensure an individual's sociometer is calibrated correctly, or if cognitive distortions are preventing a reasonably accurate interpretation of interpersonal events. Once a sociometer has been "recalibrated", clients can be encouraged to become aware of their affective, cognitive, and behavioral reactions to social rejection. First, individuals could interpret negative affect as a "signal" to focus their attention on social factors. Second, cognitive techniques could be applied to teach clients to identify cognitive distortions, and to alter maladaptive thoughts. Finally, the enhancement of social skills would help clients to increase the degree to which they are accepted by others, and thereby increase self-esteem. Although the present findings did not support the hypothesis that rejected individuals who reported eating disorder symptoms would show greater decreases in state self-esteem, the results clearly indicated that women with high EDI scores responded to rejection with significantly more negative thoughts. One implication of these findings is that the cognitions of women with eating disorders may be initially more accessible to intervention compared to direct attempts

60 to address the individual's emotional experience. Many approaches already favor cognitive interventions for the treatment of eating disorders, perhaps because these clients have such difficulty identifying their feelings (deGroot & Rodin, 1994). The present results suggest that cognitive-behavioral approaches that target cognitions related specifically to perceptions of acceptance and rejection may be particularly useful to clients with eating disordered behavior.

The results of the present study provide support for the contention that problems related to low self-esteem may be addressed productively by targeting the social experiences of individuals with negative self-feelings. The results also support the use of cognitive approaches to treat individuals with eating disorders, with the added caveat that the exploration of skewed or distorted cognitions related to social rejection may be an important key to treatment success.

61 References

Adler, A. (1929). The science of liv in g . Garden City, NY: Doubleday/Anchor. Ainsworth. M. (1989). Attachments beyond infancy. American PsvcholoQist. 4i. 709 716.' Alden, L. E. & Phillips, N. (1990). An interpersonal analysis 499^512* and depression. Cognitive Theraov and Research. 14.

Armstrong, J. G., & Roth, D. M. (1989). Attachment and separation in eating disorders: A preliminary investigation. International Journal of Eating Disorders. 8. 141-155. Bandura, A, (1977). Social learning theory. Englewood Cliffs, NJ: P rentice-H all. Baumeister, R. F. & Jones, E. E. (1978). When self-presentation is constrained by the target's knowledge: Consistency and compensation. Journal of Personality and Social Psychology. 36. 608- 618. Baumeister, R. F. & Leary, M. R. (1995). The need to belong: Desire for interpersonal attachments as a fundamental human motivation. Psychological Builetin.117. 497-529. , Baumeister, R. F. & Tice, D. N. (1990). Anxiety and social exclusion. Journal of Social and Clinical Psychology. £, 185-195. Beck, A. T. (1976). ÇQqnitiYe„th9rap.Y.fnd the.emotional disorders. New York: International Universities Press. Betz, N. IE., Wohlgemuth, E ., S erlin g , D., Harshbarger, J . & Klein, K. (1995). Evaluation of a measure of self-esteem based on the concept of unconditional self-regard. Journal of Counseling Development. 74, 76-83. X/ Bowlby. J. (1969). Attachment and loss. Vol 1: Attachment. New York: Basic Books.

. ^ Brown, J.O. (1993)., SeIf-esteem,and self-evaluation: Feeling

Bullock, J. R. (1992). Children without friends: who are they and how can teachers help? Childhood Education. 69. 92-96. 62 Carson, R. c. & Butcher, J. N. (1992). AbnomH] psychology .and modern l i f e i9th e d .). New York: HarperColiins Cooley, C. H. (1902). Human nature and the social order. New York: Scribner's. Coopersmith. S. (1967). The antecedents of self-esteem. San Frans 1SCO: W. H. Freeman.

Crandall, C. S. (1988). Social contagion of binge eating. Journal of_^er_sonallty and Socia^l Psychology. 55 . 588-598. deGroot, J. M. &,Rodin, G. (1994). Eating disorders, female psychology, and the s e lf. Journal Of tho ABBrican.Acid«iiy-g.f Psychoanalysis. 22. 299-317. DeLongis, A., Folkman, S ., & Lazarus, R. S. (1988). The impact of daily stress on health and mood: Psychological and social resources as mediators. Journal of Personality and Social Psychology. 54. 486- 495. gxç3u5.ioniAy.1aLcMBfin5atjQjryTintfrnaJ)MpDowns. D. L. (1993). so]fjJce|.ontatlojLin fSGwwnaaLtQ,agciaJ master' s thesis. Wake Forest University, Winston-Salem, N.C. Dunn, P. K., & Ondercin, P. (1981). Personality variables related to compulsive eating in college women. Journal of Clinical Psychology. 37: 43-49. E llis. A. (1962). Rgasun.aoJ «motion jn.psychotherapy. Secaucus, NJ: Lyle Stuart. Fazio, R.H., Effrein, E. A. & Falender, V. J.^(1981). Self- Sociaf l^svcHo^ogy? i f fcrSOnaJi K W Fiske, S. T., & Taylor, S. E. (1991). Social cognition. New York: McGraw-Hill. Fleming, J. S. & Courtney, B. E. (1984). The dimensionality of self-esteem. II. Hierarchical facet model for revised measurement scales. Journal of Personality and Social Psychology. 46. 404-421. Garner, D. M. & Olmstead, N. P. (1984). Eating-disorder inventory manual. New York: Psychological Assessment Resources.

