2. ’5 0 ‘l 2 0 0 <6 Tully; l\\l?§i\‘llll\ll L. LIERARY Michigan State University ll This is to certify that the dissertation entitled THE MEASUREMENT OF SHAME AND ITS RELATIONSHIP TO OTHER PERSONALITY TRAITS presented by Shing-Shiong Chang has been accepted towards fulfillment of the requirements for Ph.D. degree in PSYChOlOQ)’ %, M ajor prolessor Date 1 2/2 7/??? yo MS U is an Affirmatiw Action/Equal Opportunity Institution 0- 12771 -_—_._——_A—q—.‘__.-— fl-..“ ___ __._-_ - . _ __ 7, ,7 PLACE IN RETURN BOX to remove this checkout from your record. TO AVOID FINES return on or before date due. DATE DUE DATE DUE DATE DUE c M WKW .2 ~ I MSU Is An Affirmative Action/Equal Opportunity Institution THE MEASUREMENT OF SHAME AND ITS RELATIONSHIP TO OTHER PERSONALITY TRAITS BY Shlng-Shlong Chang A DISSERTATION Submitted to Mlchlgan State Unlverslty In partlal fulflllment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Psychology 1988 ABSTRACT THE MEASUREMENT OF SHAME AND ITS RELATIONSHIP TO OTHER PERSONALITY TRAITS BY Shing-Shlong Chang Despite its Increasingly important position in clinical literature, the construct of shame has not had adequate theoretical development and hence has not produced an empirically verified measurement model. This study was designed around several measurement issues. (1) There is the definition of shame: >What thematic expressions are Indicative of shame? (2) There is the Issue of generality or specificity of shame: ls specific shame causally prior to general shame ("conditioning theory") or Is general shame causally prior to specific shame ("vulnerability theory")? (3) What Is the relationship between shame and self-esteem?‘ (4) What are the relationships between shame and allied constructs such as embarrassment, fear of exposure, public self-consciousness, shyness, and guilt? Four positions are taken in the literature: the "identity," “variants," "shame complex," and "distinct traits" positions. Further- more, gender differences in shame and guilt proneness are assessed as are the relationships of shame and guilt to depression. Subjects were 926 college students (249 males and 677 females). In order to test various measurement models, a variety of outside variables on emotional and social functioning were Included. Path analysis, confirmatory and exploratory factor analysis, partial correlation, and meta-analysis were employed. Extensive use of hierarchical measurement fliOdOIS was made. Shing—Shlong Chang causally equivalent to each other and thus define a general factor. That general factor ls Identical to the general shame factor for the shame theme scales. It was further shown that general shame is causally antecedent to specific shame. Thus vulnerability theory was confirmed. The data showed that shame Is negative self-esteem. With regard to shame and allied constructs, all but the "distinct traits" position were rejected. Meta-analyses showed that women are slightly more prone to shame and guilt than men. Finally, shame Is much more Important than guilt In depression. In fact, gullt Is not related to depression when shame I8 held constant. Copyright by SHING-SHIONG CHANG 1988 ACKNOWLEDGMENTS This study Is like an exploration of myself. it comes out of my own life experiences that are important and meaningful personally. However, this study cannot have been completed without the support, guidance, and assistance of many people i shall forever be grateful to. My heartfelt gratitude goes to the two co-chairs of my dissertation committee: Drs. John E. Hunter and Gershen Kaufman. To you, Jack: You accommodated my fuzzy thinking In the Initial phase of my exploration most patiently. You have spent so many of your precious hours stimulating, guiding, and clarifying my thinking and going over every detail of this project with me. You have touched me deeply In your willingness to respond to my Intellectual needs. The quality of my graduate school life would have suffered greatly had I not had this unique opportunity to work with you. This study cannot have been completed without your Involvement over the past two and a half years. Thank you, Jack. To you, Gersh: Over the past four years, you have Inspired my interest In studying the shame experience through your Es1cnQlogx_91_1na_fie11_§em1naL and through your prolific writings on shame and self-esteem. I am grateful for your willingness to allow me to pursue this project quite Independently even though I fully acknowledge how you have inspired my thinking on the subject matter. l have continued to value your thinking and writing on "shame“ and related subjects. I look forward to reading your forthcoming books. I also would like to extend my thanks to two other members of my V committee: Drs. Bertram Karon and Anne Bogat. Bert: Your contribution to my development as a clinician has been great. Your respect of my sense of autonomy Is particularly appreciated. Anne: I want to thank you for being chairperson of my guidance committee. Your emotional support and guidance I remain forever grateful for. Wes Novak has also been very helpful in my whole process of carrying out this dissertation. His enthusiasm and Interest In my project ls greatly appreciated. There were times when he was a most helpful "listening ear.“ My study was quite literally built on top of the foundation he laid In his fine dissertation. Thank you, Wes. I also would like to acknowledge David Cook's contribution to my study. He generously sent me his unpublished manuscripts and scales. He has collected the most Impressive data set on his Internallzed Shame Scale. David: Your “scale 1“ Is a wonderfully constructed measure of shame. I hope the critiquing of your studies ls constructive. I also want to thank Ronda Hunter for her support and encouragement over the past two and a half years. My apology goes to her for all the Inconveniences and Interruptions l have imposed on her time with Jack. Roger Hailey has been most helpful in his substantial and concrete support of my research project. Marjorie Curtis, Seka Davis, lsabel Cordova helped me greatly with duplicating and collating questionnaires and even stuffing the envelopes. Their help was way beyond the call of duty. Suzy Pavlck and Patti Johnson have been most helpful throughout my graduate school years In the Department. I also would like to thank all the undergraduate students who volunteered for this study. You were my partners In this study. vi Without your participation, the ideas for this study should have been best considered an Intellectual ”arm-chair" exercise. I am also grateful to the instructors who made their students available to me: Drs. Norbert Kerr, Tom Carr, Ann Baumgardner, Marianne McGrath, Ms. Lori Brooks, and Ms. Ann-Marie Scheerbaum. I wish to thank my professors and psychotherapy supervisors: Drs. Albert Rabin, Donald Grummon, Robert Zucker, Norman Abeles, Joseph Reyher, John Hurley, Dozier Thornton, Bill Mueller, Al Anlsklewlcz, John Powell, Al Kirk (all of Michigan State University); Ed Rice, Bob Fink, Bela Chopp, and Amlah Barak (of Oakland University Psychological Clinic). I also wish to thank Bob Bleiski, Harold Drane, Teresa Bernardez, and John Schneider of the Psychiatry Department. My deepest gratitude goes to my wife, Margaret, for giving me support; and my apology goes to her for enduring a lot of heartaches and neglect for the past 10 years of our life together in the United States. Margaret, I dedicate this dissertation to you. I remember vividly and appreciate deeply the enthusiasm my nine year-old son, Joel, exhibited when he lnslsted on filling out the questionnaire of this study even though he never finished It. My apology to you, Joel, for not being able to play with you more often because of my preoccupation with this research project. I also wish to express my feelings of gratitude and love to friends in East Lansing. Thank you all for your friendship and a fond memory. Thank you, East Lansing, a place my wife, my son, and I have loved most dearly. And thank you, MSU. for being my alma mater. le TABLE OF CONTENTS LIST OF TABLES ................................................... LIST OF FIGURES .................................................. INTRODUCTION ..................................................... Statement of the Problem ..................................... I. Shame and Self-Esteem ............................... Il. Definitions of Shame In the Literature .............. II. The Shame Profile ................................... Competence Shame .................................... Relationship Shame .................................. Body Shame .......................................... Character Shame ..................................... Semantic Usage of the Term "Character Shame" versus "General Shame" .......................... Profile of Shame: Affect versus Target .............. lv. Contrasting Theories of Shame Development ............ v. Shame and Traits Associated with Shame .............. Four Positions Regarding Shame and Related Traits VI. Empirical Findings Regarding Shame and Related Traits .......................................... (1) Shame and Guilt ................................ Are Shame and Guilt Different? ................. Shame, Guilt, and Depression ................... Shame, Guilt, and Other Personality Variables (2) Attempts to Differentiate Shame and Other Related Traits ............................. Shame, Embarrassment, and Fear of Exposure ..... Shame and Shyness .............................. Shame and Self—Consciousness ................... MBBSUI‘OMBHI MOOO I 8 .......................................... The Shame Themes .................................... The Shame Profile ................................... 31 31 45 48 55 55 Study Goals .......................................... Study Variables ...................................... Shame Constructs ..................................... Outside Variables .................................... Statistical Methods .................................. OPERATIONAL METHODOLOGY .......................................... Subjects Procedure .................................................... instruments .................................................. RESULTS The The Shame Theme Scales ................................... Shame Profile Scales ................................. Measures of Traits or Affects Associated with Shame ...... Two Self-Esteem Scales ................................... The Internallzed Shame Scales ............................ Scales Measuring Emotional Adjustment Problems ........... Scales Measuring Social Functioning ...................... Themes In the Expression of Shame ........................ A. Six Shame Theme Scales--Development and Testing ......... 1. 2. 3. Conflrmatory Factor Analysis ......................... Exploratory Factor Analysis .......................... Parallelism .......................................... B. Hierarchical Measurement Model .......................... 1. 2. internal Consistency ................................. a. Path Model ........................................ b. Flt of the Path Model ............................. Parallelism .......................................... C. Conclusion .............................................. Shame Profile ............................................ A. Four Shame Profile Scales--Development and Testing ...... 1. 2. 3. Conflrmatory Factor Analysis ......................... Exploratory Factor Analysis .......................... Parallelism .......................................... B. Hierarchical Measurement Model .......................... 1. 2. Internal Consistency ................................. a. Path Model ........................................ b. Flt of the Path Model ............................. Parallelism .......................................... C. Conclusion .............................................. Shame and Self-Esteem .................................... Hierarchical Measurement Model .......................... IV. 1. 2. Internal Consistency ................................. Parallelism .......................................... Conclusion .............................................. Clinical Concepts Related to Shame ....................... Test for “Identity Position" ............................ Test for ”Variants Position“ ............................ Test for ”Shame Complex Position" ....................... lx 55 58 58 59 59 6O 60 60 60 6O 62 62 64 64 65 66 68 68 68 69 69 69 7O 71 71 72 73 74 75 75 76 76 76 76 79 79 79 80 82 83 84 84 86 87 87 89 90 93 The "Distinct Traits Position" .......................... 96 v. Shame and Guilt .......................................... 97 VI. Gender Differences in Shame and Guilt Proneness .......... 99 DISCUSSION ....................................................... 102 The Semantics of Shame ....................................... 103 The Shame Themes ............................................. 104 The Shame Profile ............................................ 105 The Shame Complex ............................................ 106 Cook's Studies (1988a, 1988b) ................................ 108 Gender Differences In Shame and Guilt Proneness .............. 112 Shame and Self-Esteem ........................................ 116 Shame and Guilt ........... . ........... . ...................... 120 Shame, Guilt, and Depression ................................. 122 Implications for Clinical Psychology and Personality Theory .. 124 Clinical Implications of General Shame versus Specific Shame ....................................... 124 Clinical implications for Language Describing Emotions ... 126 Clinical implications for the Treatment of Depression by Focusing on Shame or Guilt ........................... 128 Developmental Theory of Shame ............................ 128 Future Directions ............................................ 129 SUMMARY AND CONCLUSIONS .......................................... 131 The Definition and Measurement of Shame ...................... 132 Shame and Related Traits ..................................... 133 Shame and Self-Esteem ........................................ 134 Shame and Guilt .............................................. 134 Shame, Guilt, and Depression ................................. 134 Gender Differences on Shame and Gulit Proneness .............. 135 Implications for Language Use In Therapy ..................... 136 Shame Therapy: To Focus on General Shame 0r Specific Shame? .. 137 The Development of Shame ..................................... 137 APPENDICES APPENDIX A-1--An Outline of Kaufman's Developmental Theory of Shame ................................................. 138 APPENDIX A-2--inventory Filled Out by Study Subjects .......... 141 APPENDIX A-3--lndlvlduals Scales Used In This Study .......... 152 APPENDIX 8 --Tables .......................................... 162 BIBLIOGRAPHY ..................................................... 175 Table 10. 11. 12. 13. 14. 15. 16. LIST OF TABLES 119123 In Malnslaxt List of Outside Variables and Item Source .................. Means, Standard Deviations and Reliablllties of Six Shame Theme Scales and General Shame Scale ............. intercorrelatlons among Six Shame Theme Scales ............. Error Analysis for the Model In Figure 6 ................... Similarity Coefficients of Six Shame Themes ................ Means, Standard Deviations and Reliablllties of Shame Profile Scales ....................................... lntercorreiatlons of Shame Profile Factors ................. Error Analysis for the Model In Figure 7 ................... Correlations between Shame Profile and Outside Variables with General Shame Partlaled Out ...... . ......................... Correlations between Shame and Self-Esteem ................. Error Analysis for the Model In Figure 8 ................... Correlations of Shame and Self—Esteem with Outside Variables .................................................. Similarity Coefficients between Shame and Self-Esteem Scales ..................................................... lntercorreiatlons between Shame, Embarrassment, Fear of Exposure, Public Self-Consciousness, Shyness, and Guilt Error Analysis for the Model In Figure 9 ................... Partial Correlations between Embarrassment, Fear of Exposure, Public Self-Consciousness, Shyness, and Guilt and Outside Variables with General Shame Partlaled Out ................. XI Page 67 70 71 73 74 77 77 80 82 84 85 86 87 89 91 92 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. B1. 82. B3. B4. B5. 86. 87. Error Analysis for the Model In Figure 10 ................... Partial Correlations between Shame, Fear of Exposure, Embarrassment, Shyness, and Guilt with Outside Variables (with the Hypothetical Higher Order Factor Partlaled Out) Correlations of Shame and Guilt with Emotional Adjustment Problems; In Isolation and with Each Other Partlaled Out Means and Standard Deviations of College Males and Females on Shame, Self-Esteem, and Guilt Scales ....................... Correlations between Gender and Shame. Self-Esteem, and Guilt .... ................. . ............................ Correlations between Four Factors In Basic Samples .......... Correlations between Four Factors In Cross-Validation Samples .................................................... A Meta-Analysis of Correlations between Gender and Shame Proneness In Five Studies .................................. A Meta-Analysis of Correlations between Gender and Guilt Proneness In Four Studies .................................. Correlations of Internallzed Shame Scale (ISS) with Rosenberg Self-Esteem Scale (with Correction for Attenuation) ........ Iabies_1n_Annensix_B Summary of the Results of the Conflrmatory Factor Analysis for Six A Priorl Shame Theme Clusters: Correlations between Items and Factors (Decimals Omitted) ....................... Summary Data for Outside Variables and Other Key Scales Correlations of Six Shame Theme Cluster Items with Outside Variables (Decimals Omitted) ............................... Correlations of Six Shame Theme Scales with Outside Variables (Decimals Omitted) ............................... Summary of Conflrmatory Factor Analysis for Shame Profile Clusters: Correlations between Items and Factors (Decimals Omitted) ................................................... Summary of Exploratory Factor Analysis on Shame Profile Items: Factor Loadings of Shame Profile Items (Decimals Omitted) ................................................... Correlations of Shame Profile Items with Outside Variables xII 94 95 98 100 100 118 162 163 164 166 167 168 B8. 89. 810. 811. (Decimals Omitted) ......................................... Correlations between Shame Profile and Outside Variables Error Analysis for a Hypothetical Hierarchical Model Subsuming Shame, Embarrassment, Fear of Exposure, Public Self-Consciousness, Shyness, and Guilt ..................... Summary of Results of Higher Order Conflrmatory Factor Analysis on Hoblltzelle (1987) Study Data (Decimals Omitted) ................................................... Summary of Results of Higher Order Conflrmatory Factor Analysis on Harder & Lewis (1987) Study Data (Decimals Omitted) ................................ . .................. xlil LIST OF FIGURES Figure 1. Contrasting models of shame development .................... 1a. A Path Model of the Conditioning Theory of Shame ............ 1b. A Path Model of the Vulnerability Theory of Shame .......... 2. The hierarchical model for the shame theme scales showing the path model for the test for Internal consistency ........... 2a. The confirmatory factor model for the first of the shame theme constructs: disappointment with self ................. 2b. The confirmatory factor model for the shame theme cluster with general shame as the general factor or higher order construct .................................................. 2c. The combined hierarchical measurement model for all the Items of the shame theme scales .................................. 3. The hierarchical model for the shame theme scales showing the path model for the test for external consistency or parallelism ................................................ 3a. The confirmatory factor model for the first of the shame theme constructs: disappointment with self ................. 3b. The confirmatory factor model for the shame theme cluster with general shame as the general factor or higher order construct .................................................. 3c. The combined hierarchical measurement model for all the Items of the shame theme scales .................................. 4. The path model representing the “variants position" which argues that the five affects shown are all variations of shame ...................................................... 5. The path model for the "shame complex position“ which argues that (1) the traits shown are variants of shame and (2) that the general factor for the shame theme scales Is also only a variant of shame ........................................... 6. The hierarchical measurement model of general shame: A model for the test for Internal consistency ...................... 7. The hierarchical measurement model of general shame and shame profile .................................................... Page 14 14 14 36 36 36 36 43 43 43 50 51 71 79 10. 11. 12. The hierarchical measurement model for shame and self- esteem ..................................................... A hierarchical model with general shame as higher order factor: the test for Internal consistency .................. A hypothetical hierarchical model subsuming shame, embarrassment, fear of exposure, shyness, and guilt: the test for Internal consistency ................................... Stages In psychological magnification ...................... A hypothetical hierarchical model subsuming shame-related traits, Including shame .................................... XV 85 9O 93 140 INTRODUCTION The focus of this research Is shame-—how to define It and how to measure It. Shame has been shown to be a critical factor In clinical work on personality and development. This study will present empirical findings relating shame to Important measures of emotional adjustment such as anxiety, depression, loneliness, life satisfaction, and so forth. The study will also present empirical findings relating shame to Important measures of social functioning such as conformity, need for approval, egocentrlsm, exhibitionism, dominance, self-revelation, and so on. There has been a great deal of research relating emotional adjustment to self-esteem. An examination of shame and self—esteem suggests that these are actually the same construct: Shame Is the name for the negative affect at the opposite end of the same continuum whose positive pole is named self-esteem. This study will examine the relationship between shame and self-esteem measured separately. There have been great confusions as to the definition of shame. This study will consider various definitions within the context of empirical measurement. At a primary level, there Is the question about how to define shame. First, there Is the most narrow semantic question: What words or phrases are Indicative of shame? Examination of the shame literature (e.g., Alexander, 1938; Piers & Singer, 1953; Lewis, 1971; 2 Buss, 1980; Wurmser, 1981; Kaufman, 1985; Miller, 1985; Coen, 1986) and verbal reports that occur In therapy led to the development of six shame themes: disappointment with oneself, feelings of Inferiorlty, feelings of defectiveness, feelings of worthlessness, feelings of unlmportance, and feelings of falling short of one's own standards or Ideals. Second, there Is the Issue of generality or specificity of shame. Some authors have assumed that shame Is first developed In specific circumstances In regard to particular aspects of the self (Kaufman, In press). Shame then generalizes from the particular aspects of the self to a feeling of general shame about the self as a whole. In contrast, Novak (1986) noted that his data suggest the opposite: General shame ls causally prior to specific shame. His data suggest that some people are much more vulnerable to shame than others. Those who are vulnerable tend to develop shame about all or most aspects of the self. Novak noted, however, that his study design did not permit a good test of this hypothesis. The present study will test this hypothesis directly. Measures of specific shame will be developed using Kaufman's "shame profile”: shame about competence, about relationship, about character, or about the body. The questions are these: (1) How high are the correlations between different specific shame affects? (2) how high are the correlations between different specific shame affects and general shame? and (3) do the causal relations run from general shame to specific shame or from specific shame to general shame? Shame will be related to allied constructs such as embarrassment, fear of exposure. self-consciousness, shyness, and guilt. Some authors have argued that these affects define shame itself. Some authors have 3 argued that these affects are variants of shame. Some authors have argued that these affects are part of a generalized shame complex. This study will test all three of these positions. This study will also examine gender differences In shame and guilt proneness. In addition, the relationships of shame and guilt to depression and other emotional adjustment problems will be Investigated. For a review of the shame literature, the reader Is referred to Morrison (1983), erman (1984), Miller (1985). and Chang (1988). The reader Is also referred to the major sources: Piers and Singer (1953), Lynd (1958), Tomklns (1963, 1987), Lewis (1971, 1979, 1981, 1987a, 1987b), Schneider (1977, 1987), Wurmser (1981, 1987), Morrison (1983, 1987), Kaufman (1985, in press), and Nathanson (1987). Only the select shame literature pertinent to the Issues mentioned above will be reviewed here. At the outset, shame Is defined as the negative attitude and feelings one has about oneself. In shame the self becomes the focus of the negative evaluation (e.g., Lynd, 1958; Lewis, 1971, 1979, 1981). Shame feelings can be triggered by specific aspects of one's behavior or specific events; however, It Is the self that is called Into question and negatively evaluated. The feelings In shame are primarily those of self-disappointment and self-disgust (Buss, 1980). One judges oneself to be inferior, Inadequate, defective, ugly or repulsive, and feels like a failure, or feels unimportant, uniovable, or worthless. One can put oneself down for practically anything one dislikes about oneself, for a poor performance, for being ignored, rejected, or ridiculed. There Is Inevitably a drop In self-esteem when one feels ashamed (Buss, 1980; 4 Morrison, 1983). Since one focus of this research Is the relationship between the constructs of shame and self-esteem, a brief review of the self-esteem literature Is In order. The area of self-esteem research represents a wide range of theoretical orientations. Self-esteem has been operationaiized In a variety of ways: as a score on a questionnaire, as a coded behavior, as an intervening variable, as a "self-Ideal discrepancy," as an evaluation of task performance, and as a variety of other functions and tendencies. A comprehensive review of the self-esteem literature ls beyond the scope of this study. 1 will only review a number of writers, specifically, Jacobson, Rosenberg, Coopersmlth, Buss, and Branden. in her book Jacobson (1964) states: Broadly defined, the level of self esteem Is expressive of the harmony or discrepancy betweem the self representations and the wishful concept of the self, which Is by no means always Identical with the unconscious and conscious ego Ideal. Hence, disturbances of self esteem may originate from many sources and represent a very complex pathology: on the one hand, a pathology of the ego ideal or of the achievement standards and goals of the ego and, hence, of the self-critical ego and superego functions, and, on the other hand, a pathology of the ego functions and of the self representations. (p. 131) Jacobson's "self representations“ correspond to what has now been commonly referred to as “actual self Image,“ and her "wishful concept of the self" corresponds similarly to "Ideal self Image.“ Thus her definition of self-esteem Is equivalent to a currently popular one: Self-esteem Is the degree to which actual self Image matches Ideal self Image (Campbell, 1984). Rosenberg (1965) has constructed one of the best known self-esteem scales. TO I‘III‘I‘I self-esteem SIQI‘IIfIBS a DOSItIVB or negative attitude toward oneself. He states (1979): 5 When we characterize a person as having high self-esteem, we are not referring to feelings of superiority, In the sense of arrogance, conceit, contempt for others, overweenlng pride; we mean, rather, that he has self-respect, considers himself a person of worth....The person with high self-esteem has pnijgtimg, not nunLis; he does not necessarily consider himself better than most others but neither does he consider himself worse. The term “low self-esteem" does not suffer from this dual connotation. it means that the Individual lacks respect for himself, considers himself unworthy, Inadequate, or otherwise seriously deficient as a person. (p. 54) Coopersmlth (1967) sees self-esteem as a "personal judgment of worthiness that Is expressed In the attitudes the Individual holds toward himself.“ Cheek and Buss (1981) constructed a self-esteem scale that again taps opinions of self-worth. Branden (1983) defines self-esteem as follows: Self-esteem Is a concept pertaining to a Lunnamental sense of efficacy and a fundamental sense of worth, to competence and worthiness In prjnglnie....ln sum, self-esteem Is an evaluation of my mind, my consciousness, and, in a profound sense, my person. Lt Is not an evaluation of particular successes or failures, nor Is It an evaluation of particular knowledge or skills....leIng up to my own standards Is, as we shall see when we consider the factors that enhance or diminish self—esteem, an essential condition of high self-esteem. (p. 12) In short, a sense of competence and a sense of personal worth are key to his definition of self-esteem. I. Shama_and_5e11:fistaem The construct of shame has recently been linked with depression (Lewis, 1971, 1979, 1981, 1987a, 1987b; Smith, 1972; Crouppen, 1976; Morrison, 1983; Mollon, 1984; Kaufman, 1985, In press; Harder & Lewis, 1987; Hobiltzelie, 1987; Cook, 1988a, 1988b), narcissistic disturbances (Kohut, 1971, 1972, 1977, 1984; Broucek, 1982; Morrison, 1983, 1984, 1987; Mollon, 1984; Mollon & Parry, 1984; Kaufman, 1985, In press; Klnston, 1987; Lewis, 1987b) and many other emotional problems. For a much longer period of time the construct of self-esteem has held a central position In much of the psychoanalytic literature on depression (Freud, 1917; Rado, 1928; Fenichel, 1945; Blbrlng, 1953; Jacobson, 1953, . 1964, 1971, 1975). In narcissistic disturbances, problems of self- esteem have long been Implicated (e.g., Reich, 1960; Kohut, 1971, 1977; DSM—lll, 1980; DSM-lIl-R, 1987). One cannot help but wonder what kind of relationship these two constructs, namely, shame and self-esteem, have with each other. in the literature both constructs have been studied extensively, but their relationship has not yet been Investigated. Some writers on self-esteem (e.g., Branden, 1969, 1983, 1987; Wells & Marwell, 1976; Campbell, 1984; Jackson, 1984) have not even made one Indexed reference to shame In their books as If there Is no relationship whatsoever between the two constructs. Keep In mind Branden's definition of self-esteem quoted above (on page 5) and, then, consider the following statement: Shame Is a concept pertaining to a fundamental sense of lnefflcacy and a Landamantai sense of lack of personal worth, to incompetence 7 and unworthlness in nLInainla. in sum, shame is a negative evaluation of my mind, my consciousness, and, in a profound sense, my person. it Is not an evaluation of particular successes or failures, nor Is It an evaluation of particular knowledge or skills. Not living up to my own standards is an essential condition of shame. With just a few key words changed from Branden's definition of self- esteem, we have a definition of shame. In her book Jacobson (1964) states “broadly defined, the level of self esteem ls expressive of the harmony or discrepancy between the self representations and the wishful concept of the self." As Campbell (1984) has pointed out, Jacobson's definition of self-esteem Is equivalent to a currently popular one: Self-esteem Is the degree to which actual self image matches Ideal self image. According to Piers (Piers & Singer, 1953), shame Is manifest when there Is a discrepancy between one's ideals and one's actual behavior. in other words, shame Is the degree to which actual self Image or behavior mismajgnaa Ideal self image. Here we can see the striking similarity (or Identity?) between the two constructs. Let us look at Rosenberg's description of people with low self— esteem: They lack respect for themselves, consider themselves unworthy, Inadequate, or otherwise seriously deficient as persons (1979). Isn't the above description Rosenberg gives to people with low self-esteem just as applicable to people with a strong sense of shame? In other words, are these people not prone to shame? Buss (1980) states that there Is Inevitably a sharp drop in self-esteem In shame reactions. Morrison (1983) speaks of a fail in self-esteem as the hallmark of shame. They both clearly see a strong correlation between shame and self-esteem. Commenting on an empirical study of guilt and conscience 8 In the major depressive disorders that suggests that ”negative self- esteem" rather than guilt may form one of the cornerstones for depressed patients of all types (Prosen, Clark, Harrow, & Fawcett, 1983), Lewis (1987a) remarks that “shame Is the affective-cognitive state of low self—esteem." This study will examine the relationship between the constructs of shame and self-esteem. Are they basically the same construct? ls shame exactly the same as negative self-esteem at the trait level? Or are they two different constructs even though they are highly correlated? II. QQIIDJ1iQflS_QI_§him2_lfl_ln§_LllflLilnL§ An examination of the shame literature suggests that shame feelings about oneself have been talked about in many different ways. Six ways or themes can be distilled from the literature as well as from our everyday language and from clients' verbal reports in psychotherapy sessions. They are: disappointment with oneself (Lewis, 1971; Buss, 1980). feelings of Inferiorlty (Adler, 1929, 1931; Alexander, 1938; Miller, 1985; Novak, 1986), feelings of defectiveness (Wurmser, 1981; Morrison, 1983; Miller, 1985; Coen, 1986; Cook, 1985, 1988a, 1988b). feelings of worthlessness (Kaufman, 1985). feelings of unlmportance which are often expressed metaphorically as "feeling small” (Miller, 1985; Cook, 1985, 1988a, 1988b). and feelings of falling short of one's own standards or Ideals (Piers & Singer, 1953; Lynd, 1958; Morrison, 1983; Mollon, 1984). The questions are these: How are these shame themes correlated with each other? Do they measure the same underlying trait or do they 9 measure different traits? Will these shame themes fit a hierarchical measurement model with general shame as the higher order factor? The topic of hierarchical measurment model will be elaborated In the section on “Measurement Models.“ Ill. Ina_§nama_£L9111a In the past few years some writers on shame (e.g., Joffe, 1984; Kaufman, in press) have proposed "shame profiles." The profile model assumes that shame first becomes differentially associated with parts of the self or specific situations. Kaufman's theory Indicates that shame can be more meaningfully understood as a profile of activators and sources rather than a unitary construct. Joffe's profile Includes: competence shame, appearance shame, rejection shame, character shame, criticism shame, and homosexual shame. Kaufman's profile Includes: competence shame. body shame, relationship shame, and character shame. He postulates that the first stage of magnification (increased duration and Intensity) of shame affect includes four primary shame dimensions: affect shame, need shame, drive shame, and purpose shame. The second stage magnification fuses these Into: competence shame, body shame, and relationship shame. The final stage of magnification Is "character" shame. A diagram of these stages Is Included In Appendix A-1. This study focuses only on competence shame. body shame, relationship shame, and character shame. Can we measure such a profile reliably? What about the validity of such a profile? Can we find different kinds of shame at the trait level? in his dissertation Novak (1986) pointed out that there was little support for the formulation of a shame profile In 10 existing shame measures "as the profile of shame postulated (according to shame content or object of shame) Is not found." But he added: It Is important to note that the discrete clusters of shame In Kaufman's schema may exist, but to be validated requires the expansion of the domain of present shame measurement to Include the postulated states. Current shame measurement does not Include the range of shame states formulated In Kaufman's profile model. (p. 59) Novak's preliminary finding deserves extension. Therefore, items were written for this study to measure the separate components of Kaufman's shame profile and to test the validity of such a profile. A definition of the components of Kaufman's shame profile is In order. Competence.§hame People with competence shame experience themselves as Incompetent, tend to suffer from feelings of Inadequacy and tend to doubt their abilities. There Is a keen sense of being without essential talents and abilities and without skills one can be proud of. In its most Intense form, the self Is experienced as a complete failure, as being incapable of amounting to anything. The Individuals see others as being endowed with greater skills and abilities, compare themselves constantly with others, and conclude that others are more capable and competent. Such Individuals tend to be less adventurous, less willing to try new things for fear of falling. Relationsnin_sname People with relationship shame experience themselves as uniovable, and tend to doubt that anyone will ever find them acceptable, or worthy of being loved or being paid attention to. There Is a keen sense of not belonging, of being unwanted, unwelcome, and left out. For the most part relationship shame does not lead to a focusing on a specific flaw; 11 Instead, what ls reported Is a global and diffuse feeling that one Is loathsome, undesirable, unacceptable, and unlovabie. The Individuals with relationship shame believe that other people do not care for them and do not hold them In high regard. They tend to feel Insignificant to others around them. They tend to feel discounted, neglected, rejected, or held In contempt. Bodx.§hame People with body shame experience themselves as physically unpresentable and lacking In physical appeal or attractiveness. In Its most intense form, the self Is experienced as ugly or funny looking, offensive to look at, or freakish in appearance. There Is a body image problem. Common themes Include being too fat, too thin, too tall, too short, or having bodily defects, and so forth. Their attention is often drawn to their appearance and they feel that other people judge them on the basis of their physical appearance (Joffe, 1984). QhflLfl$1£L_§himfi People with character shame experience themselves as defective In their personality, as If something is basically wrong with them. They see themselves as hopelessly flawed. They experience disappointment and disgust for any character flaws they discern in themselves, be It dependent feelings, emotionallty, Irrationallty, Immaturity, unreliability or undependablilty, chronic dishonesty or phonlness, or feeling weak or spineless or being without Inner substance. Other people are seen as stronger, more mature and more Independent and In possession of true character and inner fiber (Joffe, 1984). 12 .n.. ... . .. . u .. . . ..H. . ~ -.; . ..u. It should be pointed out that there are two ways Kaufman uses the term "character shame.“ in some contexts, he uses the word "character shame" in the same way Joffe does: that Is, as shame feelings about one's "character"--be It one's Immaturity, dependency, emotionallty, Irrationallty, unreliability, undependablilty, and so forth. The focus is on shame about an aspect of oneself, not about the self as a whole. When “character shame“ Is used In this sense, It is a part of the shame profile. But there are other contexts In which Kaufman uses "character shame“ to refer to shame feelings about oneself as a whole rather than shame about some aspect of personality. In this study the term "general shame“ Is used to refer to shame feelings about oneself as a whole. Some authors believe that all shame Is general shame. Lynd (1958) and Lewis (1971) both used the term "shame” to refer to negative feelings about oneself as a whole. They both see this "whole—self Involvement" as one of the distinguishing characteristics of shame. In this study "character shame" will be reserved for shame feelings about the “character" aspect of oneself, not about oneself as a whole. “General shame" refers to shame feelings about oneself as a whole. WWW It should be obvious that the common denominator In the definitions of different components of the shame profile Is the shame affect. In other words, at the level of affect, they are not differentiated. The affect Is the same In each case; only the target Is different. The stimulus, situation, or event that triggers the shame reaction may be 13 different at different times. The result Is Still the same affective BXDBFIBHCB Of shame. IV- QQfl1Lflfi1lflfl_IflflQLlfi§_91_§hflmfi_flfilfliflnmfifll Two contrasting theories of shame development are proposed In the literature: the “conditioning theory" and the ”vulnerability theory" (Novak, 1986). It should be pointed out that some aspect of Kaufman's theory of shame (1985, In press) Is primarily a "conditioning theory;" however, Kaufman does Include the “vulnerability factor" as one of the Innate factors to account for Individual differences. An outline of his developmental theory of shame Is Included In Appendix A-1. Kaufman assumes that shame is first developed In specific circumstances In regard to particular aspects of the self. Shame then generalizes from the particular aspects of the self to a feeling of shame about the self as a whole, that Is, general shame. He uses such a model to show how shame spreads throughout the self and leads eventually to a shame—based identity (personal communication, October, 1987). In contrast, Novak (1986) noted that his data suggest the opposite: General shame is causally prior to specific shame. His data suggest that some people are much more vulnerable to shame than others. Those who are vulnerable tend to develop shame about all or most aspects of the self. Thus there are two competing theories as to the relationship between shame toward the self as a whole (here called “shame" or "general shame“) and shame toward specific aspects of the self (here referred to as "specific shame“): Kaufman's theory which asserts that general shame IS derived from SDBCITIC shame BDO Novak's theory WhICh 14 asserts that general shame Is differentiated Into specific shame. These two theories are contrasted as path models In Figure 1. Body Shame Relationship Shame ‘*~* _ General Shame: Shame L””””””’,,,”' about Self as a Whole Competence Shame Character Shame Eiguna_1a. A path model for the conditioning theory of shame. Body Shame Vulnerability to Relationship Shame Shame (General Shame) Competence Shame—I Character Shame Elguna_1n. A path model for the vulnerability theory of shame. Eiguna_1. Contrasting models of shame development. According to Novak (1986). Kaufman's theory Is at heart a "conditioning theory." Kaufman believes that shame develops from shaming experiences or what he calls “breaking of interpersonal bridge." A child (or an adult) Is made to feel ashamed about some aspect of the self by a shaming parent or peer or significant other. That shame experience then conditions the shame effect to that aspect of the self. Thus shame about each aspect of the self could develop separately and 15 independently. This Is represented In the path diagram of Figure 1a by the fact that the components of the shame profile are entered Into the causal model as ”exogenous" variables. The model In Figure 1a shows the specific shame profile components as uncorrelated. This corresponds to the fact that Kaufman presents no theory that would suggest that the aspects be correlated. But In the case of a shaming parent It Is likely that the parent would use shame as a control mechanism for many behaviors. Thus while it Is possible for the shame aspects to be uncorrelated as shown in Figure 1a, it Is also possible that there Is a correlation In the environmental events which produce the specific shame components. When the child thinks about the self and remembers shame experiences or relives shame affect, then shame Is conditioned to the self as a whole. Thus the greater the extent of any given shame profile component, the greater the generalization of shame from the specific aspect of the self to the self as a whole. This Is represented In Figure 1a by the causal arrows from each specific shame profile component to general shame. Novak's “vulnerability theory" assumes that some people are more vulnerable to shame than others. A person with high vulnerability exposed to a shaming experience will develop shame about that aspect of the self. A person with an extremely low vulnerability will not respond to the shaming experience by developing a permanent shame response toward any aspect of the self. Since shaming experiences are endemic In life, a person with high vulnerability will have a high probability of developing specific shame toward all aspects of the self, one by one, as experiences accumulate. Thus it is whether a person Is high or low In 16 shame vulnerability that Is the primary causal determinant of individual differences in specific shame profile components. According to this theory, general shame measures shame vulnerability. Thus, the vulnerability theory predicts that general shame Is a primary causal determinant of each specific shame profile component. This Is represented In the path model of Figure 1b by the fact that general shame Is the "exogenous" variable. The role of general shame in determining Individual differences on the specific shame profile components is shown by the arrows that run from general shame to the specific shame profile components. Both models In Figure I predict high correlations between general shame and each specific shame profile component variable. However, the two models make drastically different predictions about the correlations between the shame profile component variables. Conditioning theory would be satisfied even if the shame profile component variables were uncorrelated with each other. Small or even moderate correlations between the shame profile components might be explained by correlation In the frequency of shaming experiences In relation to different aspects of the self. By contrast, vulnerability theory predicts very high correlations between the shame profile components. If every child was exposed to the same number of shame experiences for each shame profile component, the vulnerability theory would predict a perfect correlation between each shame profile component and general shame. This In turn would predict perfect correlations between the shame profile components. However, Idiosyncratic experiences cause the correlation between general shame and each shame DI’OTI IO component to D8 IBSS than 1.00 and reduces 17 the correlations between shame profile components correspondingly. In fact, If the correlations between general shame and two shame profile components are rscl and rsc2' then the path model of Figure 1b shows that vulnerability theory predicts the correlation between shame components to be the product of the respective correlation between each shame component and general shame, that Is, rc1c2 ' rsc1 rsc2 For example, If the correlation between general shame and competence shame were .90 and the correlation between general shame and relationship shame were .80, then vulnerability theory would predict that the correlation between competence shame and relationship shame would be (.90)(.80) - .72. Novak (1986) noted that his study design did not permit a good test of the two contrasting theories. The present study tests these two theories directly. V. snama_and_ILaIts_Assggia1ed_wItn_snama Some writers do not differentiate shame from embarrassment, self- consciousness, fear of exposure, or shyness (Lewis, 1971; Izard, 1971; Wurmser, 1981; Morrison, 1983; Nathanson, 1987). In fact, some authors (Lynd, 1958; Schneider, 1977; Kaufman, 1985) have made embarrassment the cornerstone of their shame construct. That Is to say, they stress Intense experience of exposure and heightened self-consciousness In talking about shame affect. Furthermore, both Tomklns (1963, 1987) and Kaufman (1985) have hypothesized that experiences such as shyness In the presence of strangers, self-consciousness when talking before a large 18 group, embarrassment when seeing oneself as socially Inappropriate, or even guilt for a transgression or lmmorallty, while phenomenologically felt as distinct experiences. are. at the level of affect, variants of the shame affect. Other writers (e.g., Buss, 1980; Miller, 1985) have attempted to differentiate embarrassment from shame. Buss (1980) has differentiated four varieties of what he labels as “social anxiety“ (l.e., shame, embarrassment, audience anxiety, and shyness). All these four varieties have something In common: acute public self-awareness, In which the Individuals are aware of themselves as social objects. Shame Is viewed as a more severe and enduring emotional state than embarrassment. Here, Buss obviously does not match on the level of intensity. In other words, ls mild shame still a more severe and enduring emotional state than Intense embarrassment? Buss further states that there Is an enduring loss of self-esteem In shame, but only a temporary loss In embarrassment. Blushing, a silly smile, and nervous giggling or laughing are part of embarrassment, but not of shame. Obviously, there Is no unequivocal answer to the question whether blushing Is part of embarrassment or part of shame. Many authors certainly would disagree with Buss (e.g., Lynd, 1958; Tomklns, 1963; izard, 1977; Kaufman, 1985) as they do not differentiate shame from embarrassment. To Buss, the embarrassed persons feel foolish or silly; the ashamed persons feel disgusted or disappointed with themselves. Severe shame resembles depression. Audience anxiety ls revealed by the presence of tension, disorganization, worry, and a feeling of panic. Shyness Is Inferred from an Inhibition of expected social behavior together with feelings of 19 tension, discomfort, or awkwardness. The shy person may be concerned about being seen as III-mannered, clumsy, Intrusive, too loud, or appearing foolish, or saying the wrong thing. Miller (1985) attempts to differentiate between shame, embarrassment, humiliation, and self-consciousness. Shame carries the conviction that one Is small, Inferior, or defective. There Is a strong sense of displeasure about one's status and a desire to be different: to be smarter, stronger, more competent, or more beautiful. In her view, feeling Inferior and diminished is the core experience of the shame state. She regards the affect of embarrassment as a response to a sudden feeling that an aspect of the self has been unwillingly opened up to view by others. The core experience of embarrassment Is the excited or aroused state and the sense of being off-kllter or having lost one's composure. Humiliation, on the other hand, reflects the self being forced Into a debased position. It differs from shame In that shame Is a belief about the self (e.g., "I am worthless") whereas humiliation refers to the temporary change in status effected by someone else and not to an Identified, enduring quality of the self. Self-consciousness refers to a sense of "feeling constantly aware of the self-In—actlon." To her, self-consciousness differs from shame In that It carries less or little implication of a sense of defect or inferiority one Identifies In oneself. lzard's thinking about shame and other affects associated with shame In his differential emotions theory has changed over the past decade. It has changed from a consideration of shame as a fundamental emotion that Includes embarrassment, shyness, and guilt (1971), to 2O shame/shyness as different from gullt (1972), to all three as separate and fundamental emotions (Izard & Hyson, 1986). EouL_Eos1tIons_BenaLdIns_Sname_ann_Belated_ILaits A closer examination of the literature regarding the relationships between shame and the related traits--that ls, embarrassment, fear of exposure. self-consciousness. shyness, and gullt--lndicates that there have been four positions taken on this Issue In the literature. These positions will be called the "Identity,“ "variants," ”shame complex," and ”distinct traits” positions. First, some have argued that one or more of the related traits Is a part of the shame process Itself. These authors would expect the related traits to be perfectly correlated with shame and hence with each other. This will be called the ”Identity position." Helen Block Lewis Is one of these authors with the exception that she unequivocally differentiates shame from gullt. Wurmser (1981), Morrison (1983), Harder and Lewis (1987). Hobiltzelle (1987). and Nathanson (1987) also fit the same category. Again, they all differentiate shame from gullt a la Piers and Singer (1953). Second, some (e.g., Tomklns, 1963, 1987; Kaufman, 1985) have argued that the related traits are variants of the primary shame affect. These authors would permit the related traits to have specific factors and hence not be perfectly correlated with either shame or each other. But since the specific factors are trivial in nature, they should not correlate with each other or with the outside variables. Thus these authors would predict that the related traits would satisfy a hierarchical measurement model with shame as the general factor. This 21 will be called the "variants position." Third, some authors would argue that there Is one general shame complex that subsumes the related traits. But they would argue that the general shame scale might also be subsumed under that shame complex factor rather than be Identical to that factor. These authors would predict that a hierarchical measurement model would fit the variable cluster formed by adding shame to the list of related traits. This will be called the "shame complex position." it should be noted that Buss (1980) used the term "social anxiety" Instead of "shame complex" for the higher order factor. But his ideas fit this third position. Fourth, there are those who would simply argue that traits such as shame, shyness, and guilt measure causally different variables. Some (e.g., Novak, 1986) would also argue that embarrassment and shame are distinct as well. These authors would predict that a hierarchical measurement model would not fIt the data. This will be called the “distinct traits position." This study will test the first three positions to see If any of them are supported by the empirical data In this study. The "distinct traits position“ will not be tested as such. Rather the "distinct traits position" Is the conclusion implied by the failure of the other positions. It should be pointed out that the plgeonholing of shame authors to different positions cannot be done perfectly. Authors are not always consistent in the position that they adopt from one discussion to the I'IOXt. 22 VI. anal:|caL_EIndIngs_BagaLdina_Shims—nnfi_flnlfllfifi—IL1113 (I) Shamn_and_GuLLt There are two Issues relating shame and guilt. First, some authors believe that gullt Is a form of shame. The corresponding question Is: Can gullt be differentiated from shame? Second, there has been an argument as to whether It is guilt or shame that Is a causal factor In the causation of depression. If gullt were a form of shame, the second question would be irrelevant. Thus the discussion of depression will assume that shame and guilt can be differentiated. ALe_5hame_and_£ulL1_Qi11eLnn1? The majority of theorists have differentiated shame from guilt conceptually (Alexander, 1938; Piers a Singer, 1953; Lynd, 1958; Levin, 1967; Lewis, 1971, 1979, 1981, 1987a, 1987b; Buss, 1980; Wurmser, 1981; Morrison, 1983; Miller, 1985; Fossum & Mason, 1986). Several doctoral dissertations and studies have explored the relationship between shame and guilt psychometrlcally (e.g., Perlman, 1953; Binder, 1970; Smith, 1972; Crouppen, 1976; Korpl, 1977; Negrl, 1978; erman, 1984; Harder & Lewis, 1987; Hobiltzelie, 1987; Cook, 1988a). Perlman reported a .61 correlation between his shame and guilt scales (N - 64). Negri used the revised Perlman Anxiety Attitude Survey and reported a .47 correlation between the two scales (N - 128). Binder and Smith did not report their correlations. Wood, Pillsuk, and Uren (1973) used the Beall Shame-Guilt Test (Beall, 1972) and found a .59 correlation between shame and guilt orientation (N - 78); Crouppen used the same test but found much higher correlations (.76 to .81, N - 40 In each of two samples). Revising the Beall test, Korpl (1977) reported a .41 correlation between the shame 23 and guilt scales (N - 173). erman used the Korpl Shame-Guilt Test (Korpl, 1977) and reported a .13 correlation between the two scales (N - 102). Summing scores on three shame-related scales (l.e., the Korpl Shame Scale, the Cattell Susceptibility to Embarrassment Scale, and the Revised Stanford Shyness Survey) and summing scores on four gullt measures (three Mosher's guilt subcales and the Korpl Guilt Scale), he reported that the composite ”shame" scores correlate .32 with the composite gullt scores. He also reported a .23 correlation between the Korpl Shame Scale and the composite gullt scores. Harder and Lewis (1987) constructed the Personal Feelings Questionnaire (PFO) and obtained a .47 correlation between the shame and guilt scales (N - 120). Hobiltzelle (1987) used the Adapted Shame/Guilt Scale (ASGS) and found a .23 correlation between the two scales (N - 124). Cook (1988a) found a .61 correlation between his internalized Shame Scale (ISS) and a gullt- scale (N - 103). A simple matazanalxsia averaging weighted correlations between shame and guilt across these studies yields a .48 correlation. According to theorists who treat guilt and shame as the same thing, the correlation between them should be 1.00. An average correlation of .48 Is a substantial positive correlation, but It Is far less than 1.00. Can we then conclude that guilt and shame are different? There are a number of potential problems with past studies which might have caused the correlation to be understated. It would be premature to consider the correlation as given until those problems are considered. First, it should be noted that the correlations In these studies were not corrected for attenuation. in other words, If they had been corrected for attenuation, the average correlation would have been 24 higher than .48. However, a check of the studies shows that In all cases, the scales or combinations of scales being correlated had many Items. By comparison with scales with known reliabilities, It Is unlikely that the mean reliabilities would be less than .80. Consider a more extreme position. Suppose that the mean reliabilities were as low as .75. Then the corrected correlation would be .43 / ](.75) J(.75) - .48 / .75 - .64 While .64 is a substantial positive correlation, It Is still far below 1.00. Thus It seems unlikely that the “low" correlation could be explained on the basis of sampling error. However, there are also potential problems with construct validity In some of the studies (Perlman, 1953; Wood et ai., 1973; Crouppen, 1976; Negrl, 1978; erman, 1984; Hobiltzelie, 1987; Harder & Lewis, 1987; Cook, 1988a, 1988b). One major problem with certain shame and guilt scales (e.g., Perlman, 1953; Binder, 1970; Beall, 1972; Korpl, 1977) is that the Items In these scales do not specify the inner affective experience of the respondent In the situations. Respondents were asked to Indicate how anxious, upset, or uncomfortable they would feel in certain supposedly “consensually validated“ shame or guilt provoking situations. These Items may very well tap other affects or traits such as anxiety (e.g., "You falsify some information on a job application in order to get the job. You're worried about having lied." "You discover that even by running, you will be at least ten minutes late for class“), embarrassment (e.g., "You belch in public." "You meet a friend whose name you have forgotten"), anger (e.g., "You are criticized In front of your peers." ”You overhear your friends making 25 fun of you"). and so forth. In Novak's dissertation these Items from Perlman and Beall's scales formed a cluster labeled as the “situational shame“ factor and he reported that this factor has a low correlation with Inferiorlty (L - .26) and embarrassment (L - .28), and a moderate correlation with fear of exposure (L - .45). Novak also pointed out that the Items measuring this factor do not specify the inner affective experience of the respondent In the situations. He concluded that "perhaps the scale confounds propensity for other negative affects In response to the situation with the propensity for shame" (Novak, 1986, p. 53). A different potential problem with construct validity was found In other studies (e.g., Harder & Lewis, 1987; Hobiltzelie, 1987; Cook, 1988a, 1988b). They did not theoretically differentiate between shame and various other traits or affects that have been associated with shame. As a result, their scales contain items which measure affects other than shame. For example, Harder and Lewis's Personal Feelings Questionnaire (PFO) Includes in the shame proneness scale the following five items: (1) feeling ridiculous, or laughable, (2) embarrassment, (3) feeling humiliated, “stupid," or "childish," (4) feeling helpless, or paralyzed, and (5) feelings of blushing. Hobiltzelle's Adapted Shame/Guilt Scale (ASGS) includes the following words for the shame factor: shy, bashful, mortified, humiliated, abashed, embarrassed, depressed, chided, reproached, and ashamed. Cook's Internallzed Shame Scale (ISS) Includes embarrassment, anxiety, and emptiness Items besides shame Items. The Implicit assumption in these scales is that affects such as embarrassment and shyness are forms of shame. However, If that 26 assumption Is false, then the scales are not pure measures of shame but are contaminated by these other affects. In summary, because of the fact that (1) some scale constructors did not specify the affects measured In their shame and guilt scales and (2) some scale constructors Included many other affects (e.g., embarrassment, shyness, anxiety, or depression) in their shame scales, the correlation of .48 between "shame" and "guilt" scales obtained In the meta-analysis of nine previous studies should best be treated as preliminary. Snames_£uli1+_and_nean§SlQn Assume that shame and guilt can be differentiated. The question Is: Which is more important In the causation of depression? Psychoanalytic thinkers have long emphasized the Importance of negative self-esteem In depression (Freud, 1917; Rado, 1928; Fenichel, 1945; Blbrlng, 1953; Jacobson, 1953, 1964, 1971, 1975). It is the hypothesis of this research that negative self—esteem and shame are the same construct. Thus the psychoanalytic tradition can be Interpreted as predicting a strong link between shame and depression. When viewed from such a perspective, Helen Block Lewis's postulatlon that shame plays a very Important role in depression Is not all that original after all. in other words, Lewis substituted the terminology of shame for negative self-esteem without realizing the similarity of the two constructs. in 1987 Lewis did see the close relationship between the two constructs when she stated that "shame Is the affective-cognitive state of low self-esteem." in the clinical literature of depression, gullt has also long been 27 thought of as an Important factor (e.g., Abraham, 1911, 1916, 1924; Freud, 1917; Rado, 1928; Klein, 1935, 1937; Hamilton, 1967; Beck, 1967). In fact, the widely used clinical assessment Instrument, the Hamilton Depression Rating Scale (Hamilton, 1967). has focused on guilt feelings (item 2) to the exclusion of shame feelings. A group of researchers (Harrow a Amdur, 1971; Prosen, Clark, Harrow, a Fawcett, 1983) reviewed the clinical literature relating guilt and depression and found the empirical evidence for such a clinical position lacking. Harrow and his associates have provided preliminary evidence supporting the greater role negative self-esteem plays In depression than gullt does. Smith (1972) concluded in his dissertation that "there was strong and consistent evidence In support of the hypothesis that shame prone Individuals are more depressed than guilt prone Individuals" (pp. 92—93). Harder and Lewis (1987) and Hobiltzelle (1987) correlated shame and guilt with depression and have found that shame correlates more highly with depression than gullt does. However, existing research does not pit shame and guilt against each other. Rather existing research asks whether each Is correlated separately. The fact Is that both shame and guilt are correlated with depression. But this does not mean that both are causally relevant. Shame and guilt are correlated with each other. Thus If only one of the two traits were directly antecedent to depression, the other trait would still be correlated with depression because of the indirect linkage. There are two extreme cases: (1) shame Is a causal determinant of depression while gullt Is not or (2) gullt Is a causal determinant of depression while shame Is not. Assume the case in which shame is one of 28 the causes of depression while guilt is not. Because gullt Is correlated with shame, It will also be indirectly linked to depression. Thus gullt too would be correlated with depression. The typical literature review on the determinants of depression basically just counts the relative frequency of statistically significant correlations between depression and each predictor. From that point of view, gullt would seem just as good a predictor as shame. The fact that gullt is not a causal determinant would not come out unless someone did multiple regression with both shame and guilt as predictors or someone did partial correlation. The lrrelevance of guilt would show most dramatically In the fact that the partial correlation between guilt and depression with shame held constant would be 0. The case In which gullt Is a causal determinant of depression while shame Is not Is similar. The correlation between shame and guilt would cause shame to be correlated with depression because of Its Indirect linkage. Thus the irrelevance of shame would only show If someone computed the partial correlation between shame and depression with guilt held constant. The partial correlation would drop to O. The key comparison of shame and guilt as determinants of depression Is, therefore, a matter of partial correlation rather than zero order correlation. The past literature has shown that both guilt and shame are correlated with depression. But what would the two partlal correlations show? is gullt still correlated with depression If shame ls held constant? is shame still correlated with depression if gullt Is held constant? This study provides data on these two questions. 