New YoSrSîüêilda)!®” ’' Bf Mif In wgrYftY life, Hazan, C., & Shaver, P. (1987). Romantic love conceptualized as an attachment process. JJurna.Lftf Personalitv and Social Psvcholoov. 52n 511*524* Heatherton, T. F ., & Pol ivy, J . (1991). Development and validation of a scale for measuring state self-esteem. Journal of Personality and Social Psvcholoov. 60. 895-910. Horney, K. (1945). Our inner conflicts: A constructive theory of neurosis. New York: Norton. 63 Jones, H. H. (1990). Loneliness and social exclusion. Journal of Social and Clinical Psychology. 9. 214 220. Jones, E. E. & Pittman, T. S. (1982). Toward a general theory of strategic self presentation. In J. Suis (Ed.). Psychological geirgoectives on the self (Vol. 1, pp. 231-26È). HiTlsaÏTe: NT3.:

Kernis, M. H. (1995). Efficacy, aoencv. and self-esteem. New York: Plenum Press. Lask, B. & Bryant-Waugh, R. (1992). Early onset anorexia neryosa and related eating disorders. Journal of Child Psychology and Psychiatry. 33. 281-300. Leary, M. R. (1983). Undergtanain9-S.Q£ia.1 ■ap&jotY;. Soçial» personality, and clinical oersoecEives. beverlv Hills; Saoe. Leary,,H. R. (1990). Responses to social exclusion: Social anxiety, jealousy, loneliness, depression, and low self-esteem. Joumaj_aT_Soctal and Clinical Psychology. 9. 221-229. Leary. N. R. & Downs, D.L. (1995). Interpersonal functions of the self-esteem motiye: The self-esteem system as a sociometer. In M. Kernis (E d .). Efficacy, agency, and self-esteem . New York, Plenum.

j % n# ^ «I wiivivv ■ 9 v* i* va th9..sociQie.ten The re.lati 9nghip pgtNggD.iatsrpecMM] .appraisals state.self-esteem. Unpublished manuscript. Winston-baiem, N.C.: forest university. Leary, M. R. & Kowalski, R. M. (1990). Impression management: A literature reyiew and two-component model. Psychological Bulletin. 107. 43-47. Leary, M. R., Schreindorfer, L. S., & Haupt, A. L. (1995). The role of^low self-esteem in emotional and behayioral problems: Why is low self-esteem dysfunctional? Journal of Social and Clinical Ps.Y.ÇhQloqy.1- 14, 297-314. Leary, M. R ., Tambor, E. S ., Terdal* S. K., & Downs, D. L. (1995). The function of self-esteem feelings: Testing the sociometer hypothesis. Journal of Personality and Social Psychology. 68. 518-530. MacLeod, C., Mathews, A., & Tata, P. (1986). A ttentional bias in emotional disorders. Journal of Abnormal Psychology. 95. 15-20. Markus, H. (1980). The self in thought and memory. In D.M. Wegner & R. R. Vallâcher (Eds.), The.Jfijf _in_SOOtaJ. psychology. New York: Oxford U niversity Press. Mead, G. H. (1934). Mind, s e lf and so c ie ty . Chicago: University of Chicago Press. s e i f - e S i a : Nagel, K. L. & Jones, K. H. (1992). Sociological factors in the development of self-esteem. Adolescence. 27. 107-113. Oyserman, D. & Markus, H. R. (1993). The sociocultural self. In Suis, J. (Ed.). .Bsychgjoqiçal pecspgctlYes Qj-tbg-self;-Ih8. jolf.-In social perspective. Hinsdale. NJ: Lawrence Earlbaum Associates. 64 Parkhurst. J. T. & Asher, S. R. (1992). Peer rejection in middle school: subgroup differences in behavior, loneliness and interpersonal concerns. Developmental Psychology. 28. 231-241. Patterson, G. R. & Stoolmlller, N. (1991). Replications of a dual failure model for boys' depressed mood. Journal of Consulting and Clinical Psychology. 59. 491-498. , Pelham. B. W. & Swann, W. B. J r . (19B9). From .self-conceptions to self-worth: On the sources and structure of global self-esteem. Journal_of_Personalitv and Social Psychology. 57. 672-6B0. , „ Reis, H. T., Wheeler, L., Kernis, M.H., Spiegel, N., & Nezlek, J. (19B5). On specificity in the impact of social participation on Psycholoov?^4i^^4S6^4?l Journal of Personality and Social

u Rosers, C. R. (1961). On becoming a person. Boston: Houghton Miff I in .