29 SnamsL_QuLLLL_and_Q1hQL_££L§ODaLL11_¥1Llahis§ The same question can be asked about other personality variables. How do shame and guilt relate to trait anxiety, to loneliness, to life satisfaction, and so forth? This study will provide data on a wide variety of measures of social and emotional functioning. (2) A11amn1s_to_DI11aLsn1Ia1a_Snama_and_91naL_Be1a1ed_ILaits Here the "shame-related traits” refer to embarrassment, fear of exposure, self-consciousness, shyness, and guilt. The topic of shame and guilt has already been discussed above. W In his dissertation Novak (1986) has Identified four factors In the existing shame measures (Perlman, 1953; Beall, 1972; Cook, 1985): situational "shame", fear of exposure. embarrassment, and feelings of Inferiorlty. He states that embarrassment (or extreme self— ’ consciousness) and feelings of inferiority are not coincident at the trait level (L - .77); nor are fear of exposure and feelings of Inferiorlty (L - .76). Fear of exposure and embarrassment correlate moderately (L - .64) with each other. It should be noted that these three correlations have been corrected for attenuation. Novak concluded that shame, embarrassment, and fear of exposure are three distinct traits. erman (1984) found only a moderate correlation between shame and embarrassment, a correlation of .58 between the Korpl Shame Scale (Korpl, 1977) and the Cattell Susceptibility to Embarrassment Scale. Shame_and_fihxness There have also been studies which suggest a difference between shame and shyness. erman (1984) reported that the Korpl Shame Scale 30 (Korpl, 1977) has only a fairly low correlation (L - .33) with the Revised Stanford Shyness Survey. Mosher and White (1981) present some evidence that shyness and shame are distinct and fundamental emotions as they can be differentiated on the Differential Emotions Scale. Harder and Lewis (1987) reported a fairly low correlation (L - .31) between the PFC Shame Scale and the Stanford Shyness Inventory. Hobiltzelle (1987) reported moderate correlations between ASGS Shame Scale and three Items from the Stanford Shyness Survey (L ranged from .46 to .51). In short, the existing research data strongly suggest that shyness is quite different from shame. sname_and_SeIL:Conscicusness Two studies (Harder & Lewis, 1987; Hobiltzelie, 1987) have found that shame has a low correlation with public self-consciousness (L ranged from .05 to .47). Novak (1986) argued that embarrassment Is the extreme form of self-consciousness. If this Is true, then the embarrassment scales are self-consciousness scales. However, Novak's argument was purely theoretical and there Is not yet any empirical support for that position. Thus the correlation between shame and self- consclousness deserves further exploration. 31 The conceptual schemes defining shame In the literature span all levels of measurement methodology: from measurement models which postulate simple Identity between different aspects of shame to hierarchical measurement models with as many as three levels of shame- reiated traits. This section will briefly present the measurement models used In this analysis. For simplicity, the models will be spelled out for the six shame theme scales where only a simple model was thought to be needed. Ine_snama_lnemas The most elementay Issue Is the expression of shame In the shame literature, In our everyday language, and In clients' verbal reports. Six shame themes were Identified. They are: disappointment with oneself, feelings of Inferiorlty, feelings of defectiveness, feelings of worthlessness, feelings of unlmportance which are often expressed metaphorically as "feeling small,” and feelings of falling short of one's own standards or ideals. These shame themes were developed Into six shame theme scales. The first Issue Is the quality of the shame theme scales. Do the items in each scale measure the Intended construct? The most powerful statistical methodology for this Is confirmatory factor analysis. If the Items are linearly related to the construct that they measure, then confirmatory factor analysis can be used to answer the following question: Do all the Items on a given scale measure the same construct? 32 If all Items measure the same construct, then Item responses differ from one another only by random or unsystematlc error of measurement. Note that confirmatory factor analysis is not a complete answer to the Issue of construct validity (See Hunter and Gerblng, 1982, on this point). The statistical method asks only whether Items measure the same construct, not whether that construct Is the right construct. Whether the Items measure the right construct is largely a substantive question, that Is, a question of content which Is usually tested by looking at correlations between the scale and other constructs rather than by looking at the structure of the item correlations themselves. At the level of the shame theme scales, the Issue of the nature of the construct was dealt with solely in terms of an Item content analysis. The items In each shame theme cluster were closely examined to see If they were psychologically equivalent In both the affect expressed and In the manner In which that affect was expressed. Hunter and Gerblng (1982) note that if the right research design is used, then confirmatory factor analysis can be used to assess item equivalence in two very different ways: "internal consistency" and “parallelism" (or "external equivalence"). If Item responses differ from each other only by random error of measurement, then the item errors will not correlate with each other. The correlations between items within a scale should then satisfy a mathematical product rule discovered by Spearman (1904, cited In Hunter 8 Gerblng, 1982): his one factor model. If the correlations between Items within a scale satisfy the Spearman product rule, then the scale ls said to be "Internally consistent.“ This Is a weak criterion for Item equivalence. 33 There Is a stronger criterion for Item equivalence: parallelism In the pattern of correlations between the Items and Important “outside" variables such as the measures of emotional and social functioning used In the present study. If all Items measure the same construct, then the item errors will not correlate with any outside variable. The correlations between Items In a unldlmensional scale with any outside variable should satisfy a condition called "parallelism“ (Tryon, 1939; Tryon and Bailey, 1970; Hunter, 1973; Hunter and Gerblng, 1982). This Is a strong test for Item equivalence. If an Item is contaminated by some unintended variable, and If that contaminating variable Is one of the outside variables (or is correlated with one), then the Item will correlate more highly with that outside variable than will the other uncontaminated Items. Thus failure to find parallelism not only shows an Item to be contaminated, but It also Identifies the contaminating variable (Hunter, 1986, 1987). Parallelism can either be tested by doing a confirmatory factor analysis Including both the items and the outside variables (as noted by Hunter and Gerblng, 1982) or by computing the correlations between the Items and the outside variables and examining that correlation matrix for parallelism (as noted by Hunter, 1973, and Tryon and Bailey, 1970). Because of the large number of outside variables, this latter method was the method used In this study. Conflrmatory factor analysis was used to examine the quality of each shame theme scale. Each scale was checked for homogeneity of content, for internal consistency, and for external consistency or parallelism. To see If there might be some completely unanticipated dimension In the data, an exploratory factor analysis of the shame theme 34 Items was also run. Once the shame theme scales were shown to be unidlmenslonai, the correlations between these scales were examined to assess higher order measurement hypotheses. Consider first the strongest hypothesis: the Identity hypothesis. if the affect assessed by each theme Is the same affect (l.e., shame) and If all subjects used each semantic theme In the same way In all situations, then the shame theme constructs would be Identical to each other. The correlation between any two shame theme constructs would thus be 1.00. Note that the scales would not be perfectly correlated because of random error of measurement. However, If the correlations were corrected for the attenuation due to error of measurement, then the corrected correlations would differ from 1.00 only by sampling error. The factors of confirmatory factor analysis are the scale constructs and the correlations between factors are thus (implicitly) corrected for attenuation. Thus If two shame theme scales measure the same affect, and If all subjects have the same semantic patterns of word and phrase use for those two shame themes, then the two corresponding shame factors In the confirmatory factor analysis should be perfectly correlated. Even If the affect underlying each shame theme is the same, the shame theme factors would be less than perfectly correlated If subjects differ In their semantic patterns in Idiosyncratic ways. if each subject tends to use particular words In a somewhat Idiosyncratic way, then the shame theme factors will differ from one another because of specific factors that reflect those semantic differences. These specific factors will be trivial as causal factors since they are purely 35 semantic In nature. Nevertheless, these trivial specific factors will cause the shame theme constructs to be less than perfectly correlated. Thus If trivial semantic factors are present, the test for the identity hypothesis will fail for trivial reasons. The shame theme factors will be less than perfectly correlated not because they do not all measure shame, but because they also measure an unintended specific factor. General shame Is not measured by any single shame theme scale taken by Itself. Each shame theme scale construct differs from the shame construct because of Its specific factor. On the other hand, according to the semantic ldiosyncracy hypothesis, the shame factor Is the only factor common to all six shame theme scales. Thus the shame construct Is Identified by the set of shame theme scales even though it Is not measured by any one of the scales separately. The shame construct Is the general factor for the shame theme correlation matrix. The semantic ldiosyncracy hypothesis can be tested empirically by seeing If the shame theme scales satisfy a "hierarchical" or "second order“ measurement model (Hunter and Gerblng, 1982). The hierarchical measurement model ls mathematically the same as the confirmatory factor model except that the shame theme constructs play the role of Item while the shame construct plays the role of general factor. That Is, In confirmatory factor analysis, the scale construct Is the general factor for the items within the corresponding Item cluster. In the shame theme hierarchical measurement model, the shame theme constructs are placed in a cluster and the shame construct Is the general factor for that shame themes cluster. The hierarchical measurement model for the shame themes ls presented In the form of path diagrams in Figures 2 and 3. Figure 2 36 Disappointment with Self / / \ Item 1 Item 2 Item 3 £jgnna_2a. The confirmatory factor model for the first of the shame theme constructs: disappointment with self. General Shame M \\\. Disapp Inferior Defect Worthless I Unlmport l Short ‘ 41 - where Disapp - disappointment with self inferior - feelings of Inferiorlty Defect - feelings of defectiveness Worthless - feelings of worthlessness Unlmport - feelings of unlmportance or "smallness" Short - feeling that one has fallen short of one's Ideals Ejguna_2n. The confirmatory factor model for the shame theme cluster with general shame as the general factor or higher order construct. General Shame ,/./ \ Disapp Inferior Defect / I / I. / \ item Item ... Item Item ... Item Item 1 2 7 8 13 T4 ' Ejguna_za. The combined hierarchical measurement model for all the Items of the shame theme scales. EiguLa_2. The hierarchical model for the shame theme scales showing the path model for the test for Internal consistency. 37 presents the path models used to test the hierarchical model for Internal consistency and Figure 3 presents the path models used to test the hierarchical model for parallelism or external consistency. The hierarchical measurement model Is constructed and tested in phases. The first phase constructs confirmatory factor models for each cluster of primary variables--the shame theme scale Items In the present example. The second phase constructs a confirmatory factor model for the primary cluster factors-~the shame theme constructs In the present example. Finally, the two submodeis can be combined Into the full hierarchical measurement model. Figure 2 Illustrates these phases In the form of the path models that would be used to test for Internal consistency. Figure 2a shows the confirmatory factor model for a typical single shame theme scale: the scale for disappointment with self. The general factor for the Items In the disappointment with self scale Is the disappointment with self construct. The primary measurement model for the full confirmatory factor analysis of all of the shame theme Items would have five more models of similar form, one for each shame theme scale. That Is, the full confirmatory factor model for the shame theme Items consists of 6 submodeis, one for each shame theme scale or "cluster of Items." The combination of the six theme cluster models Is the model for the confirmatory factor analysis of the shame theme item correlation matrix. The factor analysis of the item correlation matrix is called the ”primary" or "first order" factor analysis. To see the role of confirmatory factor analysis In assessing measurment models, it Is necessary to note the ambiguity of a statement 38 such as "Item 1 measures disappointment with self." For a psychometrlcian, this sentence always has a missing final phrase: “except for random error.” That Is, to say that the Item measures a given variable Is not to say that It measures It perfectly, but to say that the deviation In measurement is random or to say that there is no systematic contamination of the Item response by some other variable. Since observed measures of a given construct are always Imperfect, the study will not observe the construct itself, only estimates of that construct. Conflrmatory factor analysis is a method for computing the correlations for constructs from the correlations for observed measures so long as the measures obey the assumed measurement model. Consider the Items written to measure the disappointment with self construct In Figure 2a. if an Item were a perfect measure of the construct, then it would correlate 1.00 with the construct. Since that Item would already measure the construct perfectly, there would be no need for further items and no need for factor analysis. But, for most constructs, there Is no perfect Item. Instead, we usually find that every Item has error of measurement. The scale of items has less error, but ls still not perfect. Even If every item differed from the disappointment construct only by random error of measurement, there would still be error of measurement in the scale (l.e., error in the average response is the average of the individual errors which Is smaller but not usually 0). That Is, the scale too would not have perfect reliability. The measurement model shown in Figure 2a Is a path diagram which shows the Intent of measurement: that each disappointment Item measures disappointment with self and only disappointment with self 39 (except for random error of measurement). in the path diagram of Figure 2a, the disappointment construct itself is not an observed variable but is to be inferred from fit of the path model to the correlations between the disappointment items with each other and with the outside variables. Conflrmatory factor analysis (a) provides the test of the measurement model and (b) estimates the correlation between each item and the construct by analyzing the correlations between items. Suppose that two of the Items were each perfect measure of the construct, then each would correlate 1.00 with the disappointment construct and hence each would correlate 1.00 with each other. If the item correlations are less than 1.00, the item responses must differ from one another. There Is vast evidence in psychology showing that any single response is usually determined in part by random elements. This random element in the item response causes the Item response to differ from the construct score and hence causes the item to correlate less than perfectly with the construct and with the other Items. The lower the correlation between Items, the greater the relative Impact of the random element in the response In comparison with the impact of the construct measured. If the causal agents which cause one response to differ from another are Just random error, then they are causally trivial. Thus the error of measurement in one item will be uncorrelated with the error of measurement in another Item. This is the hypothesis tested by the path model of Figure 2a. if that model fits the data, then the error of measurement for any one disappointment item is uncorrelated with the error of measurement for any other disappointment item. This is the meaning of "Internal consistency" for the disappointment with self 40 scale. The submodei represented by Figure 2a is the model for Internal consistency of the disappointment with self theme scale. That model asserts that the correlations between the shame theme items can be explained by a single causal factor--the disappointment with self construct. If that model fits the data, then the correlations between the disappointment Items are entirely explained by the disappointment factor and the errors In the Individual Items must be uncorrelated with each other. The path model In Figure 2a is a path model for the single theme scale written to measure disappointment with self. There are five other models for the other five theme scales. Each of the six path models represents the test for Internal consistency for that theme scale considered separately. if confirmatory factor analysis were applied to each scale separately, then each of the six models would be tested separately for Internal consistency. However, the six internal consistency models were not tested separately. Instead, one global confirmatory factor analysis was run on the items from all six scales. That Is, all six scales were considered simultaneously within the same confirmatory factor analysis. The test for fit of the global model tests for more than Just the internal consistency of each scale separately. The global test considers the error of measurement In any one item In comparison not Just to the other items In Its own scale but In comparison to all other items in all of the shame theme scales. Fit of the global model not only requires that errors be uncorrelated within each cluster but also requires that errors be uncorrelated across 41 clusters. That is, if the six factor confirmatory factor analysis model fits the data for the item correlation matrix for the combined scales, then the error of measurement for any one shame theme Item must be uncorrelated with the error of measurement with any other shame theme item; whether that Item Is In the same scale or not. Figure 2b shows the confirmatory factor model for the “second order" factor analysis which defines the general shame construct. in this model, the ”general factor“ Is the general shame theme factor and the “Items" are the six shame theme constructs placed together in a cluster. The correlation matrix for this analysis Is the factor correlation matrix from the confirmatory factor analysis of the shame theme Item correlation matrix. That is, the factor correlation matrix from the primary factor analysis becomes the Item correlation matrix for the second order factor analysis. Consider the shame theme scale construct factor correlation matrix produced by the primary confirmatory factor analysis, a 6 by 6 correlation matrix. if there were no ldiosyncracy In the shame themes, then the shame theme constructs would be perfectly correlated with each other because each would be perfectly correlated with the general shame construct. The second order factor analysis would then be unnecessary. if the shame theme constructs are not perfectly correlated, then they differ from one another. These differences can be quantified by “specific factors" where the specific factor for each shame theme construct is the algebraic difference between the construct score and the score predicted by the general shame construct. That Is, the specific factor for each theme construct is the error in predicting that 42 construct from the general theme construct. The semantic ldiosyncracy hypothesis asserts that the specific factors are substantively trivial. If the specific factors are trivial, then they will not correlate with anything else. in particular, the specific factors for the shame theme constructs will not correlate with each other. The path diagram In Figure 2b tests that assumption. If the path model in Figure 2b fits the data, then the data Is consistent with the hypothesis that the correlations between the shame theme constructs can be explained by a single antecedent causal factor (l.e., general shame). The specific factors must thus be uncorrelated with each other. Figure 2c shows the full hierarchical measurment model in the form of the path model which tests for Internal consistency. The general shame construct is the overall general factor. Under it (and causally dependent on it) are the six shame theme constructs. Under each shame theme construct are the items which measure that construct. If there Is perfect fit for the primary confirmatory factor analysis of the combined set of shame theme items, and there Is perfect fit for the second order factor analysis, then perfect fit for the combined model is guaranteed. Thus the combined model does not require an additional test for Internal consistency. if the primary model Is Internally consistent within each cluster and satisfies external consistency across ciusters--impiled in fit of the confirmatory factor analysis of the combined item correlation matrix--then fit of the second order analysis Guarantees internal consistency across the combined model. Figure 3 differs from Figure 2 In that outside variables are added to each path diagram. This corresponds to the hypothesis that 43 Other Personality Disappointment with Self Variables item 1 item 2 item 3 EiguLe_aa. The confirmatory factor model for the first of the shame theme constructs: disappointment with self. Other Personality General Shame Variables Disapp Inferior Defect Worthless } Unlmport Short | where Disapp disappointment with self inferior - feelings of Inferiorlty Defect - feelings of defectiveness Worthless - feelings of worthlessness Unlmport - feelings of unlmportance or ”smallness” ‘ Short - feeling that one has fallen short of one's Ideals EiguLe_3D. The confirmatory factor model for the shame theme cluster with general shame as the general factor or higher order construct. Other Personality General Shame Variables Disapp inferior Defect /\ Z A / item Item ... Item item ... Item item 1 2 7 8 13 14 EignLg_ag. The combined hierarchical measurement model for all the items of the shame theme scales. EiguLe_3. The hierarchical model for the shame theme scales showing the path model for the test for external consistency or parallelism. 44 errors of measurement and specific factors are causally trivial. If errors and specific factors are causally trivial, then they will not correlate with each other—-the assumption of internal consistency—-and they will not correlate with Important outside variabies-—the assumption of parallelism or external consistency. This assumption ls represented In the combined path model of Figure So by the fact that there Is no causal arrow connecting an outside variable to any item or to any of the six theme constructs. This model asserts that all causal connections between the items in a construct cluster and any outside variable are mediated by the corresponding item construct, the assumption shown explicitly In Figure 3a. This model asserts that all causal connections between the shame theme constructs and any outside variable are mediated by the general shame construct; the assumption shown explicitly In Figure 3b. if any item were contaminated by one of the outside variables, then the Items in that cluster would not be parallel and the measurement model would be disconfirmed. if any one of the shame themes were related to an outside variable differently from one of the other shame themes, then the shame theme constructs would not be parallel and the measurment model would be disconfirmed. From a substantive point of view, the test for parallelism Is both stronger and more Important than the test for internal consistency. it Is parallelism that is the basis for the use of correction for attenuation to eliminate the bias in correlations produced by error of measurement. Thus the test for parallelism is the test which directly iustifies the use of correction formulas. Since the correction formulas are Implicit In confirmatory factor analysis, it Is the test for 45 parallelism that Is the heart of the assumptions for confirmatory factor analysis. The same is true of the hierarchical measurement model. it is the test for parallelism that corresponds directly to the assertion that the specific factors are causally trivial. lhfi.§hamfi_£LQIiin Joffe (1984) and Kaufman (in press) place their main developmental focus on shame about specific aspects of the self rather than on shame toward the self as a whole. That Is, they stress the causal primacy of specific shame rather than general shame. The vulnerability theory of shame development (Novak, 1986) assumes, on the other hand, that because general shame measures shame vulnerability, general shame Is the primary causal determinant of the shame profile component variables. Figure 1b (on page 14) presented vulnerability theory as a path model. That figure can also be interpreted as a hierarchical measurement model path diagram for Internal consistency. Vulnerability theory predicts that the correlations between the specific shame component variables satisfy a product rule that Is identical to the Spearman (1904) one factor model. Thus if factor analysis were applied to the shame profile component variable correlation matrix, then there should be one general factor and that general factor would be general shame. The difference between the two models In terms of empirical assessment is that In a hierarchical measurement study the general factor Is not usually measured. Rather the general factor ls inferred from the pattern of correlations among the primary variables and Is only 46 estimated by summing the primary variable scores. in this study, general shame Is measured separately by the items of the shame theme scales. Thus the path model of Figure ID (on page 14) can be tested directly rather than indirectly through confirmatory factor analysis. There Is a second difference between vulnerability theory and a strict hierarchical measurement model that emerges when the Issue of parallelism Is considered. If the vulnerability path model fits the data for the correlations between the shame profile variables, then it is possible to define specific factors that differentiate between the shame profile variables. Each specific factor Is the difference between the corresponding shame profile construct score and the score predicted by general shame. Both models then predict that the specific factors will be uncorrelated with each other. According to the hierarchical measurement model, the specific factors are trivial in nature and hence would also be uncorrelated with any outside variable. However, the vulnerability model does not necessarily make that prediction. According to the vulnerability model, general shame differentiates to different levels of specific shame because of Idlosyncracies In shaming experiences. Consider one specific aspect of the self such as the body. Shame vulnerability determines the impact of any one potentially shaming experience. However, the amount of body shame that accumulates Is also a function of the number of such shaming experiences. If everyone had the same number of potential body shaming experiences, then body shame would be completely determined by shame vulnerability. That Is, If the number of experiences were the same for everyone, then body shame would be perfectly correlated with shame 47 vulnerability and there would be no specific factor for body shame. But the Idiosyncracles of life will produce differences In the number of potentially shaming experiences. Suppose one person experiences more potentially shaming experiences about their body than do other people. Then that person will develop more body shame than would be predicted by his or her level of shame vulnerability. Thus the specific factor score for body shame for that person would be positive. Thus according to vulnerability theory, the specific factors are not necessarily trivial in nature. Rather they measure the extent to which the person has had more potentially shaming experiences in one area than another. But shaming experiences may produce reactions other than or In addition to shame. For example, a shaming experience might produce anger or fear, and so forth. Suppose that a child Is often shamed by his or her parent for aliedged Inadequacles In social functioning. if the child Is vulnerable to shame, then he or she will develop relationship shame. However, the person might also develop fear of others. If this fear leads to avoidance of others, then it might also produce loneliness. If this were true, then it would imply a correlation between the specific factor for relationship shame and loneliness. This correlation would then be a departure from the parallelism predicted by the hierarchical measurement model. Assume that the correlations between the shame profile variables satisfy the path model for vulnerability theory. The test of the triviality of the specific factors can be computed using partial correlation. Compute the partial correlations between each shame 48 profile variable with each outside variable holding general shame constant. These are the sample partlal correlations between the specific factors and the outside variables. The hierarchical measurement model predicts that every such population partial correlation will be 0. That Is, all study partial correlations should differ from O by no more than sampling error. if one of those correlations dIffers significantly from 0, then it may be that that correlation would be explained by the differential experience hypothesis of vulnerability theory. For example, the previous paragraph presented a scenario which would predict a positive partial correlation between relationship shame and loneliness with general shame held constant. finam§_finialod_99nalLu£1§ Various authors have linked shame to certain other constructs. The "shame-related constructs" are: embarrassment, fear of exposure, self- consciousness, shyness, and guilt. There have been four positions taken relating these constructs to shame and to each other: the "Identity position,“ the "variants position," the "shame complex position," and the "distinct traits position." These positions can be differentiated empirically by testing various hierarchical measurement models. Consider the "identity position." For some authors one or another of the shame-related constructs Is a part of the shame experience itself. For example, some have taken extreme embarrassment (e.g., “I would like to crawl Into a crack In the earth") to be the epitome of shame. Consider a model which assumes that all of the shame-related constructs are aspects of the shame experience. A measure of the 49 frequency of one of these constructs would thus be identical to a measure of the frequency of shame itself. For example, a measure of the frequency of embarrassment would be a measure of the frequency of shame. This predicts that there would be a perfect correlation between shame and embarrassment. If all the shame-related constructs were aspects of the primary shame experience, then each would be perfectly correlated to shame and hence each would be perfectly correlated with the others. The "Identity position” Is thus an extreme case of a hierarchical measurement model. Once confirmatory factor analysis has eliminated the effect of error of measurement, the shame-related construct factors would be perfectly correlated with each other and with the general shame factor. Thus each of the shame-related constructs would be identical to the general factor for those constructs. Given perfect correlations between two shame—related constructs, parallelism Is mathematically guaranteed. Thus the test for internal consistency with perfect correlations Is both necessary and sufficient for this position. Consider the "variants position.” Some authors have argued that shame Is the primary affect In a number of emotional experiences Including some or all of the shame-related construct emotions. That is, these authors would argue that while the cognitive or behavioral components of the experience might differ from one to another, the affective experience Is always shame. Because the nonaffective components can occur with different frequencies for different people, the frequency of a given shame-related construct emotion need not be perfectly correlated with the frequency of shame. For example, embarrassment need not be perfectly correlated with general shame. The 50 most straightforward model for this position Is obtained by assuming that the nonaffective components of the emotional experience are the causal determinants of specific factors In the frequency of the various emotions. Thus this position could be modeled as a hierarchical measurement model with shame as the general factor. This model is presented In path analytic form In Figure 4. General Outside Shame Variable Embarrass- Fear of I Public Shyness Guilt ' ment Exposure S-C J~——~u—~—«J EiguLe_A. The path model representing the "variants position” which argues that the five affects shown are all variations of shame. Since general shame ls measured separately from the shame-related constructs, the hierarchical measurement model in Figure 4 need not be tested using confirmatory factor analysis. Rather path analysis can be used directly. The test for Internal consistency Is a typical recursive path model and can be tested In the usual way. The test for external consistency or parallelism can be tested in either of two ways. First, the correlations between the shame-related constructs and the outside variables can be computed and assessed for similarity of pattern In the usual way. Second, the partial correlations between the shame-related constructs and the outside variables can be computed with general shame held constant. The hierarchical measurement model predicts that every such partial correlation should be 0 (in the population, differing by only sampling error In the study sample). 51 The "Identity position" can be viewed as a special case of the "variants position." It Is the "variants position“ taken together with the assumption that there Is a perfect correlation between shame and each of the shame-related constructs. Consider the “shame complex position." Some authors would argue that shame Is the primary effect for the shame-related constructs. However, they might doubt whether the general shame scale measures the right general shame factor. They might worry that the shame theme scales also measure only a variant of the shame affect. In this case, the "general shame" factor of the shame theme scales would be better treated as one of the shame-related constructs rather than as the explanatory factor. The hierarchical measurement model for this position is shown in Figure 5. The difference between Figure 5 for the "shame complex” model and Figure 4 for the "variants position" Is that the shame theme factor is now one of the primary variables rather than the general factor for the complex of shame-related constructs. /\L“ General Shame Outside Factor Variable Shame Theme Embarrass- Fear of Public Shyness Guilt ' Factor ment Exposure S-C J ElguLe_5. The path model for the “shame complex position" which argues that (1) the traits shown are variants of shame and (2) that the general factor for the shame theme scales is also only a variant of shame. The general shame factor for the shame complex model is not Independently measured. Thus It must be Inferred from the pattern of 52 correlations between the shame-related constructs Including the shame theme factor. This can be tested using a second order confirmatory factor analysis. The primary confirmatory factor analysis declares six Item clusters: the Item cluster for shame theme items, for embarrass- ment, for fear of exposure, for public self-consciousness, for shyness, and for gullt. The six variable factor correlation matrix from this primary factor analysis is then used to assess the Internal consistency of the second order measurement model. That Is, the confirmatory factor analysis of the six variable shame-related construct matrix tests the internal consistency of the hierarchical measurement model that represents the “shame complex position.“ Parallelism or external consistency can be tested by assessing the matrix of correlations between the six shame-related constructs (including the shame theme factor) and the outside variables in the usual manner. if partial correlation Is to be used to test for parallelism, then confirmatory factor analysis must be used to estimate the correlations between the general factor for the complex and the outside variables. To do this a confirmatory factor analysis must be run that Includes the outside variables as well as the shame-related constructs. If there are Item clusters for each outside variable as well as for the six shame- related constructs (Including a shame theme cluster with all shame theme Items), then confirmatory factor analysis will estimate a correlation matrix with (a) the general shame factor of the shame complex, (b) the shame complex variables, and (c) the outside variables. The desired partlal correlations can be computed from that matrix. Compute the partial correlations between the shame-related constructs (Including the 53 shame theme factor) and the outside variables with the shame complex general factor held constant. The hierarchical measurement model of the "shame complex position“ predicts that each such partial correlation will not differ from O by more than sampling error. The “variants position" can be viewed as a special case of the “shame complex position." The "variants position" would make the extra assumption that there is a correlation of 1.00 between the shame complex general factor and the shame theme factor. The "variants position“ and the "shame complex position" can survive failure of the hierarchical measurement models of Figures 4 or 5 If the failure follows one of certain patterns. In particular, the hierarchical models in Figures 4 and 5 assume that all of the shame- reiated constructs are variants of the primary shame affect. It may be that one or more of the shame-related constructs is not part of the shame complex. in that case, the appropriate model would be the model for the subset of constructs which do fit the measurement assumptions. it Is also true that If a few of the partial correlations differ only slightly from O the fundamental position may be correct. it would mean that there Is some correlation between the environmental factors which cause the nonaffective components of the emotional experiences to differ In frequency. If all of the measurement models fall to fit the data, then the "distinct traits position" is confirmed by default. That is, the “distinct traits position" is defined negatively. it merely asserts that there are causal relations between the shame-related constructs that are not mediated by a general shame factor; causal relations that 54 would presumably be more easily predicted by considering the constructs in Isolation rather than as part of a complex. METHOD Slufl1_finai§ This study deals with the questions about the nature of shame using a cross-sectional design measuring the personality traits and affects of normal adults. The methodology employed Is that of confirmatory factor analysis and hierarchical measurement. There are three sets of variables assessed In the study: (a) direct measures of shame which Include shame theme scales and shame profile scales, (b) measures of shame-related traits or affects thought to be part of, to be variants of, or to be directly associated with shame, namely, embarrassment, fear of exposure, public self-consciousness, shyness, and guilt, and (c) other measures of emotional and social functioning. The study was designed around a sequence of measurement Issues. These issues can be developed in a sequence of analyses using confirmatory factor analysis and the hierarchical measurement model. The issues and analyses will be considered In sequence. The most elementary Issue Is the expression of shame In the shame literature, in our everyday language, and In clients' verbal reports. Six shame themes were identified: namely, disappointment with oneself, feelings of Inferiorlty, feelings of defectiveness, feelings of worthlessness, feelings of unlmportance which are often expressed 55 56 metaphorically as "feeling small," and feelings of falling short of one's own standards or ideals. These shame themes were developed into six shame theme scales. Conflrmatory factor analysis was employed to examine the dimensionality of each shame theme scale. Each scale was examined for homogeneity of content, for internal consistency, and for external consistency. It was tested to see If a hierarchical measurement model with a higher order factor called general shame fits the data. The hierarchical measurement model was tested for its Internal consistency as well as external consistency. if both tests are satisfied, It means that general shame factor Is causally antecedent to these shame themes and that these shame theme scales fit a second order unldimensionai construct model. The second main issue Is the structure of shame feelings about different aspects of the self, that Is, the shame profile. Lynd (1958) and Lewis (1971) both emphasize the ”whole-self" Involvement as being one of the distinguishing characteristics of shame. Joffe (1984) and Kaufman (In press) have proposed “shame profile" models. Furthermore, Kaufman has attempted to develop a developmental theory of shame. This theory assumes that shame is first developed in specific circumstances In regard to particular aspects of the self. Shame then generalizes from the particular aspects of the self to a feeling of general shame about the self as a whole. In contrast, Novak (1986) noted that his data suggest the opposite: that is, general shame ls causally antecedent to specific shame. His data suggest that some people are more vulnerable to shame than others. Those who are vulnerable tend to develop shame about all or most aspects of the self. The questions are 57 these: (1) How high are the correlations between different specific shame affects? (2) how high are the correlations between different specific shame affects and general shame? and (3) do the causal relations run from general shame to specific shame or from specific shame to general shame? Again, the analytic methodology employed is that of confirmatory factor analysis and hierarchical measurement. Once the measurement Issues concerning the shame construct are dealt with, then the focus of this study Is the relationship of shame and self-esteem. It will be examined in two ways: (1) How high Is the correlation between these two constructs? (2) are they parallel in terms of their correlations with relevant outside variables? Essentially, the question is about dimensionality. Are they the same dimension or are they not? Then the focus is the relationships between shame and allied constructs such as embarrassment, fear of exposure, public self- consclousness, shyness, and guilt. it has been noted that there have been four positions taken by authors In the shame literature on this issue. They are called: the ”Identity position," the "variants position,“ the "shame complex position," and the "distinct traits position." The methodology is that of second order confirmatory factor analysis and hierarchical measurement. Furthermore, gender differences in shame and guilt proneness are assessed as are the relationships of shame and guilt to depression and other emotional problems. 58 Slud1_¥aLiahiQS The variables measured In this study are those necessary to test the various measurement models. Each measurement model has two aspects: internal consistency and external consistency or parallelism. Internal consistency Is assessed by checking to see whether a set of Items or scales correlate with each other In the predicted manner. Parallelism Is a check to see If the Items or scales correlate with outside variables In the predicted manner. Thus, there were two kinds of variables to be measured in this study: (1) measures of various constructs thought by some authors to be directly related to shame and (2) outside variables that are thought to be causally related to shame but are not a part of the definition of shame. Shama_§QnSILu£1§ There are three levels of variables that have been postulated to be part of the shame definition by various authors. First, there are the shame themes, the elements of the phenomenology of shame. Six such themes were developed for this study. Second, there are the specific kinds of shame such as competence shame or body shame. Four specific kinds of shame were Identified for this study, that is, competence shame, relationship shame, body shame, and character shame. Third, there has been considerable disagreement as to how certain constructs, namely, embarrassment, fear of exposure, public self-consciousness, shyness, and guilt are related to shame. Are these emotional states alternate forms of shame or are they different states? All five constructs were measured for this study. 59 Qulsinc_¥aLlaniaS In order to test parallelism, the "outside" variables can be chosen on two bases: (1) to be important causal antecedents or consequences of shame or (2) to be possible contaminants In shame measurement. The main criterion used here was clinical relevance. The outside variables were those most often used In clinical research, that Is, measures of the main dimensions in emotional and social adjustment. In particular, one of the goals of the study was to assess the relationship between shame and self-esteem. For this purpose, several measures of self—esteem were included In the study. Slalifiligal_M§1nQfl§ The statistical methods used are those of path analysis and confirmatory factor analysis. Extensive use of hierarchical measurement models was made. in order to make it easier to detect possible errors In models, ordinary least squares estimation methods were used (Hunter and Gerbing, 1982). Partial correlations were used in this study as well. The statistical program used on the IBM mainframe computer was that of EAQKAQE (Hunter, 1988; Hunter, Gerbing, Cohen, & Nicol, 1980). OPERATIONAL METHODOLOGY Sublects 926 university students (249 males, 677 females) who took undergraduate psychology courses participated in this research. These students received extra credit for their participation in the study. No monetary reward was provided. ELnsnnuLa The inventory (as shown In Appendix A-2) was given to each study subject. They could fill the questionnaire out In a setting of their choice. They were assured of anonymity. instLuments All the scales used In this study are included in Appendix A—3 and are described as follows: Ihe_§name_1name_acaies The Shame Theme Scales (Chang & Hunter, 1988) have been constructed specifically for this study. The main reasons for constructing new scales instead of using one of the existing shame measures are twofold. First of all, with the exception of the Internallzed Shame Scale (Cook, 1985, 1988a, 1988b), most existing scales (modeled after Perlman's Anxiety Attitude Survey) used situational Items which did not specify the inner affective experience of the respondent in the situations. They may very well tap affects or traits other than shame (as explained 60 61 on pages 24-26). In Novak's dissertation these items formed a cluster labeled as the "situational shame" factor. He reported that this factor has a low correlation with Inferiorlty (L - .26) and embarrassment (L - .28). Novak concluded that these scales probably tapped affects other than shame: anxiety, anger, embarrassment, and so forth. Second, even though Cook's scale is an Improvement over these other scales, It measures mgLe than one factor (inferiority, embarrassment, anxiety, and emptiness). Cook as well the other scale developers (e.g., Harder & Lewis, 1987; Hobiltzelie, 1987) did not differentiate shame from other related traits. Therefore, a decision was made to construct new shame scales with these differentiations In mind. An examination of major writings in shame literature, of our everyday language, and of clients' verbal reports yielded six shame themes: namely, disappointment with oneself, feelings of Inferiorlty,- feelings of defectiveness, feelings of worthlessness, feelings of unlmportance which are often expressed metaphorically as “feeling small," and feelings of falling short of one's own standards or ideals. These shame themes were developed into six shame theme scales. An extensive effort was made to look In thesauri as well as the shame literature for words or phrases and their synonyms for the purpose of Item writing. After shame items had been written, more items were written to describe people who feel good about themselves, who see themselves as having personal worth, who are happy with themselves, who see themselves as competent people, and so forth. initially, 65 items were written altogether for six shame theme scales. Mainly on the basis of content meanings, 22 best items were retained for this study. Of 22 62 Items 4 were worded In the positive direction. Ins_§hama_ELQIiin_§£aiaa items were written to tap Kaufman's shame profile. The definition of these four different kinds of shame has been given on pages 10—11. initially, 53 Items were written. Again, on the basis of content meanings, 32 items were retained for this study. MaaSuLaa_Q1_ILal1a_QL_A11n31§_A§§Q£ialan_uilh_§namn The 8-item Embarrassment Scale was based on Novak's embarrassment cluster in his dissertation (1986). Some Items were taken from the older version of the Internallzed Shame Scale (Cook, 1985); some were rewritten or added for this study. The trait of embarrassment Indicates a disposition for Intense feelings of exposure and heightened self- consclousness along with the Impulse or desire to hide. The Self-Consciousness inventory was written by Fenlgstein, Scheler, and Buss (1975). They define self-consciousness as the consistent tendency of a person to direct attention Inward or outward. Private self-consciousness refers to the tendency to attend to one's inner thoughts and feelings, to focus attention on the private aspects of the self. Public self-consciousness Involves a particular awareness of oneself as a social object, which means that those who are high In this trait are likely to think about themselves frequently as objects of the attention of others and are likely to be concerned about others' opinions of them. The 7-Item Fear of Exposure Scale was based on Novak's cluster In his dissertation. Some Items were taken from the Beall Shame—Guilt Test (Beall, 1972); some were added for this study. The fear of exposure 63 trait Indicates a disposition to worry or to be apprehensive about others' opinions or scrutiny of one's self or aspects of one's self, as well as a desire or tendency to conform, to hide aspects of one's self, or to fit in (Novak, 1986). The Shyness Scale was constructed by Cheek and Buss (1981). They define shyness as tension, concern, discomfort, and Inhibition in social contexts. They reported that the scale has a .79 coefficient alpha. An Independent study reported a .88 alpha value (R. Hunter, 1984). Cheek and Buss further found that shyness correlates .50 with fearfuiness and -.51 with self-esteem and that the scale predicts how people will behave when meeting strangers. Briggs and Smith (1986) state that the Cheek and Buss Shyness Scale and four other self—report shyness scales all appear to measure more or less the same construct. In addition, all five scales show a remarkably similar pattern of correlations with other self-report measures. in short, the Shyness Scale has good construct validity. The Buss and Durkee Guilt Scale (1957) has been modified for the purpose of this study. The original scale has nine items. Three items have been deleted and some wording has been modified. it is obvious that Buss did not differentiate guilt from shame when he and Durkee constructed their gullt scale. Twenty three years later, he went to some lengths to distinguish shame from gullt (Buss, 1980). A 5-Item short form of the Buss and Durkee Guilt Scale has been used in a few studies before (e.g., Harrow a Amdur, 1971; Prosen, Clark, Harrow, s Fawcett, 1983). A new 9-Item guilt scale (Chang & Hunter, 1988) was constructed for this study. it should be mentioned that the definition 64 of guilt guiding the Item-writing comes from Friedman (1985). in his paper, IQwaLd_a_Beggngentuailzatign_gi_§u111, Friedman defines guilt as “the appraisal, conscious or unconscious, of one's plans, thoughts, actions, etc. as damaging, through commission or omission, to someone for whom one feels responsible" (p. 529). Guilt motivates a person to avoid harming others and to make reparation If harm is done. Iwo_SeL1:Es1aem_Scaies Two well-established self-esteem scales are included: the Rosenberg Self-Esteem Scale (1965) and the Cheek and Buss Self-Esteem Scale (1981). The former measure is regarded highly and is one of the best known self-esteem scales. Cheek and Buss found that their scale correlates .88 with Rosenberg's scale. if the correlation ls corrected for attenuation, they correlate 1.00 with each other. Iha_inlnLnaiizsfi_Snamn_§£aie A 11-Item short form of the internalized Shame Scale (Cook, 1985) was used as an alternative shame measure In this study. In his dissertation Novak (1986) used the older version of Cook's scale. On the basis of Novak's factor analysis, 11 items are selected from the internalized Shame Scale. Out of these 11 items 10 are Identical with those items labeled as the Inferiorlty factor by Novak. None of these 10 items belonged to the embarrassment factor in Novak's findings. The one extra Item Is from the new version of Cook's scale and shares similar content meaning with the other 10 Items. The two self-esteem scales mentioned above and the Internallzed Shame Scale (short form only) were used to validate the new shame scales. 65 acaies_MeasuL1nq_Emo1iona1_Adius1men1_£Lpbiems Scales that measure emotional adjustment problems include the following: (1) The Beck Depression Inventory (Beck, Ward, Mendeison, & Erbaugh, 1961; Beck & Beck, 1972): This is one of the most commonly used self-rating scales in depression research. The 13-Item version was used in this study Instead of the 21—item version. Beck and Beck (1972) did not report the reliability of the short form. They only reported that the short form correlates .96 with the original BDI. (2) The Trait Anxiety Scale (Spielberger, 1983; Splelberger, Gorsuch, & Lushene, 1970): This is a very commonly used anxiety scale. A 10-item short form was used In this study instead of the 20—Item scale. Schnarch (1976) reported a .86 coefficient alpha for the 10-Item scale. (3) The Trait Anger Scale (Spielberger, Jacobs, Russell, a Crane, 1983): The 10-Item scale correlates between .66 to .73 with the Buss- Durkee Hostility Inventory (Buss a Durkee, 1957) which Is generally regarded as the most carefully constructed psychometric measure of hostility. The trait anger scale has a very satisfactory degree of reliability (alpha about .90) In a sample of more than 1600 college students. (4) The Loneliness Scale (Russel, Pepiau, a Cutrona, 1980): Recently, Hays and DiMatteo (1987) have shortened the scale to 8 items. They have reported that the short form has a .84 coefficient alpha. They have shown that the short form Is reliable, valid, and a practical alternative to the 20—ltem scale. 66 (5) The Rathus Assertiveness Schedule (Rathus, 1973): This is the most widely used and Investigated assertion measure. The 30-item scale correlates .80 with the College Self-Expression Scale which is another very commonly used Inventory (Galassi & Gaiassi, 1980). On the basis of Rathus's item analysis and Gaiassi and Gaiassl's factor analysis, the Rathus scale was reduced to 17 items for this study. (6) The Satisfaction With Life Scale (Diener, Emmons, Larsen, & Griffin, 1985): This 5-item scale is narrowly focused to measure global life satisfaction. it has been shown to have a high reliability (.8? coefficient alpha) and to correlate moderately to highly with other measures of subjective well-being (Diener et al, 1985). (7) Buss and Durkee Resentment Scale (1957): A 3-item short scale Instead of the original 8-ltem scale was used in this study. This short form was used In two previous dissertations (Schnarch, 1976; R. Hunter, 1984) and was found to have .49 and .58 coefficient alpha respectively. We Scales that measure social functioning include the following: (1) The Conformity Scale (Fenlgstein, 1975): This scale measures a person's tendency to talk or act to match those about him or her rather than to express personal views or feelings. (2) The Egocentrism Scale (Hunter, 1983): This scale taps a person's tendency to think that other people think about him or her. (3) The Exhibitionism Scale (Hunter, 1983): This scale taps a person's propensity to enjoy being the center of attention. (4) The Dominance Scale (Saizman a Hunter, 1983): This scale measures a person's tendency to dominate social situations. 67 (5) The Self-Revelation Scale (Schnarch & Hunter, 1976): This scale measures a person's propensity to reveal personal feelings to other people. (6) The Competitiveness Scale (Gerbing a Hunter, 1979): This scale taps a person's tendency to adopt a competitive stance towards others. (7) The Need for Approval Scale (Saizman a Hunter, 1983): This scale measures a person's propensity to worry about what others might think of his or her actions. (8) The Audience Anxiety Scale (Buss, 1981): This scale taps a person's tendency to be fearful and anxious about public speaking. Table 1 lists outside variables used to validate key scales in this study. Table 1 Lis1_Q1_Qu1side_VaLiabies_and_item_SouLce Variable Name Item Source Depression Beck & Beck (1972) Anxiety Splelberger et al. (1970) Anger Splelberger et al. (1983) Loneliness Russel et al. (1980) Assertiveness Rathus (1973) Satisfaction with Life Diener et ai. (1985) Resentment Buss & Durkee (1957) Conformity Fenlgsteln (1975) Egocentrism Hunter (1983) Exhibitionism Hunter (1983) Dominance Salzman & Hunter (1983) Self-Revelation Schnarch & Hunter (1976) Competitiveness Gerbing & Hunter (1979) Need for Approval Salzman a Hunter (1983) Audience Anxiety Buss (1981) RESULTS in order to investigate the relationship between shame and self- esteem, the first step Is to examine the construct of shame. Thus the first results assess the measurement aspects of shame instruments constructed for this study. The shame instruments are of two types: shame theme scales and shame profile scales. The phrase “exploratory factor analysis” here means: (1) communailtles were estimated as the largest correlation, (2) principal axis factors were followed by VARIMAX rotation, and (3) the VARIMAX factors were used to generate a "blind“ confirmatory factor analysis. However, It is the VARIMAX factors that were used for the purposes of this research. The eigenvalue cutoff for the number of factors was initially set at 1.00, though this was often subsequently reduced to see If the factors thus forced were meaningful. I. Inames_1n_1ne_£aness1en_QL_sname A. Six Shame Theme Scales--Deveiopment and Testing During the stage of Item writing, shame Items were written according to six themes. These shame themes refer to different ways people express how they feel about themselves, to different words used in reference to shame feelings. One question for this research was whether these shame themes refer to different dimensions of shame. 68 69 1. Conflrmatory Factor Analysis Twenty two shame items were grouped in six shame theme clusters on the basis of content meanings: disappointment with oneself, feelings of inferiority, feelings of defectiveness, feelings of worthlessness, feelings of unlmportance, and feelings of falling short of one's own standards or Ideals. Then a confirmatory factor analysis was performed on these a priori clusters. The result is In Table B-1 In the Appendix. It shows that a priori clusters hold up quite well In confirmatory factor analysis; the Items within each theme cluster appear to be equivalent to each other. 2. Exploratory Factor Analysis An exploratory factor analysis was performed on the 22 shame theme Items. When eigenvalue was set at 1.00, only one factor was found. When a second factor was forced, it was found to be meaningless. That Is, a two-factor solution with eigenvalues set at .50 or .60 yielded two factors that correlate .90 with each other. As far as content meanings are concerned, these two factors are not different. The key fact is that the exploratory factor analysis did not reveal some unanticipated dimension in the shame theme items. The secondary fact Is that the exploratory factor analysis was not able to detect the shame theme specific factors. The shame theme specific factors are so small in variance accounted for that some of the random sampling error factors turned out to be larger. 3. Parallelism Items in each of the six shame theme clusters were examined in terms of how they correlate with outside variables. Summary data for 70 outside variables are listed In Table 8—2 In the Appendix. The correlations of the six shame theme cluster items with outside variables are listed in Table B-3 In the Appendix. it appears that Items within each cluster are parallel to each other In terms of their correlations with outside variables. B. Hierarchical Measurement Model Table 2 shows the means, standard deviations, and reliabilities of the six shame theme scales and the general shame scale. The general shame scale Is the sum of all the items in the six shame theme scales. All these scales have very respectable reliabilities (with coefficient alpha ranging from .71 to .95). Table 2 Meansz_S1andaLd_Daxia1ions_and_Ba1iab11111es_91_5i1_§name_lheme Scaies_and_GeneLa1_Shame_§caie Number Standard of Items Mean Deviation Reliability Disappointment with Self 4 9.13 2.68 .80 Feelings of Inferiorlty 4 8.27 2.91 .84 Feelings of Defectiveness 4 9.76 2.99 .75 Feelings of Worthlessness 6 10.33 3.88 .89 Feelings of Unimportance 2 3.41 1.49 .76 Falling Short of standards 2 5.52 1.72 .71 General Shame 22 46.41 13.65 .95 The lntercorreiatlons between six shame themes are shown In Table 3. The six shame themes are very highly correlated with each other. Furthermore, the pattern of the correlations is that for a single factor. THIS leads to the hypothesis that the SIX shame themes all 71 measure the same general shame factor. Table 3 1n1aLcQLLe1ations_amQnn_§ix_Shame_Iheme_ScaLes 2 3 4 5 6 1 Disappointment with Self .93 .95 .91 .87 .85 2 Feelings of Inferiorlty .92 .90 .89 .77 3 Feelings of Defectiveness .86 .83 .84 4. Feelings of Worthlessness .94 .68 5 Feelings of Unimportance .70 6 Falling Short of Standards 1. internal Consistency a. Path Model A hierarchical measurement model has been constructed with regard to the causal relationship between the general shame factor and the six shame themes as shown in Figure 6. A higher order confirmatory factor analysis produces the correlations between the general shame factor and six shame theme constructs. General Shame i i 7 Disappointed Inferior Defective Worthiess}{ Unimpor-[‘ Failing with Self (l) (D) (W) i tant Short ‘ (03) (U) (FS) . The hierarchical measurement model of general shame: A model for the test for internal consistency. 72 b. Fit of the Path Model Hunter and Gerbing (1982) have shown that if all the items in a cluster measure the same factor, then the correlations between the items will satisfy the Spearman (1904) "product rule for Internal consistency." That Is to say, if X and Y are two Items In the same unldimensionai cluster and T Is the cluster true score, than the correlation between the items should satisfy the product rule: rxy ' rxTryT That is, the correlation between two items In the same cluster should be the product of their correlations with the underlying trait. There are, then, two steps to test for internal consistency: (1) estimate the parameters rxT and ryT from the data, and (2) see if the product rule reproduces the Inter-Item correlations to within sampling error. That is to say, rxy - rxTryT should be zero to within sampling error. By the same token, if several scales all measure the same underlying trait, then the correlations between the scales will also satisfy the same "product rule for Internal consistency“ and likewise we can see if the product rule reproduces the Interscale correlations to within sampling error. Let us now look at the path model and examine the differences between onsengd_ggLLeiatlgns and Lennguggg_ggLLejatigns between these six shame theme scales. If the underlying trait is unldimensionai, these differences (l.e., errors) should not differ from zero by more than sampling error. We can see in Table 4 that the errors (or differences) are very close to zero. A chi-square test shows that the differences are not 73 significantly different from 0. That is, the path model has a good fit. The correlations between the six shame themes satisfy the internal consistency requirement for a general shame hierarchical measurement model. Table 4 WWW Shame_lnemes W 08 I D W U FS QRS:L1§Q.§QLL§iflllQn§ Disappointment with Self 1.00 .93 .95 .91 .87 .85 inferiority .93 1.00 .92 .90 .89 .77 Defectiveness .95 .92 1.00 .86 .83 .84 Worthlessness .91 .90 .86 1.00 .94 .68 Unimportance .87 .89 .83 .94 1.00 .70 Failing Short of Standards .85 .77 .84 .68 .70 1.00 Winn: Disappointment with Self 1.00 .95 .95 .92 .90 .80 Inferiorlty .95 1.00 .92 .89 .87 .78 Defectiveness .95 .92 1.00 .89 .87 .78 Worthlessness .92 .89 .89 1.00 .85 .75 Unimportance .90 .87 .87 .85 1.00 .74 Failing Short of Standards .80 .78 .78 .75 .74 1.00 ELLQLS Disappointment with Self -.02 .00 -.01 -.03 .05 inferiority -.02 .00 .01 .02 —.01 Defectiveness .00 .00 -.03 -.04 .06 Worthlessness -.01 .01 -.03 .09 -.07 Unimportance -.03 .02 -.04 .09 -.04 Falling Short of Standards .05 -.01 .06 -.07 —.04 2. Parallelism In order to test the parallelism of the six shame themes, two methods were employed: 74 3) Correlations between six shame themes and outside variables were calculated and are shown In Table B-4 in the Appendix. As can be seen from Table B-4 these six shame themes are parallel In terms of their correlations with outside variables; b) Parallelism was computed by using the similarity coefficient formula (Hunter & Gerbing, 1982). The similarity coefficients are shown In Table 5. They are all close to 1.00, that Is, perfect parallelism. Table 5 5imiiaLi11_QQn1IiCififll§_QI_§ll_§ham§_IthQ§ 2 3 4 5 6 1. Disappointment with Self 1.00 1.00 .99 .99 .99 2. Inferiorlty 1.00 .99 1.00 .99 3. Defectiveness .99 .99 .99 4. Worthlessness 1.00 .98 5. Unimportance .98 6. Falling Short 0. Conclusion The results show clearly that each of the six shame theme scales Is unldimensionai. However, both the pattern of high correlations between shame theme scales and the exploratory factor analysis show that the six scales are not Independent of each other. Rather those analyses suggest that the six shame theme scales measure a general shame construct and each shame theme scale construct differs from that general factor by 75 only a small and trivial specific factor. To check this suggestion, a hierarchical measurement model was tested. First, the six shame theme clusters form a correlation matrix that is internally consistent. Second, the shame theme clusters are externally consistent In their correlations with outside variables. it is, therefore, concluded that the general shame scale measures only one useful construct. The different shame themes do not refer to different emotions or different dimensions of shame. ll. Shame_£Loiiie A. Four Shame Profile Scales-~Deveiopment and Testing 32 items were written to tap the shame profile: competence shame, relationship shame, character shame, and body shame. These shame profile clusters were tested using confirmatory factor analysis. This revealed a problem with the character shame scale. it had two subciusters. Four character shame items (l.e., "i feel defective as a person, as if something is basically wrong with me"; "i feel there Is something defective in my character"; ”I see myself as Intact and without personal defects”; "I am down on myself for any character flaws In me") were too general. The other six items deal specifically with dependency issues (e.g., “I feel ashamed for not being more independent“; "i feel ashamed for being too dependent on others"). Therefore, the four nonflttlng items were dropped and character shame cluster was renamed dependency shame. All further analyses were done on the 28 remaining items. 