Sanford, L. T. & Donovan, M. E. (1984). .Woyn,and. Sglf-QStpsa; Kglan)'"fenguin do3s°'^ and feel about ourselves.

Scarano, G. N. & Kalodner-Nartin, C. R. (1994). A description of the continuum of eating disorders: Implications for intervention and research. Journal of Counseling & Development. 72. 356-361. Schlenker, B. R. (1980).. iBBrggsiorLnanagefent.;. The .« If- SrooKS?Co1eCia identity , and interpersonal re la tio n s . Honterev. CA:

Schlenker, B. R. & Leary, N. R. (1982a). Social anxiety and self-presentation: A conceptualization and model. Psychological Bulletin. 92 . 641-669. ^ Schlenker, B. R. & Leary, M. R. (1982b). Audience's reactions to self-enhancing, self-denigrating, and accurate self-presentations. Journal of Experimental Social Psychology. 18. 89-104. Schneider. D. J. (1969). Tactical self presentation after success and failure. Journal_of^Personalitv and Social Psychology. Ü , ZOc-cOC. Sedikides, C., Olsen, N.. & Reis, H. T. (1993). Relationships as natural categories. Journal of Personality and Social Psychology. a l, 71-82. S hisslak, C. M., Pazda, S. L. & Crago, N. (1990). Body weight and bulimia as discriminators of psychological characteristics among anorexic, bulimic, and obese women. Journal of Abnormal Psychology. âi, 380-384. Shraugher, J . S. & Schoeneman, T. J . (1979). Symbolic interaction 1s t view of self-concept: Through the looking glass darkly. Psychological Bulletin. 86. 549-573. Smelser, N.J. (1989). Self-esteem and social problems. In Mecca. A. H.. Smelser. N. J . , & Vasconcellos, J . (Eds.) The social importance of self-esteem. Berkeley, CA: Univ. of California Press. 65 Sours, J.A. (1974). The anorexia nervosa syndrome. International Journal of Psychoanalysis. 55. 567-576. Spivey, E. (1990). Social exclusion as a cq—on factor in

âi V Sullivan, H. S. (1953). Ihfi.lntfifpersonal theoa .of-Bsychiatry. New York: Norton. Swindle, R.H. & Moos, R. H. (1992). Life Domains in stressors, coping and adjustment. In W. B. Haish, R. H. Price, & K. H. Craik

cflmon.factor.Jn.social,anxlett,..lone]Inessa.jealopsy».ggpressM absL t^oresl EntversTiy manuscript. Winston-Salem, nC: Wake

Tedeschi, J. T., & Norman, N. (1985). Social power, self-

! S m S " i p î " n I S W : . Terdai, s., & Leary, M. R. (1990). Se If ;estegii. .and. perce i ved social exclusion. Paper presented at the meeting of the Southeastern Psychological Association, Atlanta. Terdai, S., & Leary, M. R. (1991). Social exclusion, self- esteem. and dysphoria. Paper presented at theimeeting of the southeastern Psycho logical Association, New Orleans. Tice, D. M. (1992). Self-concept change and self-presentation:

Tiggemann, H., Minefield, H. R., Minefield, A. H. & Goldney, R. D. (1994). Gender differences in the psychological correlates of body weight in young adults. Psvcholoov-and-Health. 9. 345-351. Matson, 0., Clark, L. A., & Tellegen, A. (1988). Development and validation of brief measures of positive and negative affect: The PANAS scales. Journal of Personality and Social Psychology. 54. 1063-

Williams, G. J., Power, K. G., Millar, H. R., Freeman, C.P., Yellowlees, A., Dowds, T., Walker, M., Campsie, L., MacPherson, P., & Jackson, M. A. (1993). Comparison of eating disorders and other dietary/weight groups on measures of perceived control, assertiveness, self-esteem, and self-directed hostility. International Journal of ü tin s Dlsordgrst 14. 27-32.