76 1. Conflrmatory Factor Analysis Conflrmatory factor analysis result is shown In Table B—5 In the Appendix. The confirmatory factor analysis shows that the four shame profile clusters hold up very well; the Items within each cluster appear to be equivalent to each other. 2. Exploratory Factor Analysis Once the four too-general "character" shame items were dropped, the character shame cluster was renamed dependency shame. Then all the shame profile items were submitted for an exploratory factor analysis as a check on the confirmatory factor analysis results. A four-factor solution was obtained when the eigenvalue was set at .50. The four VARIMAX factors suggest clusters that exactly match the a priorl clusters. That is, when the Items were blindly grouped together using the highest loading from the four varimax factors, the clusters thus formed were exactly the shame profile clusters. The four-factor solution is Included in Table 8-6 in the Appendix. 3. Parallelism Items in each of shame profile clusters were examined In terms of how they correlate with outside variables. These correlations are listed In Table B-7 in the Appendix. This table shows that items within each cluster are parallel to each other In terms of their correlations with outside variables. B. Hierarchical Measurement Model Table 6 shows the means, standard deviations and reliabilities of the four shame profile scales. These scales all have high reliability as measured by coefficient alpha (ranging from .80 to .91). 77 Relationship shame and dependency shame scales have only six Items each but they both have respectable reliability (.84 and .80). Table 6 Maansa_StandaLd_Da1Ja119ns_and_Bei1abi1111es.o£.§hame_EL21iLe Sfiflififi Number Standard of Items Mean Deviation Reliability Competence shame 8 17.93 5.38 .89 Relationship shame 6 11.46 3.90 .84 Dependency shame 6 12.28 3.99 .80 Body shame 8 20.91 6.62 .91 The lntercorreiatlons between the four shame profile factors are shown in Table 7. The correlations are all positive and show the Spearman pattern for one common factor. Therefore, the hypothesized hierarchical measurement model was tested. Table 7 lfl1flLQQLLfllaIiQfl3_91.5flflmfl_ELQIllfl_EifilflLfi 2 3 4 1. Competence Shame .82 .71 .67 2. Relationship Shame .62 .58 3. Dependency Shame .38 4. Body Shame 78 Vulnerability theory makes two predictions: First, the shame profile scales will fit a hierarchical measurement model. Second, the general factor for that model will be the construct for the general shame theme scales, that is, general shame. if both hypotheses are true, then there is no need to use confirmatory factor analysis to define a general factor for the shame profile. instead, the general shame factor measured by the shame theme Items would act in that role. That Is, the general factor for the shame profile would be Independently measured rather than computed by factor analysis. The Internal consistency test for the hierarchical model would be tested using ordinary path analysis, a path model with the measured variable general shame as the causal determinant of the four shame profile scales. That path analysis was carried out. The path diagram predicted by vulnerability theory Is presented in Figure 7. It Is also possible to test the two hypotheses separately. This is done with a two cluster confirmatory factor analysis with one cluster defined by the shame profile cluster and a second cluster defined by the general shame scale. According to the first hypothesis, the confirmatory factor model should show perfect fit for the shame profile cluster (fit for the general shame cluster was already tested). According to the second hypothesis, the correlation between general shame factor and the shame profile general factor should be 1.00. That would then show that the factor defined by the confirmatory factor analysis Is In fact identical to the general shame factor. This analysis was also carried out. 79 1. internal Consistency a. Path Model A hierarchical measurement model was constructed for the causal relationships between the shame profile factors. if the general shame factor measured by the shame theme scales Is substituted for the general factor of the profile scales, then the measurement model becomes the simple path diagram as shown in Figure 7. The correlations between the general shame and shame profile are shown. A separate higher order confirmatory factor analysis (data not included) shows that the higher order factor correlates 1.00 with general shame. in other words, general shame is the higher order factor for the shame profile. General Shame .9 / .69 .68 Competence Relationship Dependency Body Shame ‘ Shame (CS) Shame (RS) Shame (08) (BS) Eiguce Z. The hierarchical measurement model of general shame and shame profile. b. Fit of the Path Model Let us look at the path model and examine the differences between QQEQLIQQ_£QLL§lflliQnS and LfinLanfifid_£QLLfilallQn§ between the shame profile. if the underlying trait Is unldimensionai, these differences (l.e., errors) should not differ from zero by more than sampling error. The computations for testing this model are shown in Table 8 and we can see that the errors are very close to zero. A chi-square test shows that the differences are not significantly different from 0. That Is, 80 the path model has a good fit. The correlations between the shame profile factors are internally consistent with a general shame hierarchical measurement model. Table 8 ELLQL_Ana1xsis_19L_1ha_Mode1_in_£isuLa_1 Shame_£Loliie Shame_ELoiiie CS RS 08 BS QbSfiL1£d_£QLL§ialionS Competence Shame 1.00 .82 .71 .67 Relationship Shame .82 1.00 .62 .58 Dependency Shame .71 .62 1.00 .38 Body Shame .67 .58 .38 1.00 BQDLQQUQQQ_§QLL§laliQn§ Competence Shame 1.00 .86 .66 .65 Relationship Shame .86 1.00 .62 .61 Dependency Shame .66 .62 1.00 .47 Body Shame .65 .61 .47 1.00 .ELLQLS Competence Shame -.04 .05 .02 Relationship Shame -.04 .00 -.03 Dependency Shame .05 .00 -.09 ,Body Shame .02 —.03 -.09 2. Parallelism Vulnerability theory predicts that the shame profile correlations will satisfy the parallelism condition for a hierarchical measurement model. According to vulnerability theory, the only reason that a shame profile scale correlates with any other variable Is that the shame profile scale Is correlated with general shame. If general shame were held constant, there WOUId b6 no correlation between any shame profile 81 scale and any other personality variable. Two tests of the parallelism prediction were done: a) Correlations between the shame profile and outside variables were calculated and are shown in Table B-8 in the Appendix. As can be seen from the table the shame profile factors are parallel In terms of their correlations with outside variables. b) The general shame factor was partlaled out of the correlations between the shame profile and outside variables. The result is shown In Table 9. Table 9 shows that the specific factors in the shame profile are not correlated with most of the outside variables once the general shame factor is partlaled out. Out of 104 partial correlations only 11 have absolute values equal to or greater than .29. Once the general shame factor is partlaled out, we do observe that the specific factor in competence shame still correlates to a slight degree with variables having to do with handling of Interpersonal situations. PeOpie suffering from competence shame seem to have some problem with shyness (L - .32), lack of assertiveness (L - .29), trait anxiety (L - .29) and audience anxiety (L - .29); tend not to want to be center of attention (exhibitionism, L - -.34), or be.a leader (dominance, L - —.35). The specific factor in relationship shame correlates quite highly with loneliness (L - .67) and moderately with lack of self-revelation (L - .44). In other words, people with relationship shame (greater than predicted by their level of general shame) suffer from loneliness and intimacy problems. In addition, people with dependency shame have some difficulty with guilt (L - .33) 82 and trait anxiety (L - .30). Table 9 W W Competence Relationship Dependency Body Shame Shame Shame Shame Embarrassment .27 -.08 .07 .19 Fear of Exposure .18 .05 .11 .06 Public Self-Consciousness .03 —.17 .10 .18 Shyness .32 .09 .08 .05 Guilt (1) .10 -.11 .27 .14 Guilt (2) .04 -.11 .33 .09 Self-Esteem (1) —.03 .33 .20 —.10 Self-Esteem (2) -.13 .11 .09 -.11 Depression -.15 .06 .06 -.08 Trait Anxiety .29 .19 .30 -.13 Trait Anger -.02 .13 .13 .00 Loneliness -.06 .67 .07 -.20 Assertiveness -.29 .06 -.12 -.12 Satisfaction with Life -.02 -.05 .12 -.04 Resentment -.03 .03 .17 .05 Conformity -.12 -.11 .23 -.01 Egocentrism .06 -.13 .13 .21 Exhibitionism -.34 -.10 .02 -.01 Dominance -.35 .04 .01 -.06 Self—Revelation .17 -.44 .09 .03 Competitiveness -.20 .20 .03 -.11 Need for Approval .19 -.26 .26 -.03 Audience Anxiety .29 -.19 .05 .05 Note: Guilt (1) is measured by Buss and Durkee Guilt Scale; Guilt (2) is measured by Chang and Hunter Guilt Scale. Self—esteem (1) is measured by Rosenberg Self-Esteem Scale; Self-esteem (2) Is measured by Cheek and Buss Self-Esteem Scale. C. Conclusion When a higher order factor analysis was performed on the shame profile scales, the higher order factor correlates 1.00 with the general shame factor of the shame theme scales. That Is to say, the higher order factor for the shame profile scales is general shame. When tested for Internal consistency, the shame profile scales are consistent with a 83 unldimensionai construct model. There are specific factors In the shame profile. The specific factors for competence shame and relationship shame are very small In magnitude while the specific factors for dependency shame and body shame are much larger. The test for parallelism shows that the specific factors for body shame and dependency are not significantly correlated with any of the outside measures of social and emotional functioning. Thus the specific factors for dependency shame and body shame are large but trivial. The test for parallelism revealed a large number of correlations of about .30 between competence shame and the measures of social and emotional functioning. However, since the correlation between competence shame and general shame is .95, these small relationships would be Invisible to most analyses without massive sample size. Thus competence shame does appear to have a not-completeiy-trivlal specific factor but It is of no clinical significance. Finally, we come to the specific factor for relationship shame. it Is small in magnitude since reiatlonahip shame Is correlated .90 with general shame. However, that specific factor has two large correlations with two outside variables: .67 with loneliness and -.44 with self- revelatlon. This means that those who develop relationship shame to a greater extent than would be predicted from their level of general shame also have a problem forming intimate relationships. Ill. Shame.and.§eii:£staam The correlations between shame and self-esteem can be seen In Table 10. Shame is measured by two scales (Chang & Hunter, 1988; Cook, 1985) 84 and self-esteem is measured by two widely used scales (Rosenberg, 1965; Cheek & Buss, 1981). Table 10 shows that shame and self-esteem are very highly correlated (with an average correlation of -.92). They are not only very highly correlated with each other, but they also form a correlation matrix that Is internally consistent from the point of view of a one factor measurement model. Table 10 CoLLeiations_betwaen_Shama_and_Sail:£steem Cook Shame Rosenberg Cheek & Buss Self-Esteem Self-Esteem General Shame .99 -.95 -.94 Cook Shame -.90 -.90 Rosenberg Self-Esteem 1.00 Hierarchical Measurement Model 1. internal Consistency We can use a more rigorous test for internal consistency by computing the differences between the observed correlations and the reproduced correlations. The differences (or errors) should not differ from zero by more than sampling error. First, a higher order confirmatory factor analysis produces the following correlations as shown in Figure 8. 85 —. Higher Order Factor Shame & Self-Esteem .99 .95 —.98 -.97 General Shame Cook Shame Rosenberg Cheek & Buss } (GS) (CS) Self-Esteem Self-Esteem i J (ass) (CBSE) l EiguLe_§, The hierarchical measurement model for shame and self-esteem. The computations for testing such a model is provided In Table 11. The errors are within sampling error of 0. in other words, shame and self-esteem scales form a correlation matrix that Is Internally consistent for a one factor model. Table 11 Shame.and.§eii:£steem ‘ Shamn_ann_§nilzfisleam GS CS RSE CBSE QDSlend_QQLLaia1lQn§ General Shame 1.00 .99 -.95 -.94 Cook Shame .99 1.00 -.90 -.90 Rosenberg Self—Esteem -.95 -.90 1.00 1.00 Cheek a Buss Self-Esteem -.94 -.90 1.00 1.00 BenLQduced_CoLLe1a119ns General Shame 1.00 .94 -.97 -.96 Cook Shame .94 1.00 -.93 -.92 Rosenberg Self-Esteem -.97 -.93 1.00 .95 Cheek & Buss Self-Esteem -.96 -.92 .95 1.00 ELLQLS General Shame .05 .02 .02 Cook Shame .05 .03 .02 Rosenberg Self-Esteem .02 .03 .05 Cheek & Buss Self-Esteem .02 .02 .05 86 2. Parallelism Parallelism of shame and self-esteem was assessed by examining how they correlate with outside variables. By prediction, if shame and self-esteem are equivalent constructs they should correlate with outside variables In a parallel manner. Shame and self-esteem are parallel in their correlations with outside variables, as can be seen In Table 12. The parallelism was further assessed In terms of similarity coefficients by using Hunter's PACKAGE and Is shown in Table 13. Table 12 and 13 show very clearly that shame and self-esteem are parallel with each other in terms of their correlations with outside variables. Table 12 CQLLela1ions_n1_Shame_and_Sa11:Es1eem_witn_0ulside_xaLiabics General Cook Rosenberg Cheek & Buss Shame Shame Self-Esteem Self-Esteem Depression .64 .61 -.66 -.71 Anxiety .86 .86 -.80 -.84 Anger .17 .23 -.10 -.14 Loneliness .77 .81 —.69 -.72 Assertiveness -.59 —.61 .61 .62 Life Satisfaction -.73 —.72 .77 .82 Audience Anxiety .35 .37 —.38 -.40 Embarrassment .58 .60 -.57 -.61 Fear of Exposure .65 .67 -.62 -.66 Shyness .62 .65 -.64 -.67 Public S-C .17 .23 -.10 -.16 Resentment .46 .49 -.42 -.51 Conformity .42 .44 -.33 —.40 Egocentrism .50 .54 -.47 -.53 Exhibitionism -.41 -.42 .49 .48 Dominance -.43 —.44 .51 .50 Self-Revelation -.40 -.39 .42 .43 Competitiveness .02 .02 .03 —.02 Need for Approval .53 .52 -.59 -.57 87 Table 13 Sim1JaLi1x_C9e11Js1an1s_be1wean_Shame_and_Sei1:Esteem_Sca1es Cook Rosenberg Cheek s Buss Shame Self-Esteem Self-Esteem General Shame 1.00 -1.00 -1.00 Cook shame -.99 -1.00 Rosenberg Self-Esteem 1.00 Conclusion it has been shown that shame and self—esteem are identical constructs as measured by two shame scales (Chang & Hunter, 1988; Cook, 1985) and two self-esteem scales (Rosenberg, 1965; Cheek s Buss, 1981). They are not only almost perfectly correlated, but also parallel in their correlations with outside variables. Thus they are unldimensionai by both the test for internal consistency and the test for external consistency (or parallelism). it is, therefore, proposed that shame and self-esteem are the same dimension. That is, there Is only one dimension of shame and self-esteem. iV. cILnIca1_Concents_ae1a1ed_to_sname This section will consider the relationship between shame and the related constructs: embarrassment, fear of exposure, self- consciousness, shyness, and guilt. For this study "self-consciousness" is defined by the Fenlgstein, Scheler, and Buss (1975) Public 88 Self—Consciousness Scale. There have been four positions taken on this Issue in the literature. First, some have argued that one or more of the related constructs Is a part of the shame process Itself. These authors would expect the related constructs to be perfectly correlated with shame and hence with each other. This will be called the “Identity position." Helen Block Lewis Is one of these authors with the exception that she differentiates shame from guilt. Second, some (e g., Tomklns, 1963, 1987; Kaufman, 1985) have argued that the related constructs are variants of the primary shame affect. These authors would permit the related constructs to have specific factors and hence not be perfectly correlated with either shame or each other. But since the specific factors are trivial in nature, they should not correlate with each other or with the outside variables. Thus these authors would predict that the related constructs would satisfy a hierarchical measurement model with shame as the general factor. This will be called the "variants position." Third, some authors would argue that there is one general shame complex that subsumes the related constructs. But they would argue that the general shame scale might also be subsumed under that shame complex factor rather than be Identical to that factor. These authors would predict that a hierarchical measurement model would fit the variable cluster formed by adding shame to the list of related constructs. This will be called the "shame complex position." it should be pointed out that Buss (1980) used the term "social anxiety" as the name for the higher order factor. He did not use the term “shame complex." But It seems that his Ideas fit this third position. Fourth, there are those who would simply argue that constructs such as shyness, 89 shame, and guilt measure causally different variables. Some would also argue that embarrassment and shame are distinct as well (Novak, 1986). These authors would predict that a hierarchical measurement model would not fit the data. This will be called the "distinct traits position." The “distinct traits position" will not be tested as such. Rather it is the conclusion implied by the failure of the other three positions. The "Identity position" predicts that the related constructs will each be perfectly correlated with shame and hence perfectly correlated with each other. This hypothesis falls by a wide margin as shown in Table 14. The highest correlation between shame and one of the related constructs Is the correlation of .65 with fear of exposure. The correlation between shame and public self-consciousness Is only .18. Embarrassment and shyness correlate .58 and .62 with shame respectively. Guilt correlates .52 with shame. The "identity position" Is thus rejected. Table 14 2 3 4 5 6 1. Shame .58 .65 .18 .62 .52 2. Embarrassment .87 .46 .74 .61 3. Fear of Exposure .48 .82 .65 4. Public Self-Consciousness .24 .47 5. Shyness .46 6. Guilt Note; Guilt is measured by Buss & Durkee Guilt Scale and Chang & Hunter Guilt Scale combined as one scale. 90 n r u Some authors have postulated that shame is the higher order factor for these other related constructs. To test this “variants position," a hierarchical measurement model with general shame as the higher order factor was constructed. Figure 9 shows the path model for the test for internal consistency. Genera l Shame , \ .65 .6 .58 .52 17 Fear of Shyness Embarrass- Guilt Public Exposure ment -——— S-C ELQULQ_9. A hierarchical model with general shame as higher order factor: the test for Internal consistency. The computations for testing the above model are shown in Table 15. Table 15 shows that the path model did not fit the data as the errors between observed and reproduced correlations are far from zero. That is, embarrassment, fear of exposure, shyness, guilt, and public self- consclousness do not satisfy the internal consistency requirement for a hierarchical measurement model with general shame as the higher order factor. in other words, embarrassment, fear of exposure, shyness, guilt, and public self—consciousness are much more highly correlated with each other than would be predicted from their correlations with shame. Thus each measure differs from shame by something more than a trivial specific factor. 91 Table 15 ELLoL_Ana11s1s_19L_1ns_Modei_in_EinuLe_9 MaLiabLes laLLabLes FE S E 6 P80 QDSQL1§Q_£QLLnialiQn§ Fear of Exposure 1.00 .82 .87 .65 .48 Shyness .82 1.00 .74 .46 .24 Embarrassment .87 .74 1.00 .61 .46 Guilt .65 .46 .61 1.00 ,47 Public Self-Consciousness .48 .24 .46 .47 1.00 BQDLQQUQQQ_QQLLelaliQnS Fear of Exposure 1.00 .40 .38 .34 .12 Shyness .40 1.00 .36 .32 .11 Embarrassment .38 .36 1.00 .30 .10 Guilt .34 .32 .30 1.00 .09 Public Self-Consciousness .12 .11 .10 .09 1.00 ELLQLS . Fear of Exposure .42 .49 .31 .36 Shyness .42 .38 .14 .13 Embarrassment .49 .38 .31 .36 Guilt .31 .14 .31 .38 Public Self-Consciousness .36 .13 .36 .38 Table 16 shows that the model did not satisfy the test for external consistency either. if general shame were the higher order factor for these related traits, we would predict that the partial correlations between specific factors of these traits and the outside variables would not differ from zero by more than sampling error. Table 16 shows that 37 out of 72 partial correlations in absolute values are greater than .30, and 12 partlal correlations In absolute values are equal to or greater than .50. For example, embarrassment still correlates negatively with assertiveness (L - -.50), exhibitionism (L - -.35), and 92 dominance (L - —.35); but positively with blushing (L - .51), audience anxiety (L - .52), resentment (L - .39), conformity (L - .43), egocentrism (L - .59), and need for approval (L - .45). Fear of exposure correlates negatively with assertiveness (L - -.52), exhibitionism (L - -.31), and self-revelation (L - -.51). in addition, public self-consciousness, shyness, and guilt also still correlate significantly with many outside variables when general shame is partlaled out. In conclusion, the "variants position" Is rejected as the hypothetical hierarchical measurement model did not satisfy the tests for internal consistency and external consistency. Table 16 EaL11a1_CoLLeia1iQns_be1wean_EmbaLLassmentz_EeaL_21_ExnosuLez.EubLLc ; —0|‘ 0 ‘g; ‘ I g- ‘ ago ..00 ‘ o- . .0 s . g GeneLa1_Shame_£aL11aied_Qu1 Embarrass- Fear of Public Shyness Guilt ment Exposure S-C Depression -.04 .03 -.14 .07 -.04 Anxiety .18 .20 .14 .14 .21 Anger .17 .16 .11 -.04 .20 Loneliness .05 .28 -.11 .46 -.10 Assertiveness -.50 -.52 -.19 -.76 —.23 Life Satisfaction -.05 -.24 —.05 -.18 .05 Resentment .39 .47 .33 .29 .31 Conformity .43 .55 .46 .28 .35 Egocentrism .59 .56 .71 .26 .47 Exhibitionism -.35 -.31 .10 -.59 -.05 Dominance -.35 -.29 .01 -.58 .05 Self-Revelation -.18 —.51 .02 -.32 -.06 Competitiveness .02 .10 .03 .03 .01 Need for Approval .45 .43 .33 .40 .24 Audience Anxiety .52 .35 .22 .49 .21 93 In the previous section, It was found that shame is not the higher order factor for embarrassment, fear of exposure, public self- consciousness, shyness, and guilt. The next question is: Is it possible that shame, embarrassment, fear of exposure, public self-consciousness, shyness, and guilt can be subsumed under a higher order factor? in other words, is there a hierarchical measurement model that subsumes all these traits, including shame? The correlations between the six shame-related constructs are shown In Table 14. This correlation matrix was used to test the hierarchical measurement model for internal consistency. The results are presented in Appendix B-9. The fit was good for all variables except public self- consclousness. For public self—consciousness, there were major departures from unidimensionallty. Therefore, further testing of the hierarchical measurement model excluded public self-consciousness from the model. Figure 10 shows the hypothetical hierarchical measurement model for the five shame-related traits (excluding public self- consciousness). Higher Order Factor .97 .89 .82 71 .66 Fear of Embarrass- Shyness Shame Guilt Exposure (FE) ment (E) (SHY) (S) (G) EiguLe_1Q. A hypothetical hierarchical model subsuming shame, embarrassment, fear of exposure, shyness, and guilt: the test for Internal consistency. 94 Computations for testing the internal consistency of the model in Figure 10 are shown in Table 17. We can see in Table 17 that all the errors are very close to zero. That is, embarrassment, fear of exposure, shyness, shame, and guilt satisfy the internal consistency requirement for a hierarchical measurement model. Table 17 ELLQL_Ana1xsis_19L_1ha_Modei_in_£inuLe_1Q Variables MaLiabies FE E SHY S G QbsnL1£fi_QQLLfiia1iQnS Fear of Exposure 1.00 .87 .82 .65 .65 Embarrassment .87 1.00 .74 .58 .61 Shyness .82 .74 1.00 .62 .46 Shame .65 .58 .62 1.00 .52 Guilt .65 .61 .46 .52 1.00 BQDLQQUQQQ_QQLLQLaLLQnS Fear of Exposure 1.00 .86 .80 .69 .64 Embarrassment .86 1.00 .73 .63 .59 Shyness .80 .73 1.00 .58 .54 Shame .69 .63 .58 1.00 .47 Guilt .64 .59 .54 .47 1.00 ELLQLS Fear of Exposure .01 .02 -.04 .01 Embarrassment .01 .01 —.05 .02 Shyness .02 .01 .04 -.08 Shame -.04 -.05 .04 .05 Guilt .01 .02 —.08 .05 Hunter and Gerbing (1982) have pointed out that the test for Internal consistency Is a weak test for unidimensionallty. We must also test the model for external consistency before making any conclusion regarding whether a hierarchical measurement model fits the data. Thus 95 a test for parallelism was performed. The higher order factor was partlaled out of the correlations between embarrassment, fear of exposure, shyness, shame, and guilt and outside variables. if the hierarchical measurement model were to fit, then the partial correlations with outside variables should not differ from zero by more than sampling error. Table 18 shows these partial correlations. Table 18 e a . e.‘ a; ,;;| ..n; -. e _00‘ ;. “a. . u; I a; on! . I 0 QLQQL_EACLQL_EaLllai§Q_Qu1) Shame Fear of Embarrass— Shyness Guilt Exposure ment Public Self— -.23 .19 .15 -.26 .25 Consciousness Depression .44 -.40 -.33 .01 -.07 Anxiety .69 -.78 -.42 -.15 .00 Anger .00 -.01 .05 -.21 .14 Loneliness .52 -.40 —.54 .29 -.24 Assertiveness -.03 .60 .19 -.54 .12 Life Satisfaction -.50 .22 .39 -.02 .17 Resentment .01 -.05 —.07 -.08 .12 Conformity —.08 .14 -.03 -.15 .16 Egocentrism -.05 -.12 .15 -.29 .22 Exhibitionism .00 .35 .02 -.44 .21 Dominance -.05 .33 —.01 -.45 .29 Self-Revelation -.05 -.51 .25 -.11 .15 Competitiveness -.01 .20 -.01 .01 -.06 Need for Approval .07 -.28 .00 .07 -.03 Audience Anxiety —.11 -.47 .20 .21 -.02 Table 18 shows that the hypothetical hierarchical measurement model does not fit the data. Many partlal correlations are far from zero. There are 7 partial correlations (In absolute value) greater than .50. There are 14 more partial correlations (In absolute value) in the range of .30 to .50. Furthermore, the specific factors do not make much 96 substantive sense as many partial correlations are simply not lnterpretabie. in particular, some partial correlations between fear of exposure and embarrassment and outside variables are meaningless. For example, the partial correlations between fear of exposure and some outside variables are as follows: -.40 with depression, -.78 with anxiety, and .60 with assertiveness. The partial correlations between embarrassment and some outside variables are as follows: -.33 with depression, -.42 with anxiety, and -.54 with loneliness. In other words, even though a higher order factor was created mathematically, the results were not substantively meaningful. Summarizing results from Tables 17 and 18, we can see that the correlations between shame, fear of exposure, embarrassment, shyness, and guilt are internally consistent with a hypothetical hierarchical measurement model; but when the higher order factor Is partlaled out many partial correlations with outside variables are different from zero by more than sampling error. That Is to say, these traits are not parallel In terms of their correlations with outside variables. It is, therefore, concluded that there is no hierarchical measurement model that subsumes shame, fear of exposure, embarrassment, shyness, and guilt. Thus the "shame complex position" is rejected. Three positions In regard to the shame complex have been tested. First, the “identity position" was tested. The analysis showed that embarrassment, fear of exposure, public self-consciousness, shyness, and guilt are not Identical to shame. Second, the "variants position" was tested. This analysis showed that shame Is not the higher order factor 97 for embarrassment, fear of exposure, public self—consciousness, shyness, and guilt. Third, the "shame complex position" was tested. This analysis showed that there Is no higher order factor that subsumes all the shame-related traits Including shame. In other words, the "Identity position," the "variants position,” and the "shame complex position" are all rejected. The "distinct traits position" is, therefore, accepted by default. What this means is that shame ls causally differentiated from embarrassment, fear of exposure, public self-consciousness, shyness, and guilt. V. Shame_and_GuLLt The key question for this section Is the relationship between shame, guilt, and depression. Are both shame and guilt correlated with depression? Are both causal determinants of depression? These same ‘ questions can be asked for any measure of emotional adjustment. Therefore, the analysis was done for all eight measures of emotional adjustment Included In the study. The relevant correlations are presented In Table 19. Table 19 presents first the simple or zero order correlations between shame and guilt with each of the 8 measures of emotional adjustment. Shame is highly correlated with 6 of the 8 measures, all except audience anxiety and anger. Guilt is more moderately correlated with all but the same two variables. Guilt is more highly correlated with audience anxiety and anger than Is shame. in particular, shame correlates .64 with depression and guilt correlates .31 with depression. 98 Table 19 o a - on‘ o I-ll; -|- . I l0 on. a0 llél ' 00 3i ° I Shame Guilt Shame* Guiit** zero zero partlal partlal; Depression .64 .31 .59 -.03 Anxiety .86 .52 .80 .17 Loneliness .77 .35 .73 -.08 Life Satisfaction —.73 -.34 -.45 .06 Assertiveness -.59 -.49 -.69 -.26 Resentment .46 .49 .28 .34 Audience Anxiety .36 .38 .21 .25 Anger .17 .26 .05 .20 * Guilt partlaled out ** Shame partlaled out Table 19 also presents the partial correlations needed to check for indirect (or "spurious") correlation. The column for shame partial correlations shows the correlations for shame with guilt held constant. If shame were only Indirectly linked to a variable through guilt, the corresponding partial correlation would be 0. The only partial correlation that Is 0 Is that for anger. But the zero order correlation between shame and anger was only .17 to begin with. Thus shame Is not Indirectly linked to any of the other variables by guilt. The column for guilt partlal correlations shows the correlations for guilt with shame held constant. if guiit were only indirectly 99 linked to a variable through shame, the corresponding partial correlation would be 0. The guilt partial correlations for depression, loneliness, and life satisfaction are in fact 0. Thus guilt is not a causal determinant of any of these three variables. The guilt partial correlation for anxiety drops from .52 to .17. Thus gullt also makes little or no contribution to the determination of anxiety. 0n the other hand, the guilt partials are nowhere near 0 for assertiveness, resentment, audience anxiety, or anger. Each of these variables ls associated with responses to people and are thus more social in nature than the first four variables. it would appear that while guilt is only Indirectly connected to feelings of emotional well- being, it may be directly connected to measures of social interaction. In particular, consider the results for depression. The partial correlation for shame (with guilt partlaled out) is .59, only slightly smaller than the zero order correlation of .64. Thus the link between shame and depression does not depend on guilt. The partial correlation for guilt (with shame partlaled out) is -.03, virtually 0. Thus guilt appears to be indirectly linked to depression through its positive correlation with shame. Vl- GQflflflL_QlI1QLflflQQfi_lD_5hiflfl.lflfi.§flll1.fi£9flflflfi§i Table 20 shows the means and standard deviations of college males and females on shame, self-esteem, and guilt scales. Table 21 shows the correlations between gender and these scales. 100 Table 20 ,,-._n ..o .... o I; . e. e e as: U. ;‘ ..e 'u. ;‘ e. l- Maies (N-249) Females (N-677) Number Scales of items M SD M SD 1. General Shame 22 45.83 13.09 46.63 13.85 2. Competence Shame 8 17.33 5.17 18.15 5.44 3. Relationship Shame 6 11.72 3.82 11.37 3.92 4. Dependency Shame 6 12.00 3.71 12.38 4.08 5. Body Shame 8 18.33 5.60 21.86 6.73 6. Cook Shame 11 24.63 6.62 25.21 7.25 7. Rosenberg Self-Esteem 10 38.65 6.46 37.95 7.06 8. Cheek & Buss Self—Esteem 6 23.05 4.02 22.55 4.51 9. Buss a Durkee Guilt 6 17.70 3.86 18.76 3.91 10. Chang & Hunter Guilt 9 27.65 5.57 29.00 6.02 Table 21 CoLLaLa11ons_ba1waen_GendeL_and_Sname._Se11:Esteem._and_fiuilt Gender Gender (with General Scales Shame Partlaled Out) 1. General Shame .02 -- 2. Competence Shame .07 .16 3. Relationship Shame -.04 -.15 4. Dependency Shame .05 .04 5. Body Shame ‘ .24 .31 6. Cook Shame .04 -- 7. Rosenberg Self—Esteem -.04 -- 8. Cheek a Buss Self—Esteem -.05 -- 9. Buss & Durkee Guilt .15 .16 10. Chang & Hunter Guilt .11 .11 Note; For the variable of gender, 1 - Male, 2 - Female. 101 Table 21 shows that there are no gender differences in general shame proneness among college students. College females do report experiencing more body shame and slightly more competence shame and guilt. 0n the other hand, college males report experiencing slightly more relationship shame. DISCUSSION Much of the contemporary literature on shame is directed to the questions of definition. What is shame? Does shame include traits such as embarrassment, fear of exposure, public self—consciousness, shyness, or guilt? These questions are answered by the present research. This research considered the concept of shame at three levels. First, there is the phenomenology of shame. This study produced six themes used by people In describing their shame experiences. Do these shame themes describe different affects or do they describe the same shame affect? Second, there is a literature on the specificity of shame. Some have argued that shame Is first developed toward particular aspects of the self and then generalized to the self as a whole. Others have argued instead that specific aspects of shame differentiate from a general shame vulnerability. Those who are highly vulnerable to shame develop shame toward nearly all aspects of the self while those who are very low in vulnerability do not become ashamed of any aspect of the self. This argument was crystalized In terms of contrasting path models for the "shame profile." These contrasting models were labeled as "conditioning theory" and "vulnerability theory" and were tested In this research. Third, several concepts closely related to shame were considered: embarrassment, fear of exposure, public self-consciousness, shyness, and guilt. Four theoretical positions towards these 102 103 shame-related traits were developed from the literature. These positions were tested In this research. The first section of the discussion will review the findings of the study In regard to the questions: What is shame? What Is not shame? Ih§_§£man1J£§_Ql_§namn It is critical to distinguish between semantic issues and substantive issues in defining shame. it is always possible to expand the word “shame" as broadly as is desired by semantic definition. This is broadness by fiat. However, it Is both a scientific and a logical maxim that a word ceases to be meaningful unless the things denoted by that word are equivalent In some important sense. in psychological measurement, this principle ls registered as the search for equivalence. If two concepts are to be referred to by using the same word, then there should be some Important way In which they are psychologically equivalent. There are two criteria which were used in this study. First, variables are equivalent if they are causally equivalent. Second, variables are equivalent if they are clinically equivalent, that Is, if they relate to all Important clinical variables In the same way. Consider causal equivalence. When are two variables causally equivalent? If we build a causal model that includes the two variables, then the model works best if the two variables are tied together as measures of one underlying construct (Hunter a Gerbing, 1982). This is the psychometric principle of perfect parallelism. Parallelism within a set of equivalent measures is called "Internal consistency" or factor analytic unidimensionallty. Parallelism with respect to variables 104 outside the domain of equivalence Is called "external consistency" or "parallelism." In this sense, embarrassment would only be part of the shame concept If the measure of embarrassment were causally equivalent to the measure of shame. If they are not causally equivalent, then the two concepts are better considered separately. Consider clinical equivalence. Two variables are clinically equivalent if they relate in an Identical manner to all Important clinical variables. For example, If embarrassment Is to be regarded as an aspect of shame, then the correlations between embarrassment and Important clinical variables should exactly match the correlations for shame. If there were a match, then maximal correlations would be obtained by merging the concepts. If there is not a match, then maximal multiple correlations would require that the concepts be kept separate. Mathematically, clinical equivalence Is a special case of causal equivalence. Ine_sname_1nemes How do people describe their shame experiences? A search was made of the clinical literature, clients' verbal reports In psychotherapy sessions, and everyday language. Six shame themes were distilled from this search: disappointment with oneself (Lewis, 1971; Buss, 1980), feelings of Inferiorlty (Adler, 1929, 1931; Alexander, 1938; Miller, 1985; Novak, 1986), feelings of defectiveness (Wurmser, 1981; Morrison, 1983; Miller, 1985; Coen, 1986; Cook, 1985, 1988a, 1988b), feelings of worthlessness (Kaufman, 1985), feelings of unlmportance which are often expressed metaphorically as “feeling small" (Miller, 1985; Cook,.1985, 1988a, 1988b), and feelings of falling short of one's own standards or 105 Ideals (Piers & Singer, 1953; Lynd, 1958; Morrison, 1983; Mollon, 1984). Each shame theme was translated into a scale. The analysis showed that each scale was unldimensionai. Thus each affect defined by a shame theme was successfully measured. The six shame theme clusters are very highly correlated with each other. These shame themes are internally consistent and parallel In terms of their correlations with outside variables (l.e., they are externally consistent). Thus the six shame theme scales satisfy a one factor hierarchical measurement model. That factor ls here called the general shame factor. The specific factors which differentiate the shame theme constructs from the general shame theme factor are both very small in magnitude and trivial In their causal connections to emotional and social functioning. Thus while six distinct phenomenological patterns to the description of shame can be Identified, they all refer to the same underlying affect. Therefore, the combined shame theme scales define the general shame affect. In£_§bamfi_ELQIlln This study has shown that the shame profile clusters form an Internally consistent correlation matrix and that they are parallel In terms of their correlations with outside variables. 'ln other words, they define a unldimensionai hierarchical measurement model. The underlying trait has been shown to be the same general shame factor as for the six shame theme scales. This study further shows that the specific factors In the shame profile become trivial In terms of their correlations with most of the outside variables once the general shame factor Is partlaled out. That Is to say, the general shame factor Is 106 the major causal factor for the significant correlations between the shame profile factors and the outside variables. There are a few exceptions. Once the general shame factor ls partlaled out, we do observe that the specific factor In competence shame still correlates to a slight degree with variables having to do with handling of Interpersonal situations; and the specific factor in relationship shame correlates moderately with loneliness and lack of self-revelation. People suffering from competence shame seem to have some problem with shyness, lack of assertiveness, and audience anxiety; tend not to want to be center of attention, or be a leader. People with relationship shame suffer from loneliness and Intimacy problems. Although some of the specific factors of the shame profile are correlated with some of the emotional and social variables, the overall pattern Is that the specific factors are of little Importance. Thus the shame profile has a higher order factor which is general shame and the specific factors in the shame profile become trivial once the general shame factor ls partlaled out. In other words, It Is the susceptibility to general shame that causes a person to focus on body shame when he or she has been teased or shamed about his or her body or appearances, or causes a person to experience competence shame when he or she has been laughed at for being Incompetent. It Is not the specific types of shame that lead to general shame. In§_§ham§_99mnln1 Some writers have postulated that shame is the higher order factor for other shame-related constructs, namely, embarrassment, fear of BXDOSUI’O, publ lC self-consciousness, shyness, and 91.” It. SUCh a 107 position Is not supported. When a hierarchical model was constructed with general shame as the higher order factor, the path model did not fit as the residuals between observed and reproduced correlations are far from zero. When the effect of general shame was partlaled out, these shame—related traits still correlate with many outside variables. in other words, the specific factors In embarrassment, fear of exposure, public self-consciousness, shyness, and guilt are not trivial once the general shame ls partlaled out. Furthermore, a question was raised regarding the possibility that there might be a hierarchical measurement model that could subsume shame, embarrassment, fear of exposure, public self-consciousness, shyness, and guilt. That hierarchical measurement model did not fit the data. Thus it was shown that shame ls well differentiated from embarrassment, fear of exposure, public self- consclousness, shyness, and guilt. In the literature, some writers emphasize inferiority as the core definition of shame (e.g., Alexander, 1938; Miller, 1985; Novak, 1986); some emphasize embarrassment or exposure (Lynd, 1958; Schneider, 1977; Wurmser, 1981; Kaufman, 1985); some even include feelings of emptiness and fragility (Cook, 1988a, 1988b). Kaufman's definition of shame Includes feelings of exposure and Inferiorlty (or deficiency). Some of these authors have been shown to have extended the label of “shame" to affects that are not psychologically equivalent to shame. In particular, these variables, namely, embarrassment, fear of exposure, emptiness, and fragility relate to important clinical variables very differently from shame. Therefore, It Is very unwise to confuse these CODCODtS VI l th shame . 108 9QQKL§_§1udiea_ilafifiaa_lfififinl Cook (1988a) claimed to have evidence showing the shame concept to be multidimensional. in a more recent study, Cook (1988b) has changed his position and created one "Internallzed shame scale" summing across all the Items In his original inventory. The evidence from the first study on the factor structure of the original Inventory will be examined here. The conclusions of both the first and the second study will then be reconsidered. Cook (1988a) Identified four dimensions In his internalized Shame Scale. They are named: (1) Inadequacy and deficiency, (2) emptiness and loneliness, (3) feelings of exposure and self-criticalness, and (4) feelings of Insignificance and fragility. However, Cook did not start with a theoretically defined measurement model. He used an exploratory factor analysis procedure to cluster his shame Items and then tried to name the factors after the fact. One of the potential problems with exploratory factor analysis Is that It automatically groups Into a factor (or cluster) any items that correlate highly with one another regardless of whether or not they measure the same construct. Indeed the Items need not even have similar content. This blind reliance on size of correlation also makes exploratory factor analysis more vulnerable to sampling error because of capitalization on chance. Let us examine each factor Cook has identified. Items In Factor 1 (Inadequacy and deficiency) share a common meaning: that of Novak's (1986) Inferiorlty factor and the general shame factor Identified In this study. Nine out of ten items In Factor 1 are the same items picked to form the 11-ltem short form Of the ISS for thls study. Thus his 109 Factor 1 does measure shame. Factor 2 Is pure and clear in terms of content meaning: The four items that Cook names "emptiness and loneliness" are "i always feel like there is something missing"; "i have this painful gap within me that i have not been able to fill"; "I feel empty and unfulfilled"; "My loneliness Is more like emptiness.“ However, Novak (1986) in his study left out this entire group as they do not share the same content meaning with shame or inferiority. That Is, Novak rejected Cook's Factor 2 on the basis of inappropriate content. This study presented clear evidence showing that such closely related concepts as embarrassment, fear of exposure, public self-consciousness, shyness, and guilt are distinct from shame. It seems unlikely that a concept as different as emptiness or loneliness would be causally or clinically equivalent to shame. Although the content of these Items is more "emptiness” than "loneliness," this study does have direct data on the relationship between loneliness and shame. The correlation between shame and loneliness Is .77 (corrected for attenuation) which is well less than 1.00. Thus Cook's Factor 2 does not measure shame. Factor 3 is not clear. it is a mixture of four Inferiorlty Items (e.g., “i see myself striving for perfection only to continually fall short"; "I think others are able to see my defects"). two embarrassment items (“I would like to shrink away when I make a mistake“; “When I feel embarrassed, I wish I could go back in time and avoid that event"), one public self-consciousness Item (“I seem always to be either watching myself or watching others watch me“) and one neurotlcism (7) Item ("i become confused when my guilt ls overwhelming because i am not sure why 110 I feel guilty"). Thus only 4 of the 8 items are shame Items. Factor 4 is also mixed. it has three neuroticism (or anxiety) Items ("At times I feel like I will break into a thousand pieces“; "l replay painful events over and over In my mind until i am overwhelmed"; "i feel as if I have lost control over my body functions and my feelings"), one inferiority item ("Sometimes I feel no bigger than a pea"), one embarrassment item (“At times i feel so exposed that i wish the earth would open up and swallow me"), and one residual (l.e., unclassiflable) Item ("i really do not know who I am"). The majority content of Factor 4 Is neuroticism, though it Is not a unldimensionai scale for neuroticism either. To summarize, on the basis of content analysis and the findings in Novak's dissertation, Cook's iSS Items can be regrouped into four clusters: (1) Inadequacy and deficiency (or inferiority), (2) embarrassment, (3) neuroticism (or anxiety), and (4) emptiness. Thus It Is understandable that Cook found 4 factors In his inventory. Cook's factors did not line up perfectly with the four constructs measured because of sampling error. However, this analysis shows that Cook's results do not show that shame Is a multidimensional concept but that Cook's Inventory contained some items that are related to concepts other than shame. Even those authors who Include the concept of embarrassment as part of their semantic definition of shame do not Include the concepts of emptiness and neuroticism. There Is a second point to be made about his subscales. Cook did not correct for attenuation and thus understated the correlations between the four factors. Once we correct his data for attenuation (see 111 Tables 22-23). these factors correlate with each other much higher than what Cook has thought, especially in the cross-validation samples. The fact that most of the correlations for the normal population reach .90 Indicates that the exploratory factor analysis did not correctly assign Items to clusters. Thus the potential distinctions between the four constructs Included in his Inventory were blurred by sampling error. Table 22 Table 23 W W W ‘ WW Factors Factors 1 2 3 4 1 2 3 4 1 _ 82 85 .82 1 _ 84 90 84 2 79 _ 75 84 2 76 _ 87 so 3 83 72 _ .87 3 85 7o __ so 4 75 77 81 _ 4 75 76 79 Note; Correlations below diagonal are from non-clinical samples and correlations above diagonal are from clinical samples. A reexamination of the results In Cook's (1988a) factor analysis shows that his Inventory is multidimensional but that the shame concept Is unldimensionai. The multidimensionallty in Cook's inventory stems from the presence of a scattering of Items that do not measure shame but Instead measure other concepts such as emptiness, embarrassment, and anxiety. Cook's first cluster Is a pure shame scale. it was also these Items which Novak (1986) selected as shame Items on the basis of a 112 content analysis. These Items were used In the present study as the Cook Shame Scale. In his more recent study, Cook (1988b) responded to criticisms of his earlier work by rejecting the hypothesis of multidimensionallty. However, he did not then focus his scale on the pure shame Items. Rather he summed across all the Items in the Inventory. Thus his "internalized Shame Scale" Includes the non-shame items measuring emptiness, embarrassment, and anxiety. It is likely that these Items have little impact on the total score which should be very highly correlated with shame. However, the presence of these Items greatly confuses the literature on the substantive nature of shame by blurring the affective nature of shame. Cook's more recent paper (1988b) Is also confusing in Its relation to self-esteem. In the main body of the results, Cook argues that shame and self-esteem are separate constructs. Yet In several places, Cook recommends adding 7 self-esteem Items to the Internallzed Shame Scale. He does not address the self-contradictory nature of this recommenda- tion. This Issue will be considered further in the later section on the relationship between shame and self-esteem. GendeL_D111eLences_in_§nama_and_fiu111_ELoneness This study found no gender differences in general shame proneness (L - .02). College females do report experiencing more body shame (L - .24, or .31 with general shame partlaled out), slightly more competence shame (L - .16 with general shame partlaled out), and more guilt (L - .15, or .16 with general shame partlaled out). On the other hand, college males report experiencing slightly more relationship shame 113 (L - -.15 with general shame partlaled out). Four other empirical studies that examined gender differences In shame and guilt proneness were located. Table 24 shows the correlations between gender and shame proneness In these studies. in the case of Cook (1988a). only the data for non-clinical samples was used. Furthermore, since the total scale contains many non-shame items, the data for "Scale 1"--the pure shame measure--was used. Table 24 f. 115 .—;.. ‘ e e 2 ; 0" 2; .32. 2.0; ..e ..u; ° egsgs . W Ihs_£lxe_§1unina investigator Date Sample Size Correlation Negri 1978 448 .07 - Mirman 1984 102 -.11 Harder a Lewis 1987 120 .20 Cook 1988a 1296 .19 This study 1988 926 .02 Ine_Mela:Anaixsis Total sample size - 2892 Average sample size - 578.4 Ave(r) - .107 Var(r) - .0077 Var(e) - .0017 Ave(p) - .11 SD(p) - .08 Chi-square - 22.88 with Q; - 4 The weighted average correlation between gender and shame proneness 114 across these studies with a total sample size of 2892 is .11. This would suggest that women are sijgnlly more shame prone than men. However, the results are not homogeneous across studies. The chi-square statistic Is 22.88 (d1 - 4) which Is highly significant. The key differences are those for the present result (L - .02, N - 926) which is .09 below the average and the result from Cook (L - .19, N - 1296) which is .08 above the average. These departures are small by ordinary standards, but the sample sizes are 926 and 1296 respectively and thus both deviations are statistically significant. Furthermore, the two studies differ from each other: L - .02 versus L - .19, a statistically significant difference of .17. Although there appears to be no difference in how the studies were run, there is a large difference In the study results. if both studies are removed, the meta-analysis on the remaining studies yields an average correlation of .07. This Is not far off from the average of the present finding and the Cook finding: L - .105. Thus It Is likely that both studies are off by sampling error. The Cook finding is too high by more than the usual oi- .05 level and the present study finding Is too low by more than the usual c1- .05 level. After all, extreme sampling errors must happen some of the time. That Is, In 1 in 1000 cases, the p - .001 level of sampling error occurs. It may be that very large sampling errors occurred in both studies. If this hypothesis Is correct, then the meta-analysis provides the best estimate: L - .11. By whatever standard, the gender difference on shame proneness Is very small in magnitude. Table 25 shows the correlations between gender and guilt proneness In four studies. A meta-analysis was performed on these results. The 115 weighted average correlation between gender and guilt proneness across these four studies with a total sample size of 1596 is .13. Furthermore, these results are homogeneous across studies. The estimated variance of population correlations Is not positive and the chi—square statistic is 2.76 (d1 - 3) which Is right at the chance level. Thus women are slightly more prone to guilt feelings than men. This is certainly contrary to what Helen Block Lewis would predict. Table 25 WW EQHL_§1ufiiQ§ Iha_£ouL_Sludies Investigator Date Sample Size Correlation Negri 1978 448 .15 Mirman 1984 102 .01 Harder a Lewis 1987 120 .10 This study 1988 926 .13 Ihe_Meta:Anaists Total sample size - 1596 Average sample size - 399 Ave(r) - .126 Var(r) - .00108 Var(e) - .00243 Ave(p) - .13 SD(p) - 0 Chi-square - 2.76 with g1 - 3 She has postulated that "because of their lesser sociability and greater aggression, and their unfair position of superiority in the world of 116 power, [men] are more susceptible to guilt [than women)" (Lewis, 1987a). The meta-analysis across several studies shows just the opposite: Women are slightly more gullt prone than men. To summarize, meta-analyses have shown results consistent with the findings of this study. Women are just slightly more prone to shame than men (L - .11) and also just slightly more prone to gullt than men (L -.13). Women are somewhat higher than men on the body shame specific factor (partial L - .31). This Is not unexpected since women grow up under greater societal pressure to be pretty and to be thin and hence have a greater tendency to be dissatisfied with their weight and looks (see Silbersteln, StriegeI-Moore, & Rodin, 1987). Men are slightly higher than women on the relationship shame specific factor (partlal L - -.15). No previous studies have examined the gender differences on relationship shame. The currently popular thinking Is that women are more socialized around relationship building than men and men are, therefore, more deficient In their ability to establish relationships. Another possibility may be that parents tend to handle their girls with kid gloves but tend to treat their boys with harsher measures. Boys are socialized as If their feelings do not get hurt as easily as girls and hence they may feel less lovable. Again, It should be pointed out that the correlation is very small In magnitude (L - -.15). Shame_and_§eiizfisteem The present study shows that shame and self-esteem are the same construct. They form two ends of the same dimension. That Is, shame Is 117 negative self-esteem. Shame and self-esteem as measured by two shame scales and two self—esteem scales correlate —.92 with each other. These four scales form a very highly correlated and highly Internally consistent correlation matrix; and they are perfectly parallel with each other In terms of their correlations with outside variables. If we examine the semantic meanings of shame and self-esteem items, they all share very similar content meanings. It Is, therefore, concluded that shame and self-esteem are identical constructs. Cook (1988a, 1988b) claims that his data show that shame and self— esteem are distinct constructs. His evidence from both studies will be reexamined here. Consider first the earlier study. in this study Cook (19888) did not correct his correlations for attenuation and thus those correlations are understated. Cook measured self-esteem with the Rosenberg Self- Esteem Scale. Cook did not report the reliability for the Rosenberg scale, but It was used In this study and the coefficient alpha was .90. Table 26 (on next page) presents Cook's correlations corrected for attenuation. Cook's “Scale 1" is the pure measure of shame. it measures inferiority (or Inadequacy and deficiency as Cook names It). That scale correlates -.90 with self-esteem In a non-clinical sample and -.99 In sex offenders sample. Thus corrected for attenuation, Cook's pure shame scale is perfectly correlated with self-esteem. The scales which do not correlate perfectly with self-esteem are those which Include extraneous Items measuring emptiness, embarrassment, or anxiety. 118 Table 26 . . 0 — . _ e ‘. . e.‘ e g ; g. ..e I-II- . , | .o -.e- e .. — Rosenberg Self-Esteem Scale Non-Clinical Sample Sex Offenders Sample iSS Total -.88 -.95 Scale 1 -.90 -.99 Scale 2 -.75 -.94 Scale 3 -.82 -.92 Scale 4 -.72 -.89 in his more recent paper, Cook (1988b) offers a series of multiple regression analyses as evidence that shame and self-esteem are different constructs. Unfortunately, he did not correct for attenuation before doing these analyses and thus the results were spurious. in essence, each analysis took the following form: He created a shame scale summing across his inventory but without any self—esteem Items. He created a self-esteem scale using 7 of the Rosenberg self—esteem Items. He then computed the incremental multiple regression of some dependent variable onto the shame and self-esteem scales. His test for multidimensionallty was to see if adding the second scale would add to the prediction accuracy generated by the first scale. He did both analyses: shame followed by self-esteem, and self-esteem followed by shame. He consistently found that the multiple correlation was larger than either of the two single scale zero order correlations. He then concluded that 119 shame and self-esteem account for different variance In the dependent variable and are therefore different constructs. The error In this reasoning stems from the fact that he did not correct his correlations for attenuation due to error of measurement. Thus the increase In the multiple correlation could be due to a reduction In error of measurement rather than due to capturing some second causal agent. Consider an analogous example. Suppose we start- with a 20 Item vocabulary scale that has reliability of .81 and a correlation of .50 with performance In a psychology class. We then create two 10 Item scales by summing either the even Items or the odd Items. The reliability of each 10 item scale drops from .81 to .68. The correlation with class performance drops from .50 to .46. Consider now Cook's (1988b) multiple regression analysis for multidimensionallty applied to the odd and even vocabulary tests. We first compare the multiple correlation adding the even Item test to the odd Item test. The multiple regression will combine the two subtests In a manner equivalent to adding them back to form the original total test score. Thus the multiple correlation will be .50 while the correlation for the odd item test alone is .46. Since the even item test adds to the prediction of class performance, Cook would then conclude that the even item test measures some construct different from the construct measured by the odd item test. That conclusion would be false. This example shows that Cook's regression analysis lmpiicitiy assumes that the two constructs to be compared are measured without error. Since Cook did not correct for attenuation due to error of measurement, his shame and self-esteem scales were both short of perfect 120 measurement. Thus the multiple correlation could be Increased due to the fact that the combination of the two scales is a more reliable measure of shame than is either scale alone. Cook's hierarchical regression findings are thus not evidence of multidimensionallty between shame and self—esteem as he claimed. Consider all the evidence. This study found shame and self-esteem to be virtually perfectly correlated. Even more important, shame and self-esteem are perfectly parallel in their correlations with outside variables. Thus they are causally and clinically equivalent. The pure Cook shame scale (his ”scale 1” In cook, 19888) also correlated perfectly with self-esteem. Examination of the correlations with outside variables In both Cook's papers (1988a, 1988b) showed that shame and self-esteem were perfectly parallel In this data too. Thus Cook's data Is not Inconsistent with the present findings. Finally, Cook (1988b) did not correct for attenuation due to error of measurement and hence his hierarchical regression analyses could confound an Increase In reliability with multidimensionallty. Thus Cook's findings are all consistent with the hypothesis that shame and self-esteem are the same construct. in short, even though Cook (1988a, 1988b) has argued that shame and self-esteem are distinct constructs, a reanalysis of his data showed that he too found them to be virtually perfectly correlated. Shame_and_au111 This study has shown that shame and guilt are distinct from each other. At the level of zero order correlation, they correlate .52 (corrected for attenuation) with each other. This Is consistent with 121 the correlation of .48 (uncorrected for attenuation) between "shame" and "guilt" scales obtained In the meta-analysis of nine previous studies. Furthermore, from the point of view of a hierarchical measurement model, It has been shown that shame and guilt are not equivalent to each other. This Is consistent with the clinical literature spearheaded by Alexander (1938), Piers and Singer (1953), and Lewis (1971) that has emphasized the distinction between shame and guilt. it appears that the role of guilt as a contributing factor In emotional difficulties has been overrated in the clinical literature. Researchers have been misled by considering correlations for gullt In Isolation rather than in comparison to results for shame. Whenever guilt is correlated with emotional problems without taking shame or self-esteem into account, gullt correlates substantially with emotional difficulties. However, once shame ls held constant, many of the partial correlations for gullt drop to O. The reverse is not true. When guilt is partlaled out, the correlations for shame drop by very little. In predicting feelings of weii-belng--depresslon, anxiety, loneliness, and life satisfaction—-guiit appears to be relevant only because It is an indirect indicator of shame. That is, It appears that trait gullt Is a causal consequence of shame and that It is not a causal antecedent of feelings of emotional well-being. if shame is measured, then guilt will make no further contribution to the understanding of any of these measures. In particular, the present results suggest that guilt will not make an Independent contribution to the understanding of depression. The partial correlations for gullt do not vanish for the emotional 122 measures that are most closely related to social functioning: assertiveness, resentment, audience anxiety, and anger. Thus guilt may be a causal determinant of some of these measures of social functioning. Shamez_Guli12_and_nenLesslon There Is a long history of psychoanalytic theorizing that negative self-esteem Is a key factor In depression (Freud, 1917; Rado, 1928; Fenichel, 1945; Blbrlng, 1953; Jacobson, 1953, 1964, 1971, 1975). This hypothesis predicts a negative correlation between self-esteem and depression. That prediction has been borne out (Harrow & Amdur, 1971; Smith, 1972; Prosen, Clark, Harrow, a Fawcett, 1983). if we equate shame with negative self-esteem, as this study has shown, then that evidence shows a positive correlation between shame and depression. This study found a correlation of .64 between shame and depression. There has also been a long history of theorizing that gullt is a causal determinant of depression. This would Imply that guilt and depression are positively correlated. This too has been found in the literature (Harrow & Amdur, 1971; Prosen, Clark, Harrow, a Fawcett, 1983). This study found a correlation of .31 between guilt and depression. The findings relating guilt to shame for personality In general suggest that while gullt Is correlated with depression at the zero order level, this correlation is spurious (In the language of path analysis). If shame Is held constant, then guilt should not be correlated with depression. At least one group of researchers have suggested that "negative self—esteem" (l.e., shame) rather than guilt forms a cornerstone in depressed patients of all types (Prosen, Clark, Harrow, & 123 Fawcett, 1983). The findings of this study are consistent with their conclusion. Other studies have found results for guilt and shame consistent with the findings of this study. A higher order confirmatory factor analysis was performed on Hobiltzeile's data (Hobiltzelie, 1987) and it was found that guilt as measured by her Adapted Shame/Guilt Scale (ASGS) correlates .20 with depression with shame partlaled out whereas shame (as a higher order factor for ASGS shame, Beall Shame-Guilt Test shame, and Fear of Appearing incompetent) correlates .68 with depression with guilt partlaled out. Another higher order confirmatory factor analysis was performed on Harder and Lewis's data (Harder a Lewis, 1987) and It was found that guilt (as a higher order factor for Personal Feelings Questionnaire guilt and Situational Upset Scale gullt) correlates -.19 with depression with shame partlaled out whereas shame (as a higher order factor for PFO shame, Kaplan's self-derogation, Eagiy's negative self-esteem, and Rosenberg's instability of self-image) correlates .74 with depression with guilt partlaled out. The results of these two higher order factor analyses are Included in Table B-10 and B-11 In the Appendix. in the two studies reanalyzed, the partial correlations between guilt and depression with shame held constant were +.20 and -.19, an average value of 0. The partial correlations for shame with depression with guilt held constant were .68 and .74, an average value of .71. The present study found a partial correlation of -.03 between guilt and depression with shame held constant (Table 19), a value exactly consistent with the other studies. The present study found a partial 124 correlation of .59 between shame and depression with guilt held constant (Table 19), a value not significantly different from the average of the other studies. Imn1ica1Ions_1oL_C11n1saJ_Esxcho1ndx_and_£aLsonaLitx_IneoL1 Four key issues will be discussed with regard to the Implications the findings in this study have for clinical treatment and personality theory. One issue ls related to the causal relationship between general shame and specific shame (l.e., shame profile). Should psychotherapy focus on specific shame or on general shame? The second Issue Is the use of language in psychotherapy and everyday life In regard to experiences related to shame, embarrassment, fear of exposure, public self-consciousness, shyness, and guilt. The third Issue Is whether to focus on guilt or on shame In the treatment of depression. The last issue is the developmental theory of shame. Clip1caI_ImpI1ca1Ions_o1_GenaLa1_Sname_1eLsus_SnecL11c_Shame The hierarchical measurement model relating general shame and the specific shame profile shows that general shame ls causally antecedent to specific shame. In the long run, any changes In general shame will be passed on to all or most of the specific shame dispositions. However, the reverse is not true. In the long run, the specific shame profile components adjust to be consistent with general shame. Thus changes on specific shame will tend to be unstable. What are the Implications for psychotherapy? There are Implications for both diagnosis and treatment. Consider clinical diagnosis. In the Initial interview, the client who suffers from shame often emphasizes shame with respect to some aspect of the self, namely, 125 that specific shame which is most intense or most persistent, say body shame. The therapist should be alert to the fact that people who are high on any one dimension of specific shame are usually high on general shame. Thus the client probably experiences shame about most aspects of the self. The specific aspect stressed in the Initial Interview is likely to be the most extreme of the specific shame factors since that Is what makes it salient to the client. But many other aspects may be nearly as painful. Furthermore, the total pain due to shame experience across the minor aspects may be greater than that for the one most extreme aspect. Consider the Implications for treatment. The causal Implications of the hierarchical measurement model are that changes in specific shame may be transitory while changes In general shame generalize to all the specific aspects. If general shame can be reduced, then over time there should be a corresponding decrease In all specific shame aspects. Thus the therapist may be better off to treat the client for general shame rather than the specific shame emphasized by the client in the Initial clinical interview. Consider first the implications for treatment focused on the extreme aspect emphasized In the Initial Interview. Focused treatment of specific shame aspects may not generalize to general shame. Thus the client will be treated for only that one aspect of shame and not for the many other aspects of shame that are probably present. Furthermore, If the specific factors tend to be transitory over time, then there may be a tendency for the focal aspect to return to its original high level after therapy. 