66 APPENDIX A SELF-ESTEEM INVENTORY

67 Name:______SS#:______Age: ____ Sex: M F Phone:______Race/Ethnicity: African American Caucasian (White) American Indian __ Hispanic Asian __ Other______

Please place an "X" on the scale following each statement that is generally most true for you. Please answer as honestly as you can. 1. I feel that I am a person of worth, at least on an equal basis with others. strongly Agree Strongly Disagree 2. I feel that I have a number of good qualities. strongly Agree Strongly Disagree 3. All in all, I am inclined to feel that I am a failure. strongly Agree Strongly Disagree 4. I am able to do things as well as most other people. strongly Agree Strongly Disagree 5. I feel I do not have much to be proud of. Strongly Agree Strongly Disagree 6. I take a positive attitude toward myself. 1 1 1 1 1 1 1 1 strongly Agree Strongly Disagree 7. On the whole, I am satisfied with myself. strongly Agree Strongly Disagree 8. I wish I could have more respect for myself. strongly Agree Strongly Disagree 9. I certainly feel useless at times. 1 1 1 1 1 1 1 1 Strongly Agree Strongly Disagree 10. At times I think I am no good at all. strongly Agree Strongly Disagree Please Note

Copyrighted materials in this document have not been filmed at the request of the author. They are available for consultation, however, in the author’s university library.

Appendices Band 0

UMI APPENDIX D

AFFECTIVE REACTIONS

70 In the next series of questions are designed to assess how you are feeling RIGHT NOW. Please rate your current reactions on the following scales as honestly as you can by placing an "X" in the appropriate blank. Your responses will MQZ b9 ghown. tg-.thg-gfchsr-jgiahiggfr»

tense relaxed strong weak not embarrassed embarrassed angry not angry calm nervous powerless powerful carefree : worried secure : insecure flustered : : composed anxious : not anxious dominant submissive proud ashamed not vulnerable vulnerable poised : humiliated peaceful hostile APPENDIX B

COGNITIVE REACTIONS

71 Now we would like you to describe in your own words what you were thinking as you talking with the other subjects. Below are 12 spaces. Write one statement about what you were thinking per space, ignoring spelling, grammar and punctuation. Do not feel that vou have to fill every space. Just write enough statements or thoughts to give a clear impression of what you were thinking about at the time. Your comments will WOT be shown to the other subject.

3.

6 .

7.

9.

1 0 .

11.

12 APPENDIX F INFORMED CONSENT FORM

72 Infonned Consent Form Purpose and Procedure. This study is designed to investigate how people perceive interact. First, you will speak with two other subjects for about 5 minutes. Next, you will answer questions about the other subjects and record your reactions and observations during the interaction. These questions and answers will not be shown to the other subjects. Risks. There are no unusual risks associated with participating in this study. Confidentiality. All of your responses will be treated with the utmost confidentiality. Also, your anonymity will be further protected because the written results of these data will not contain individual scores, rather findings will be reported as averages across all subjects. Also, to ensure anonymity, after your data are entered into the computer for analysis, all information that identifies you personally will be discarded. If you have questions. The researcher will answer any questions you may have. This research is being supervised by Dr. Rich Russell of the Psychology Department. You may contact the researcher Deb Downs any time if you have additional questions (home phone: 424-6607; work phone: 242- 5766) . Freedom to decline to participate. You are under no pressure to participate in this study, and you are free to discontinue your participation during the study itself at any time. Your signature below indicates that you have read the description of the study above and freely agree to participate.

Signature

Researcher: Deborah Downs, Richard K. Russell, Ph.D. Graduate Student Advisor 424-6607 APPENDIX 6 DBIOGRAPHXC P H J Æ R XTHIS

73 Please complete the following information. The information you provide will be confidential.

Name:

Social Security Number:

Hometown :

Major(s) or potential majors:

Year in school:

Race/Ethnicity: (please check all that apply) Asian ___ American Indian ___ Black ___ Caucasian ___ Other (please specify) APPENDIX H NANIPOLATION CHECK

75 The purpose of the following questions is to determine your reactions to talking with the other subject who asked you the questions f Subject B) and the other subject who also emswered questions (Subject C). Please answer as ACCURATELY and HONESTLY as possible. Your responses will be confidential. 1. To what degree do you wish you would have talked about something else? Not at all Slightly Moderately Very Extremely

2. How positively did you think Subject B reacted toward you?

Not at all Slightly Moderately Very Extremely

3. How positively did 5ybjflsfe_B react toward Subject C?

Not at all Slightly Moderately Very Extremely

4. How rejecting do you think suhiaet b was toward you?

Not at all Slightly Moderately Very Extremely

5. How rejecting do you think suhlAct B was toward ç ?

Not at all Slightly Moderately Very Extremely

6. How positively do you think (Syhj@g&_2i reacted toward you?

Not at all Slightly Moderately Very Extremely

7. How negatively do you think fffuhjorrt Cl reacted toward you?

Not at all Slightly Moderately Very Extremely