126 CIinica1_1mniication_EoL_Lananane_nescnibins_fimotions Language that descrlbs emotions is used at three different levels In clinical work. First, there Is communication between professionals on personality theory. Second, there Is the language used by therapists when talking to clients. Third, there is the language used by clients talking to therapists. This study has findings relevant to all three levels. Consider communication between professionals. This can be regarded as the focus of the present study: to clear up confusions and ambiguities as to what Is and what is not shame. This study has shown that most of the cllnlclal literature has been too broad, that too many different concepts have been Inappropriateiy grouped under the label of "shame.“ Some clinicians have referred to shame-related traits or emotions such as embarrassment, fear of exposure, public self- consclousness, and shyness as a "shame family of emotions“ (e.g., Wurmser, 1981; Nathanson, 1987). Morrison (1983) used shame to subsume many "closely enough related“ emotions. The findings of this study show that at the level of zero order correlations shame is differentiated from embarrassment, fear of exposure, public self-consciousness, shyness, and guilt. Furthermore, It has been shown that shame Is no; the higher order factor for these related traits or emotions. This is contrary to the "shame family of emotions“ hypothesis. Consider now the language used by therapists talking to clients. Empathlc understanding has been shown to be very Important in counseling and psychotherapy (e.g., Carl Rogers, Heinz Kohut). Having a clinical language that speaks to and further clarifies clients' experiences ls 127 most essential in helping relationships. There are two ways that current language use can confuse understanding. First, the client may use the right word but the therapist makes a false Inner translation. Second, the client may use the wrong word and the therapist may not ask the follow-up questions that would properly identify the affect. Consider false translation by the therapist of the correct word used by the client. Some contemporary therapists are prone to translate many other affects Into shame. Suppose the client refers to feelings of "embarrassment" or “shyness." Just because someone reports feeling embarrassed it does not necessarily mean that they feel ashamed. A person can feel shy without feeling ashamed; or guilty without feeling ashamed. Thus the current tendency to treat many negative affects as shame leads to a high probability of false Interpretation by the therapist. it is also likely that clients would be confused If they talk about "shyness"-—most often a fear of strangers--and their therapist responds with questions about "shame." Consider false acceptance by the therapist of a wrong word used by the client. The word "shame“ is often used by ordinary people for other affects such as embarrassment or gullt. Thus, some clients may say "I feel ashamed of myself" when they actually mean embarrassment. A therapist preset to translate many negative affects into shame will not think to question the word used by the client. It Is wise for the therapist to try out other words or descriptive phrases to make sure that the actual affect Is identified accurately. Even after science teaches therapists how to use their words with precision, ordinary people will continue to use words in the rough 128 manner as the present. There are two implications. First, therapists should use clarifying responses to correctly Identify the affect. Second, therapists may be able to train clients to Improve their emotional language. There was once a therapy derived from General Semantics which had some success just by focusing on Identifying illogical beliefs and by making clients aware of them (Korzybskl, 1933). That therapy did look for logical errors In addition to wrongful word use, but incorrect word use was a frequent cause of Iliogicai generalization and accompanying false beliefs. Many traditional treatments of depression focus on guilt. The findings of this study are that guilt is only related to depression because it Is an Indicator for shame. As shame has been shown to be a more Important factor in depression than gullt the treatment of depression should focus on shame rather than guilt. Lewis (1971, 1979, 1987a) has written amply on the focus of shame In treating depression. Describing the clinical features of affective disorders in CganenensLLe Iexlbggk_91_fisygnlaLLy, Lehmann (1985) stated that "feelings of personal inadequacy (l.e., shame] seem to have replaced gullt feelings" (p.794). Daxe1opmen1a1_1naoLx_91_Sname Two contrasting theoretical models of shame have been proposed In the shame literature (Novak, 1986). One is the shame profile model (Joffe, 1984; Kaufman, in press) which says that shame affect first becomes associated with particular aspects of oneself or with particular situations. it then generalizes to shame about the whole self. The 129 second model suggests, instead, that differences In shame vulnerability lead to differences In the extent of specific kinds of shame. According to this theory, general shame measures shame vulnerability. Therefore, "vulnerability theory” can be rephrased to say that the specific kinds of shame (l.e., shame profile components) develop from general shame. The first model is called "conditioning theory"; and the second model "vulnerability theory." The findings in this study support the ”vulnerability theory" of shame and disconfirm the predictions of "conditioning theory." Wm A large body of research literature on self-esteem has been accumulated over the past half century. Only in the past decade or so have writers and researchers started to focus on shame. The findings of this study show that, at the trait level, shame and self-esteem are Identical constructs. Thus researchers can start to bridge the two fields. Quantitatively oriented researchers can utilize the meta- analytic methods to sort through such a rich jungle of empirical findings. The major focus of this study was the measurement of shame. The next step is to build causal models of shame. Some very Important research questions are: What are the causal antecedents to shame? What makes one person susceptible to shame while another ls Immune to it? Are there genetic bases for shame? What are the consequences of shame? Are there different sets of causal antecedents for shame, embarrassment, fear of exposure, public self-consciousness, shyness, and guilt? How are these traits different from each other substantively? 130 As scales have been constructed to measure shame, embarrassment, shyness, audience anxiety, and public self-consciousness, the Buss (1980) model of self-consciousness and social anxiety can be put to empirical test. Some questions can be asked: What are the relationships between shame, pathological narcissism, and healthy narcissism? What about shame and Irrational beliefs? Shame and alexithymia? Shame has been shown to be related very substantially to emotional problems. Thus, it is of great practical importance to explore what to do about or how to cope with shame. Much of the focus In psychotherapy SOSSIOI‘TS deals Wlth ISSUOS about shame 811d SSIT-OStOOM. SUMMARY AND CONCLUSIONS in the recent past the concept of shame has assumed an Increasingly Important position In the clinical literature on psychopathology and psychotherapy. However, the concept of shame has not had adequate theoretical development and hence has not produced an empirically verified measurement model. Therefore, this study had two goals. Definition and measurement of shame were the first priority of this study. Second, this study sought answers to a number of clinically relevant questions about shame. What is the relationship between shame and self-esteem? What is the relationship between shame and guilt? How does shame develop? Should treatment for shame focus on specific shame aspects or on general shame vulnerability? Are there large gender differences on shame and guilt proneness? This study was a cross-sectional study of correlations based on a non-clinical sample of 926 college students. A large number of items were generated In order to fully investigate a wide variety of concepts associated with the concept of shame. Scales were found or constructed to measure six shame themes, four components of the shame profile, and five shame-related traits (l.e., embarrassment, fear of exposure, public self-consciousness, shyness, and guilt). The clinical relevance and causal equivalence of variables were assessed by measuring variables that assess the quality of emotional and social functioning. Variables 131 132 that measure emotional well-being or problems are: depression, anxiety, anger, loneliness, assertiveness, satisfaction with life, and resentment. Variables that measure social functioning are: conformity, egocentrism, exhibitionism, dominance, self-revelation, competitiveness, need for approval, and audience anxiety. Ine_ne1Ini1Inn_and_MeasuLemen1_21_sname This study sought a scientific definition of shame that would include concepts that are causally and clinically equivalent to each other but exclude concepts that are not equivalent. Six themes were distilled from clinical literature and clients' reports of their shame experiences. Six scales were constructed, one for each shame theme. The data showed each scale to be internally and externally consistent. it was shown that at the level of a hierarchical measurement model, the six shame theme constructs are causally equivalent to each other and thus unambiguously define a general factor. The general factor for these shame themes ls called “general shame." Some authors have argued that there are many shame traits: one for each aspect of the self toward which the person might feel shame. This has been crystalized in the concept of "shame profile“: four major shame traits for competence shame, relationship shame, body shame, and character shame. A scale was constructed for each component of the shame profile. The results showed that the character shame scale defined only a narrower shame aspect called dependency shame. The four scales were internally and externally consistent. Flt of a hierarchical measurement model showed that the four components of the shame profile are causally equivalent to each other. Thus they unambiguously define a 133 general factor. That general factor is identical to the general factor for the shame theme scales. Thus the same construct of general shame captures all the clinically relevant Information in the Items of the six shame theme scales and the four shame profile scales. aname_and_aelated_ILa11s Various authors have postulated a number of theories relating shame to embarrassment, fear of exposure, public self-consciousness, shyness, and guilt. These five concepts are referred to as the "shame-related traits.“ Scales were found or constructed for each of the five shame- related traits. Four theoretical positions were Identified from the literature and were tested in this study. Some authors have argued that one or more of the shame-related affects are actually forms of shame and thus would be Identical to shame at the trait level. The results showed that not one of the shame-related traits is identical to shame. Some authors have argued that one or more of the shame-related affects Is a variant of shame. Thus, while that trait might not be identical to shame at the trait level, it would be causally equivalent. The data showed that none of the shame-related traits is causally equivalent to shame. Some authors have argued that shame and the five related traits form a larger complex of variations from an underlying higher order factor (e.g., social anxiety). At the level of a hierarchical measure- ment model, these traits would be causally equivalent to each other. The data showed that this, too, is false. in other words, these shame- reiated affects are no; variations of shame at any level. The affects of embarrassment, fear of exposure, public self-consciousness, shyness, and guilt are all distinct from shame at both the state and trait level. 134 Shame_and_Sei1:esteem The most salient clinical variable that would be expected to be related to shame Is self-esteem. This study hypothesized that shame and self-esteem are simply the two ends of the same bipolar dimension. Self-esteem was measured by the two most commonly used scales in the literature: the Rosenberg scale and the Cheek and Buss scale. The data showed that shame and self-esteem are identical constructs. They are almost perfectly negatively correlated (L - -.92, corrected for attenuation). More lmportantiy, they are perfectly parallel In their correlations with all outside variables. Thus they are causally and clinically equivalent. From the point of view of a hierarchical measurement model, they are unldimensionai by both the test for Internal consistency and the test for parallelism. in short, shame and self- esteem are the same dimension. Shamn_and_fiulll This study found a correlation of .52 (corrected for attenuation) between shame and guilt, a correlation close to the average correlation of .48 (uncorrected for attenuation) In a meta—analysis across several studies. Furthermore, from the point of view of a hierarchical measurement model, It was shown that shame and guilt are not equivalent to each other. It was, therefore, concluded that shame and guilt are two distinct traits. ShameL_GuLL1._and_DenLassion This study found that shame correlates more highly with depression than gullt does. The method of partial correlation showed that gullt does not correlate with depression when shame Is partlaled out whereas 135 shame still correlates quite highly with depression when guilt is partlaled out. Similar results consistent with this study were found by applying the method of partial correlation on the data of two other studies (Harder & Lewis, 1987; Hobiltzelie, 1987). it was concluded that the role of guilt has been overstated In the literature of depression. GandeL_nllieLsncas_Ln_Snama_and_Gu111_ELonenass Gender differences were examined. in this study there was no difference between men and women on shame proneness. However, a meta- analysls showed that the average correlation across several studies Is .11. It was concluded that the present study probably had a value that was low by sampling error. Thus women are silgnliy more vulnerable to shame than men. Gender differences on gullt proneness were also examined. The difference In this study was small (L - .13). This exactly matched the value found In a meta-analysis. Thus women are also slightly more vulnerable to guilt than men. This is contrary to what Helen Block Lewis would predict. Gender differences on specific aspects of shame were also examined. Women are higher on body shame than would be predicted from the difference on general shame (partlal L - .31). This presumably reflects the societal focus on good looks In women. Men are slightly higher on relationship shame than would be predicted from the difference on general shame (partlal L - -.15). The currently popular thinking Is that women are more socialized around relationship building than men and men are, therefore, more deficient In their ability to establish 136 relationships and hence slightly more prone to feel unlovable. Another possibility may be that parents tend to handle their girls with kid gloves but tend to treat their boys with harsher measures. Boys are socialized as if their feelings do not get hurt as easily as girls and hence they may feel less lovable. 1mnI1ca11ons_1eL_Lansuaqe_Use_In_InaLanx This study has shown that In the cilniciai literature too many concepts have been grouped Inappropriateiy under the label of "shame." Some clinicians have referred to shame-related traits or emotions such as embarrassment, fear of exposure, public self-consciousness, and shyness as the "shame family of emotions" (Wurmser, 1981; Nathanson, 1987). The findings of this study show that at the level of zero order correlations shame is differentiated from embarrassment, fear of exposure, public self-consciousness, shyness, and guilt. Furthermore, It has been shown that shame Is no; the higher order factor for these related traits or emotions. This is contrary to the “shame family of emotions" hypothesis. The Implication for language use In psychotherapy is that therapists need to pay more attention to the feeling words that clients use in describing their emotional experiences. it Is unwise for therapists to use these feeling words too loosely or interchangeably. This often may confuse clients. For example, clients may be confused If they talk about "shyness"--most often a fear of strangers--and their therapist responds with questions about “shame." Accurate and empathic understanding of clients' affective experiences has been shown to be essential in psychotherapy (e.g., Carl Rogers, Heinz Kohut). 137 Shame_InnLanx4__IQ_EQnua_Qn_GannLal_§namn_QL_§nnnlli£_§nama7 The hierarchical measurement model relating general shame and the specific shame profile shows that general shame Is causally antecedent to specific shame. In the long run, any changes in general shame will be passed on to all or most of the specific shame dispositions. However, the reverse is not true. That is, treatment of specific shame would have no impact on the underlying shame vulnerability. In other words, changes In specific shame may be transitory and may not generalize to general shame. Thus the therapist may be better off treating the client for general shame rather than the specific shame emphasized by the client In the initial clinical interview. Treatment of specific shame would have little long term benefit. Ine_nexeionment_oi_ahama Two contrasting theoretical models of shame have been proposed in‘ the shame literature. One Is the shame profile model which says that shame affect first becomes associated with particular aspects of oneself or with particular situations. It then generalizes to shame about the whole self. The second model suggests, Instead, that differences in shame vulnerability lead to differences In the extent of all specific kinds of shame. In other words, specific kinds of shame develop from general shame. The first model is called "conditioning theory;" and the second model ”vulnerability theory.“ The findings in this study support the "vulnerability theory” of shame and disconfirm the predictions of "conditioning theory." APPENDIX A-l An Outline of Kaufman's Developmental Theory of Shame An outline of Kaufman's theory is given here. The interested- reader is referred to his forthcoming book, Ine_fisygngiggy_91_§nameL InQQLx_and_ILsalmsn1_QL_§namn:Bascd_§xndLQme§ (In press) and Shame;__lfl§ W (1985)- it Is beyond the scope of this study to give a comprehensive treatment of Kaufman's theory of shame. Only four aspects of his theory will be mentioned here. First, Kaufman postulates some 1nnaie_1aLLQLs which Include different Innate thresholds for the activation of shame affect for different individuals, variations In the actual strengths of innate shame affect In different Individuals, and temperament (e.g., introverslon vs. extroversion). These factors are used to account for Individual differences. Kaufman's innate factors include an Innate predisposition or susceptibility to shame independent of shaming experiences; however, this is not the dominant position he takes. Second, he conceptualizes the inleLnaiizaliQn process of shame affect: He sees shame Internallzation as a very significant developmental process. “Prior to Internallzation, shame remains a feeling which Is generated and then passes on, whereas following Internallzation, shame can be prolonged Indefinitely“ (Kaufman, 1985, p. 7). Through this internalizing process, shame can spread throughout the self, ultimately shaping one's emerging identity. He postulates four contributing sources of lnternalization which Include developments of affect-shame binds, drive-shame binds, Interpersonal need-shame binds, and purpose-shame binds. For example, any affect that meets with sufficient shaming can develop Into an affect-shame bind; that is to 138 139 say, the expression of that particular affect becomes bound and controlled by shame. The same process takes place with regard to expressions of drive, Interpersonal needs, and purpose; that Is, they become drive-shame binds, interpersonal need-shame binds, and purpose- shame binds. In other words, expressions of these four basic human motivational systems are now bound and associated with shame. This aspect of Kaufman's theory Is clearly a “conditioning theory." This Is the dominant position Kaufman has taken on shame development. Third, Kaufman postulates the magniLLLalign process: Shame about particular attributes of the self becomes magnified to Include shame about the whole self via imagery and language. These isolated shame experiences are initially Induced Interpersonally and they are stored in memory in the form of scenes. Through language, one comes to interpret separate shame experiences as signifying essential meanings about the self. Scenes of shame fuse together, thereby magnifying each other and generating a "script" (in Tomklns's sense) of shame about the self. Lastly, Kaufman postulates a dexeionmentai_nLocess of the shame profile: He views the shame profile as emerging across time. He postulates that the first stage of magnification of shame affect includes four primary shame scene dimensions: purpose shame, affect shame, drive shame, and Interpersonal need shame. The second stage of magnification fuses these into: competence shame, body shame, and relationship shame. The final stage of magnification ls character shame (or shame-based identity). The three stages of magnification are diagramed as follows (Kaufman, In press): 14O Stage III Character Shame Stage lI Competence Body Relationship Shame Share Share Stage i Purpose- Affect- Drive- Interpersonal Shame Shame Shame Need-Shame Scenes Scenes Scenes Scenes EiguLe_11. Stages in psychological magnification. Kaufman advocates a longitudinal methodology to study the developmental theory of shame; that Is, data on changes in developing children over time would be more direct and relevant in testing his developmental theory of shame. it should be pointed out that this study did not test all aspects of Kaufman's theory which is much more elaborate and extensive. This study focused on some aspects of his developmental theory, i.e., the concept of shame profile and the Issue of generality vs. specificity of shame . APPENDIX A—2 Inventory Filled Out by Study Subjects Directions: Please Read Carefully! And ng_u91_§Kln_Any_11gm§! Mark your responses to the questionnaire Items on the 1W0 enclosed machine-scoreabie answer sheets. Please use EENQLL as these sheets will be scanned by machine. Each section of the questionnaires Is preceded by a direction and a response key. Read them first. Then respond to the items In that section using that key. On the E1351 answer sheet (nine), please enter your responses to Items 11_1nLgugn_ngQ and stop right on 11gm_gzgg on the answer sheet. Then for your reponses to 11em_1291_1nL9ugn_1211, please enter them on the SECOND answer sheet (Led). That is to say, answer 01 is for your response to Item 4201; answer 02 Is for your response to item «202... answer «77 Is for your response to item «277. 11£m_£ AnfiflfiL_i 1 1 2 All on the ILLS; . answer sheet (bins) 200 200 201 > 1 202 — > 2 All on the second . . answer sheet (Leg) 277 > 77 - Please be sure that the item numbers correspond to the answer numbers on the answer sheets as illustrated above. Do n91 enter your name or student number anywhere on the answer sheets. Please return the questionnaires and Lug answer sheets to: Room . D g o- .. I... 11-. e - .0 so .H— go‘s. “0.0.. o.‘ Inank_1ou_and_hona_1h1s_n111_bs_a_maan1n91u1_aana11ance_IoL_xou! Before you get started, please complete the following: Today's Date Your Age On bOth machine-scoreabie answer sheets, please blacken E In the SEX column if you are a female student, and blacken u,lf you are a male student. 141 142 A number of statements which people have used to describe themselves are given below. Read each statement and fill in the proper place on the answer sheet to Indicate how you geneLaliy feel. There are no right or wrong answers. W SgalgmgnL but give the answer that describes how you geneLaLLL feel. EicaSQ_DQ_N91_§Kln_An1_11ems. Use the following key: WWWWW 1. i am "calm, cool, and collected." 2. i don't like my physical appearance. 3. i feel like I am never quite good enough. 4. I have a fiery temper. 5. I feel ashamed for being an emotionally weak person. 6. i lack companionship. 7. i feel somehow left out. 8. i feel I am just as good as my friends. 9. i am happy. 10. I think that peOple look down on me. 11. I see others as being more mature than me emotionally. 12. i feel unlovabie. 13. Compared to other people I feel like I somehow never measure up. 14. I am quick-tempered. 15. I feel ashamed for being overly emotional. 16. I scold myself and put myself down. 17. i feel ashamed for not being more independent. 18. i see myself as being very small and insignificant. 19. i feel disappointed In myself. 20. i say to myself, "how could anyone really love me or care about me?" 21. I can easily describe myself as self-confident. 22. i feel defective as a person, as if something is basically wrong with me. 23. I am down on myself for any character flaws in me. 24. i feel as If there is something repulsive about my body. 25. i feel like i don't belong. 26. i feel ashamed for needing a lot of advice and support from others. 27. I am worthy of being loved. 28. When someone ignores me, I cannot help wondering what is wrong with me. 29. i feel unwelcome, unwanted by others. 30. I feel competent as a person. 31. I can't seem to do anything right. 32. I put myself down for being incompetent. 33. I always seem to fall short of what I aspire to be. 34. There is no one i can turn to. 35. I feel very Inadequate as a person. 36. i have disturbing thoughts. 37. I doubt that anyone will ever love me or accept me. 38. I am a hotheaded person. 39. I feel ashamed for being too dependent on others. 40. i am an outgoing person. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78. 79. 80. 81. 82. 83. 84. 85. 86. 87. 88. 89. 90. 143 i am a worthwhile person. I am content. it makes me furious when I am criticized In front of others. I feel left out. I feel like a failure. i feel secure. i get angry when i'm slowed down by others' mistakes. I feel Isolated from others. I feel like i am not worthy of being loved. l I I I become tense and upset when i think about my present concerns. feel like I am good for nothing. get infuriated when I do a good job and get poor evaluation. can find companionship when i want it. Compared to other people I feel like I don't measure up. I feel that difficulties are piling up so that I cannot overcome them. I feel so unimportant, just like a nobody. fly off the handle. am unhappy being so withdrawn. feel worthless as a person. worry too much over something that really doesn‘t matter. feel that i look ugly. see myself as intact and without personal defects. don't have confidence In myself. find that I don't live up to my own standards or ideals. feel annoyed when I am not given recognition for doing good work. don’t like myself. don't feel competent In most things. People are around me but not with me. I feel like a useless person. I feel intensely inadequate and full of self doubt. Some unimportant thought runs through my mind and bothers me. i can easily describe myself as good-looking and attractive. When I get mad, i say nasty things. i feel I am a complete failure as a person. I feel pleasant. i feel ashamed of myself. I see myself striving for perfection only to continually fall short. I I feel very unattractive. feel unsure of myself. I am pleased with myself. When I compare myself to others i am just not as important. I am ashamed of my body. When i get frustrated, i feel like hitting someone. It is difficult for me to accept a compliment. I feel like I am just not quite good enough. I feel ashamed of my weight. i feel Inferior to most of my friends. I feel there Is something defective In my character. i feel so insignificant to others that I feel as if i aminvlsibie. i hate certain parts of my body because they make me look unattractive. 144 You are to read each of the following statements and decide if the statement is true of you or not. You then fill out the corresponding space on the enclosed machine scoreable answer sheet using the following key: Strongly Disagree Disagree Uncertain or indifferent Agree Strongly Agree mauve-n IIIII 91. On the whole, I am satisfied with myself. 92. I'm concerned about what other people think of me. 93. The few times I have cheated, i have suffered unbearable feelings of guilt. . 94. Most people seem to be more aggressive and assertive than i am. 95. At times i think i am no good at all. 96. i usually worry about making a good impression. 97. it bothers me that i did not do more for my parents. 98. l have hesitated to make or accept dates because of "shyness." 99. I feel that l have a number of good qualities. 100. I'm concerned about the way i present myself. 101. l have done many things that made me feel guilty afterwards. 102. When the food served at a restaurant Is not done to my satisfac- tion, I complain about it to the waiter or waitress. 103. i am able to do things as well as most other people. 104. I'm self-conscious about the way I look. 105. i am concerned about being forgiven for my sins or wrongdoings. 106. If a salesperson has gone to a lot of trouble to show me something which I don't really like, i have a hard time saying “No.“ 107. i feel I do not have much to be proud of. 108. i'm usually aware of my appearance. 109. When I do wrong, my conscience punishes me severely. 110. I enjoy starting conversations with new acquaintances and strangers. 111. l certainly feel useless at times. 112. One of the last things i do before leaving the house Is look in the mirror. 113. I often feel that l have not lived the right kind of life. 114. l often don't know what to say to attractive persons of the opposite sex. 115. i feel that I'm a person of worth, at least on an equal plane with others 116. I'm concerned about my style of doing things. 117. i sometimes have bad thoughts which make me feel awful about myself. 118. i will hesitate to make phone calls to business establishments and Institutions. 119. i wish i could have more respect for myself. 120. I reflect about myself a lot. 121. 122. 123. 124. 125. 126. 127. 128. 129. 130. 131. 132. 133. 134. 135. 136. 137. 138. 139. 140. 141. 142. 143. 144. 145. 146. 147. 148. 149. 150. 151. 152. 153. 154. 155. 145 i would rather apply for a job or for admission to a college by writing letters than going through with personal Interviews. All In all, I am inclined to feel that i am a failure. i'm generally attentive to my Inner feelings. l often feel guilty. Strongly Disagree Disagree Uncertain or Indifferent Agree Strongly Agree Olbwfo-i IIIII I find it embarrassing to return merchandise. i take a positive attitude toward myself. I'm always trying to figure myself out. i have felt very guilty for letting down those close to me. I have avoided asking questions for fear of sounding stupid. I have a low opinion of myself. i'm constantly examining my motives. It bothers me that l have not done more for my parents or family members. I am open and frank about my feelings. l often wish i were someone else. i'm alert to changes In my mood. I have felt very guilty for not being there when someone close to me needed me. I often have a hard time saying "No." Things are all mixed up in my life. I tend to scrutinize myself. When i let my anger out i often feel very guilty afterwards. I tend to bottle up my feelings rather than make a scene. I'm fairly sure of myself. Generally, I‘m not very aware of myself. I feel horrible for having hostile feelings toward other people. I complain about poor service In a restaurant and elsewhere. I am a failure. i’m aware of the way my mind works when I work through a problem. Sometimes I cannot forgive myself for how i have treated others, for the hurt feelings I have caused. If a couple near me In a theater or at a lecture were conversing rather loudly, I would ask them to be quiet or take their conversation elsewhere. I am basically worthwhile. i'm often the subject of my own fantasies. In the past few years i have felt so guilty that I may do something terrible to myself. Anyone attempting to push ahead of me in a line is In for a good battle. I sometimes have the feeling that I'm off somewhere watching myself. Sometimes i hurt people I love or care for and feel very guilty about it afterwards. 156. 157. 158. 159. 160. 161. 162. 163. 164. 165. 166. 167. 168. 169. 170. 171. 172. 173. 174. 175. 176. 177. 178. 179. 180. 181. 182. 183. 184. 185. 186. 187. 188. 189. 190. 191. 146 I am quick to express an opinion. I often cannot forgive myself for having caused deep pain in those i love or care for. l often feel guilty for being better off than my family members. l'd like to shrink away when i make a fool of myself in public. l often feel like I don't want to be seen. I am socially somewhat awkward. When I become embarrassed, i would like to go hide in the corner. i often worry that i might do something Inappropriate in social situations. Strongly Disagree Disagree Uncertain or indifferent Agree Strongly Agree «sun- IIIII I find It hard to talk to strangers. When i am embarrassed, i wish the earth would open up and swallow me. i am afraid that my faults or weaknesses will be exposed in front of others. I feel tense with people i don't know well. Sometimes I feel so exposed, so self-conscious that I wish i could sink into the ground. I often hesitate to reveal myself to others. When conversing, l worry about saying something dumb. Sometimes I feel like there are 1,000 eyes staring at me. i worry about making foolish mistakes, and wonder what other people would think. i feel nervous when speaking to someone in authority. I often feel embarrassed by compliments people pay to me. I worry that others might think some of my Ideas are ”crazy." I am uncomfortable at parties and other social functions. i blush easily. I keep secrets and worry that they might be discovered. I feel Inhibited In social situations. i get embarrassed when people ask me very personal questions. I have trouble looking someone right In the eye. in certain situations i feel like melting away. i am very concerned about the impression i make on others. I am more shy with members of the opposite sex. i blush easily when I am embarrassed. i am often afraid of exposing the sensitive, intimate and vulnerable aspects of myself to others. I get easily embarrassed when people tease me. When i look back on what's happened to me, I can't help feeling resentful. in different situations and with different people, i often act like very different persons. in public places, i am always aware of people watching me. i like to tell jokes. 192. 193. 194. 195. 196. 197. 198. 199. 200. 147 I feel that I can dominate a social situation. i feel unable to tell anyone all about myself. The most Important thing in life is winning. Before I raise my hand in class, i always worry about what other classmates might think of my question. i feel very relaxed before speaking in front of a group. it makes me mad when I see other people getting things that they don't deserve. When I am uncertain how to act in social situations, I look to the behavior of others for cues. i worry about what other people are saying about me. I am satisfied with my life. *‘* BEM1NQEB: Starting on Item 0201, please switch to the SECOND answer sheet (Leg). Use Answer 01 for your response to Item 0201; Answer 02 for your response to Item s202...Answer 077 for your response to Item 0277. 118m_g AnsneL_g_on the Second Answer Sheet Land) 201 > 1 x 202 > 2 All on the Seggnd - AnsweL_§nca1!! 277 > 77 201. 202. 203. 204. 205. 206. 207. 208. 209. 210. 211. 212. 213. 214. 215. 216. 217. 218. 219. 220. 221. 222. 223. 224. 225. 226. 227. 228. 229. 230. 148 1 - Strongly Disagree 2 - Disagree 3 - Uncertain or indifferent 4 - Agree 5 - Strongly Agree I enjoy being the center of attention. in most social situations, I emerge as the leader. i reveal my deepest feelings to my friends. I would rather cooperate and share than compete and win. i usually maintain my position even when my superiors disagree. I feel anxious when i speak In front of a group. i am bitter about not getting what is coming to me. Even if i am not enjoying myself, i often pretend to be having a good time. The conditions of my life are excellent. i often wonder about what gossip there Is about me. i would enjoy speaking to a large group. I feel that I can control a social situation, even though It may not be obvious to other people. if someone Is Important to me, It Is easy for me to let them know It. My life would not be meaningful without a chance to compete and do better than others. i usually avoid doing something that might provoke criticism. i feel very nervous while performing In front of people. i'm not always the person that i pretend to be. in most ways my life Is close to my ideal. I try to be inconspicuous. I easily change my mind after I hear what others have to say. My voice never shakes when i recite in class. I often put on a show to Impress or entertain people. i seldom try to call attention to myself. i feel comfortable being different from those around me. Sometimes my body trembies when i speak up in class. At parties and social gatherings, I do not attempt to do or say things that others will like. i enjoy entertaining other people. i find it hard to do anything that my parents would disapprove of. I may deceive peOple by being friendly when i really dislike them. I'm likely to discontinue doing something that others think Is not worthwhile. 231. 232. 233. 149 1 - Strongly Disagree 2 - Disagree 3 - Uncertain or Indifferent 4 - Agree 5 - Strongly Agree So far I have gotten the important things I want in life. When i take a stand i tend to hold on to It even If others disagree. if I could live my life over, I would change almost nothing. The following statements mention ways parents sometimes react. Read each one and decide how It relates to the way your parents gnaLagleLlsllgaiiy reacted to you before the time you were Inelxe_yeaLs ILL!!- 234. 235. 236. 237. 238. 239. 240. 241. 242. 243. 244. 245. 246. 247. 248. 249. 250. 251. 252. 253. 254. Strongly Disagree Disagree Uncertain or indifferent Agree Strongly Agree G‘UN-e IIIII My parents were often critical of what I did. My mother avoided looking at me when l disappointed her. My parents never seemed satisfied with what i did. i was sometimes punished with a stick or switch. My father avoided looking at me when I disappointed him. My parents had a lot of expectations of me growing up. I was often spanked. Sometimes when my mother disapproved of something I did, she was cold and distant. When I hurt my mother's feelings, she stopped talking to me until I pleased her again. My parents rarely tried to make me conform to their values. When I upset my mother, she did not have anything to do with me until I found a way to make up. My parents were easygoing and rarely commented on what i was doing. My parents believed in spanking when I broke the rules set for me. My parents gave me lots of warmth and affection. Sometimes when my father disapproved of something I did, he did not speak to me for awhile. My parents expressed affection toward me more than most parents. My parents believed that physical punishment builds character. When i hurt my father's feelings, he stopped talking to me until I pleased him again. My parents were somewhat reserved towards me. My parents were not particularly Interested in what I did. When i upset him, my father didn't have anything to do with me until I found a way to make up. 150 The following statements mention what some people went through while they were growing up. Read each one and decide how It relates to how people around you geneLallx treated you before the time you were 12 111L3_Qlfl- Strongly Disagree Disagree Uncertain or Indifferent Agree Strongly Agree (”.5de IIIII 255. i was often ignored by some kids of my age. 256. l was often mocked and laughed at by my peers. 257. I was often ridiculed by other kids. 258. i was looked up to by most of my peers. 259. l was well accepted, well liked by other kids. 260. I was often rejected by other kids. 261. i was often looked down upon. 262. l was often scolded or yelled at by some grown-ups. 263. I was often rejected by some adults. 264. I was often laughed at or made fun of by some grown-ups. Following are groups of statements. Please read the entire group of statements In each category. Then pick out the one statement In that group which best describes how you have felt In the fiASI_IflQ_MQflIfl§. Please mark your answer In the appropriate place on the answer sheet. Enter the number (1, 2, 3, or 4) In front of the statement In each category that best describes you. 265. 1 I do not feel sad. 2 Sometimes i feel sad or blue. 3 I am blue or sad all the time and I can't snap out of it. 4 I am so sad or unhappy that I can't stand it. 266. 1 i am not particularly pessimistic or discouraged about the future. 2 I feel discouraged about the future. 3 I feel i have nothing to look forward to. 4 i feel that the future ls hopeless and that things cannot Improve. 267. 1 i do not feel like a failure. 2 i feel i have failed more than the average person. 3 As i look back on my life, all i can see is a lot of failures. 4 I feel i am a complete failure as a person. 268. 1 i am not particularly dissatisfied. 2 i don't enjoy things the way I used to. 3 i don‘t get satisfaction out of anything anymore. 4 I am dissatisfied with everything. 151 I don‘t feel particularly guilty. i feel bad or unworthy a good part of the time. 3 i feel quite guilty. I 4 feel as though I am very bad or worthless. 270. 1 i don't feel disappointed In myself. 2 I am disappointed In myself. 3 i am disgusted with myself. 4 i hate myself. 271. 1 I don't have any thoughts of harming myself. 2 i feel I would be better off dead. 3 I have definite plans about committing suicide. 4 i would kill myself if I had the chance. 272. 1 l have not lost interest in other people. 2 i am less Interested in other people than i used to. 3 i have lost most of my Interest in other people and have little feeling for them. 4 i have lost all of my interest in other people and don't care about them at all. 273 1 make decisions about as well as ever. 2 try to put off making decisions. i l 3 I have great difficulty In making decisions. i 4 can't make any decisions at all anymore. 274. 1 I don't feel i look any worse than I used to. 2 i am worried that I am looking old or unattractive. 3 I feel that there are permanent changes in my appearance and they make me look unattractive. 4 I feel that I am ugly or repulsive looking. 275. 1 I can work about as well as before. 2 It takes extra effort to get started at doing something. 3 I have to push myself very hard to do anything. 4 I can't do any work at all. 276. I don't get any more tired than usual. 2 i get tired more easily than i used to. I I 3 get tired from doing anything. 4 get too tired to do anything. 277. 1 My appetite is no worse than usual. 2 My appetite is not as good as it used to. 3 My appetite is much worse now. 4 i have no appetite at all anymore. APPENDIX A-3 Individual Scales Used In This Study 66. 80. 19. 76. 85. 87. 54. 22. 88. 23. 62. 59. 69. 51. 74. 45. 41. 56. 89. 33. 64. N911: Ibe_Sname_Ineme_§caies_ichand_&_flunteL+_1a§§1 (1) Disappointment with Self don't like myself. am pleased with myself. feel disappointed In myself. feel ashamed of myself. (2) Feelings of Inferiorlty I feel like I am Just not quite good enough. I feel Inferior to most of my friends. Compared to other people I feel like I don't measure up. I feel I am Just as good as my friends. (3) Feelings of Defectiveness I feel defective as a person, as If something Is basically wrong with me. I feel there Is something defective In my character. I am down on myself for any character flaws In me. I see myself as Intact and without personal defects. (4) Feelings of Worthlessness feel worthless as a person. i I feel like a useless person. I feel like I am good for nothing. I feel I am a complete failure as a person. I feel like a failure. i am a worthwhile person. (5) Feelings of Unimportance I feel so unimportant, Just like a nobody. i feel so Insignificant to others that I feel as If I am invisible. (6) Falling Short of Own Standards or ideals i always seem to fall short of what I aspire to be. I find that I don't live up to my own standards or Ideals. The W (1988) Is made up of all the above six shame theme scales. 152 63. 35. 79. 32. 21. 31. 67. 30. 82. 24. 61. 90. 78. 86. 72. 49. 12. 37. 29. 25. 27. 39. 15. 17. 26. 11. \1 153 Sname_ELn1l1a_sca1e3.19nanp_&_flunteLs_19§&1 (1) Competence Shame I don't have confidence In myself. I feel very Inadequate as a person. I feel unsure of myself. I put myself down for being Incompetent. I can easily describe myself as self-confident. i can't seem to do anything right. I don't feel competent In most things. I feel competent as a person. (2) Body Shame I am ashamed of my body. I feel as If there Is something repulsive about my body. I feel that I look ugly. I hate certain parts of my body because they make me look unattractive. I feel very unattractive. I don't like my physical appearance. I feel ashamed of my weight. I can easily describe myself as good-looking and attractive. (3) Relationship Shame I feel like I am not worthy of being loved. I feel unlovable. I doubt that anyone will ever love me or accept me. I feel unwelcome, unwanted by others. I feel like I don't belong. I am worthy of being loved. (4) Dependency Shame I feel ashamed for being too dependent on others. I feel ashamed for being overly emotional. I feel ashamed for not being more Independent. i feel ashamed for being an emotionally weak person. I feel ashamed for needing a lot of advice and support from others. I see others as being more mature than me emotionally. 154 W 91. On the whole, I am satisfied with myself. 95. At times I think i am no good at all. 99. I feel that I have a number of good qualities. 103. I am able to do things as well as most other people. 107. I feel I do not have much to be proud of. 111. l certainly feel useless at times. 115. I feel that I'm a person of worth. at least on an equal plane with others. 119. I wish i could have more respect for myself. 122. All In all, i am lncllhed to feel that I am a failure. 126. I take a positive attitude toward myself. WM 130. I have a low opinion of myself. 134. I often wish I were someone else. 138. Things are all mixed up In my life. 142. I‘m fairly sure of myself. 146. I am a failure. 150. i am basically worthwhile. MW 3. feel like I am never quite good enough. 7. feel somehow left out. 10. think that people look down on me. I i I 13. Compared to other peOple I feel like I somehow never measure up. I I i l 16. scold myself and put myself down. 18. see myself as being very small and insignificant. 20. say to myself, "how could anyone really love me or care about me?" 22. feel defective as a person, as If something ls basically wrong with me. 70. I feel Intensely Inadequate and full of self doubt. 77. I see myself striving for perfection only to continually fall short. 81. When I compare myself to others I am Just not as Important. W 165. When I am embarrassed, I wish the earth would open up and swallow me. 162. When I become embarrassed, I would like to go hide In the corner. 168. Sometimes I feel so exposed, so self-conscious that I wish I could sink Into the ground. 182. In certain situations I feel like melting away. 159. I‘d like to shrink away when i make a fool of myself in public. 171. Sometimes i feel like there are 1,000 eyes staring at me. 187. I get easily embarrassed when people tease me. 180. 166. 172. 163. 175. 186. 178. 169. 92. 96. 100. 104. 108. 112. 116. 183. 161. 164. 167. 170. 173. 176. 179. 181. 184. 155 i get embarrassed when people ask me very personal questions. e .00 ' . ; ”0'. ;e e" 1;. o ...g I. I- I am afraid that my faults or weaknesses will be exposed in front of others. I worry about making foolish mistakes. and wonder what other people would think. I often worry that I might do something inappropriate in social situations. I worry that others might think some of my Ideas are "crazy." I am often afraid of exposing the sensitive, Intimate andvul- nerable aspects of myself to others. I keep secrets and worry that they might be discovered. I often hesitate to reveal myself to others. Eubi1c_sa11:9pnssi9usnasa_LEnni95121n_21_31‘a_121§1 I'm concerned about what other people think of me. i usually worry about making a good Impression. I'm concerned about the way I present myself. I'm self-conscious about the way I look. i'm usually aware of my appearance. One of the last things I do before leaving the house Is look In the mirror. I'm concerned about my style of doing things. I am very concerned about the Impression I make on others. fin1n£§§_§£aln_L§hnnK_&_Bu§3‘_lfifill i am socially somewhat awkward. I find It hard to talk to strangers. i feel tense with people I don't know well. When conversing, i worry about saying something dumb. I feel nervous when speaking to someone In authority. I am uncomfortable at parties and other social functions. I feel Inhibited in social situations. I have trouble looking someone right In the eye. I am more shy with members of the opposite sex. 93. 101. 105. 109. 113. 117. 157. 148. 155. 136. 128. 140. 144. 132. 158. 8|lll3lll|3§ 5" 156 . W The few times i have cheated, l have suffered unbearable feelings of guilt. I have done many things that made me feel guilty afterwards. I am concerned about being forgiven for my sins or wrongdoings. When i do wrong, my conscience punishes me severely. I often feel that l have not lived the right kind of life. I sometimes have bad thoughts which make me feel awful about myself. WW I often cannot forgive myself for having caused deep pain In those I love or care for. Sometimes I cannot forgive myself for how I have treated others, for the hurt feelings I have caused. Sometimes I hurt people I love or care for and feel very guilty about it afterwards. I have felt very guilty for not being there when someone close to me needed me. I have felt very guilty for letting down those close to me. When i let my anger out, i often feel very guilty afterwards. I feel horrible for having hostile feelings toward other people. it bothers me that I have not done more for my parents or family - members. I often feel guilty for being better off than my family members. WM do not feel sad Sometimes i feel sad or blue 0) 7. am blue or sad all the time and I can't snap out of It am so sad or unhappy that I can't stand It I am not particularly pessimistic or discouraged about the future I feel discouraged about the future i feel I have nothing to look forward to I feel that the future ls hopeless and that things cannot Improve i do not feel like a failure I feel I have failed more than the average person As I look back on my life, all I can see Is a lot of failures I feel I am a complete failure as a person (parent, husband, wife, etc.) 157 am not particularly dissatisfied don't enjoy things the way I used to don't get satisfaction out of anything anymore am dissatisfied with everything don't feel particularly guilty feel bad or unworthy a good part of the time feel quite guilty feel as though I am very bad or worthless don't feel disappointed In myself am disappointed In myself am disgusted with myself hate myself don't have any thoughts of harming myself feel I would be better off dead have definite plans about committing suicide would kill myself if I had the chance have not lost Interest In other people am less Interested In other people than I used to have lost most of my Interest In other people and have little feeling for them have IOSt all Of my lhtOf’OSt in other DGODIO and don't care about them at all 01 make decisions about as well as ever try to put off making decisions have great difficulty In making decisions can't make any decisions at all anymore don't feel I look any worse than i used to am worried that I am looking old or unattractive feel that there are permanent changes In my appearance and they ake me look unattractive feel that i am ugly or repulsive looking can work about as well as before It takes extra effort to get started at doing something IIIBI O) IIIIBI have to push myself very hard to do anything can't do any work at all don't get any more tired than usual get tired more easily than I used to get tired from doing anything get too tired to do anything 14. 38. 43. 47. 52. 57. 65. 73. 83. 34. 40. 44. 48. 53. 58. 68. 158 My appetite Is no worse than usual My appetite Is not as good as It used to My appetite ls much worse now I have no appetite at all anymore WW I am "calm, cool, and collected.“ I am happy. I have disturbing thoughts. I am content. i feel secure. I become tense and upset when I think about my present concerns. I feel that difficulties are piling up so that i cannot overcome them. I worry too much over something that really doesn’t matter. Some unimportant thought runs through my mind and bothers me. I feel pleasant. WW I have a fiery temper. i am quick-tempered. I am a hotheaded person. It makes me furious when i am criticized In front of others. I get angry when I'm slowed down by others' mistakes. I get Infurlated when I do a good Job and get poor evaluation. I fly off the handle. I feel annoyed when i am not given recognition for doing good work. When I get mad, I say nasty things. When I get frustrated, I feel like hitting someone. WW3 WW I lack companionship. There Is no one I can turn to. i am an outgoing person. I feel left out. I feel Isolated from others. I can find companionship when I want It. I am unhappy being so withdrawn. People are around me but not with me. 94. 98. 102. 106. 110. 114. 118. 121. 125. 129. 133. 137. 141. 145. 149. 153. 156. 200. 209. 218. 231. 233. 188. 197. 207. 159 WWW L5DQLl_IQLm_§£J££1§fl_hl_9flflflg+_lflfifll Most people seem to be more aggressive and assertive than I am. I have hesitated to make or accept dates because of "shyness." When the food served at a restaurant Is not done to my satisfac— tion, I complain about it to the waiter or waitress. If a salesperson has gone to a lot of trouble to show me something which I don't really like, I have a hard time saying "No." I enjoy starting conversations with new acquaintances and strangers. I often don't know what to say to attractive persons of the opposite sex. I will hesitate to make phone calls to business establishments and Institutions. I would rather apply for a Job or for admission to a college by writing letters than going through with personal interviews. I find It embarrassing to return merchandise. I have avoided asking questions for fear of sounding stupid. I am open and frank about my feelings. I often have a hard time saying “No." I tend to bottle up my feelings rather than make a scene. I complain about poor service In a restaurant and elsewhere. if a couple near me In a theater or at a lecture were conversing rather loudly, I would ask them to be quiet or take their conversation elsewhere. Anyone attempting to push ahead of me In a line Is In for a good battle. I am quick to express an opinion. W I am satisfied with my life. The conditions of my life are excellent. In most ways my life Is close to my ideal. So far i have gotten the important things I want In life. If i could live my life over, I would change almost nothing. WM When I look back on what's happened to me, I can't help feeling resentful. it makes me mad when I see other people getting things that they don't deserve. I am bitter about not getting what Is coming to me. 189. 198. 208. 217. 222. 226. 229. 190. 199. 210. 191. 201. 211. 219. 223. 227. 192. 202. 212. 193. 203. 213. 160 W In different situations and with different people, i often act like very different persons. When I am uncertain how to act In social situations, I look to the behavior of others for cues. Even If I am not enjoying myself, I often pretend to be having a good time. I'm not always the person that l pretend to be. I often put on a show to Impress or entertain people. At parties and social gatherings, i do not attempt to do or say things that others will like. I may deceive people by being friendly when i really dislike them. W In public places, I am always aware of people watching me. I worry about what other people are saying about me. i often wonder about what gossip there Is about me. ExbJb1119n15m_LHunteL+_1a§31 like to tell jokes. enjoy being the center of attention. would enjoy speaking to a large group. try to be inconspicuous. seldom try to call attention to myself. enjoy entertaining other people. Dominance_iflunteLa_12&31 I feel that I can dominate a social situation. In most social situations, I emerge as the leader. i feel that I can control a social situation, even though it may not be obvious to other people. 51W i feel unable to tell anyone all about myself. I reveal my deepest feelings to my friends. If someone Is Important to me, It Is easy for me to let them know It. 194. 204. 214. 195. 205. 215. 220. 224. 228. 230. 232. 196. 206. 216. 221. 225. 161 W The most Important thing In life Is winning. i would rather cooperate and share than compete and win. My life would not be meaningful without a chance to compete and do better than others. WW Before I raise my hand In class, I always worry about what other classmates might think of my question. I usually maintain my position even when my superiors disagree. I usually avoid doing something that might provoke criticism. I easily change my mind after I hear what others have to say. I feel comfortable being different from those around me. I find It hard to do anything that my parents would disapprove of. I'm likely to discontinue doing something that others think Is not worthwhile. When I take a stand I tend to hold on to It even If others disagree. WW I feel very relaxed before speaking In front of a group. I feel anxious when I speak In front of a group. i feel very nervous while performing In front of people. My voice never shakes when l recite In class. Sometimes my body trembles when i speak up in class. APPENDIX B Tables Table 8-1 Summary of the Results of the Conflrmatory Factor Analysis for Six A Priorl Shame Theme Clusters: Correlations between items and Factors (Decimals Omitted) Correlations of items (Grouped by Factor) with Factors DS I D W U FS Disappointment_ultb_§e11_LQ&1 55 80 72 74 74 69 60 89 73 69 67 66 64 61 19 68 63 68 58 56 69 15 64 6O 62 61 58 51 EQQiiflfl§_91_in1fiLlQLi11_Lil 85 80 79 78 71 69 67 51 67 79 69 71 72 56 51 74 78 75 72 72 63 B 58 65 56 57 53 45 EQQilnQ§_QL_Qn1§QLi1nnn§3_LDl 22 69 67 77 68 65 58 58 7O 72 72 68 68 62 23 65 58 67 53 51 64 52 48 47 48 39 35 37 EflQilDRE—QI_HQLJDJ£§SDQSS_Lfll 59 71 72 67 85 80 51 $9 75 75 70 84 78 52 51 69 68 68 81 76 52 15 64 63 61 76 71 45 $5 76 71 69 68 65 66 $1 63 65 6O 64 59 45 EQQLin9§_91_UnianLlan£Q_LUl 55 74 72 70 79 79 70 39 63 68 62 68 79 53 Elliln9_§hQLl_LE§l 33 63 57 64 51 54 75 55 65 58 62 51 50 75 162 163 Table 8-2 Summary Data for Outside Variables and Other Key Scales Number Standard Personality Scale of Items Mean Deviation Reliability Embarrassment 8 24.20 5.94 .86 Fear of Exposure 7 20.72 4.97 .80 Public Self-Consciousness 8 30.13 4.77 .81 Shyness 9 24.47 6.36 .84 Buss & Durkee Guilt 6 18.47 3.92 .66 Chang & Hunter Guilt 9 28.63 5.93 .81 Combined Guilt 15 47.10 8.91 .84 Beck Depression 13 18.92 5.83 .90 Splelberger Trait Anxiety 10 24.94 5.84 .84 Splelberger Trait Anger 10 28.38 6.19 .83 Loneliness 8 17.54 5.42 .85 Assertiveness 17 52.80 10.13 .84 Satisfaction with Life 5 16.52 3.87 .80 Resentment 3 9.03 2.16 .53 Conformity 6 22.51 3.82 .64 Egocentrism 3 9.65 2.42 .68 Exhibitionism 6 19.33 3.65 .68 Dominance 3 9.38 2.42 .78 Self-Revelation 3 10.49 2.54 .64 Competitiveness 3 7.28 2.28 .67 Need for Approval 7 18.88 4.05 .74 Audience Anxiety 5 17.17 3.80 .76 Rosenberg Self-Esteem 10 38.14 6.91 .90 Cheek & Buss Self-Esteem 6 22.69 4.38 .85 Cook Internallzed Shame 11 25.05 7.08 .90 164 Table 8—3 Correlations of Six Shame Theme Cluster Items with Outside Variables (Decimals Omitted) Outside_XAL1abias (see note) 5 6 7 8 9 1O 11 12 .5 N 0) fi 66 44 50 17 46 33 49 69 12 57 ~44 ~63 37 EQ 41 45 7 43 29 52 69 12 56 ~44 ~65 33 19 37 42 18 39 38 42 62 18 51 ~37 ~53 25 16 42 46 13 43 33 41 56 11 48 ~42 ~47 32 EfifiiinQ§_QI_JD1§LiQLl11 £5 48 55 17 50 37 49 67 10 58 ~51 ~58 35 31 43 45 15 48 35 47 56 11 52 ~47 ~48 27 54 47 5O 20 50 35 46 6O 13 59 ~48 ~54 37 8 35 41 15 42 30 38 50 2 47 ~44 ~47 19 Eeeiinps_91_nelactixeness 22 4O 47 11 44 33 47 66 15 62 ~38 ~54 37 88 42 52 17 47 36 48 62 15 57 ~41 ~51 38 23 45 49 23 36 33 31 57 19 46 ~36 ~46 37 52 34 33 19 30 25 26 42 7 32 ~31 ~42 24 Efifiiiflflfi.QI.flQL1fllfi§§D§§fi 59 42 45 6 46 30 54 61 12 59 ~40 ~51 39 69 45 49 7 46 34 49 63 14 60 ~45 ~53 29 51 36 40 5 41 29 47 61 12 55 ~38 ~49 36 15 32 36 ~2 4O 24 51 52 8 48 ~36 ~44 29 45 43 45 13 44 29 45 63 12 58 ~41 ~53 35 Al 32 4D ~2 41 22 49 59 5 54 ~41 ~49 22 Eeeiing§_91_unimpleance 56 42 48 11 48 29 49 65 14 63 ~43 ~55 39 82 41 47 7 48 32 48 56 11 61 ~41 ~47 39 EaIIInd_SnQL1_c1.0wn_§1andaLds_pL_LdeaLs 33 39 44 15 40 35 36 56 15 45 ~39 ~47 31 El 38 42 2O 36 33 36 57 18 44 ~35 ~48 28 Note: 1 - Embarrassment, 2 - Fear of Exposure, 3 - Public Self- Consclousness, 4 - Shyness, 5 - Guilt, 6 - Depression, 7 - Anxiety, 8 - Anger, 9 - Loneliness, 10 - Assertiveness, 11 - Satisfaction with Life, 12 - Resentment Table B-3 (cont'd) 165 Correlations of Six Shame Theme Cluster with Outside Variables Items (Decimals Omitted) Outsidn_¥aLiapins (860 DOtO) 13 14 15 16 17 18 19 20 Qisanpgintmeni_nitn_§9l1 99 28 38 ~30 ~31 ~27 ~1 38 25 99 26 33 ~35 ~36 ~34 1 41 28 19 36 38 ~21 ~24 ~16 ~3 33 26 26 31 37 ~23 ~30 ~23 ~7 37 24 Eneiinns_pi_in19LigLitx 85 31 43 ~34 ~39 ~29 ~2 46 29 91 30 40 ~30 ~35 ~30 —6 45 28 55 35 42 ~32 ~33 ~30 -1 49 30 a 27 35 ~29 ~36 ~29 ~9 39 29 Eeniings_91_nelectixeness 22 28 36 ~30 ~29 ~32 8 32 23 88 35 44 ~28 ~28 ~32 6 40 24 23 35 38 ~19 ~23 ~21 2 35 23 92 21 33 ~16 ~24 ~13 ~4 30 26 E99iJn99_91_WpL1ninssness 59 29 33 ~31 ~29 ~33 8 38 23 99 28 35 ~34 ~37 ~35 3 41 26 51 24 28 ~29 ~24 ~28 5 35 24 14 19 29 ~30 ~27 ~26 9 34 19 $5 26 34 ~30 ~34 ~25 ~5 38 31 41 22 25 ~35 ~35 ~33 9 34 22 Eeeiinas.91.fln1mpthance 99 28 33 ~32 ~30 ~34 6 41 22 99 28 36 ~31 ~33 ~34 1 34 22 Ea1i1np_5nQL1_Q1_an_91andALds_QL_ideais 23 35 32 ~21 ~25 ~22 -2 31 26 95 33 33 ~19 ~22 ~25 9 30 21 Note: 13 - Conformity, 14 - Egocentrism, 15 - Exhibi- tlonlsm, 16 - Dominance, 18 - Competitiveness, 20 - Audience Anxiety 17 - Self-Revelation, 19 - Need for Approval, 166 Table B-4 Correlations of Six Shame Theme Scales with Outside Variables (Decimals Omitted) DS I D W U FS Embarrassment 58 57 61 50 53 52 Fear of Exposure 64 64 69 56 60 57 Public Self-Consciousness 19 22 27 6 12 23 Shyness 6O 63 59 57 61 51 Buss & Durkee Guilt 59 53 60 45 46 53 Chang & Hunter Guilt 46 45 48 36 38 46 Beck Depression 65 60 57 64 62 48 Splelberger Trait Anxiety 90 78 86 78 77 75 Splelberger Trait Anger 19 12 21 14 15 22 Loneliness 75 72 75 73 79 59 Assertiveness ~58 ~63 ~55 ~53 ~53 ~50 Satisfaction with Life ~80 ~69 ~73 ~65 ~65 ~63 Resentment 45 39 52 41 49 39 Conformity 42 41 45 32 36 45 Egocentrism 51 53 57 40 44 43 Exhibitionism ~38 ~42 ~35 ~41 ~40 ~27 Dominance ~43 ~48 ~40 ~40 ~40 ~31 Self-Revelation ~35 ~39 ~37 ~39 ~43 ~31 Competitiveness ~4 ~6 5 6 4 5 Need for Approval 52 59 52 48 48 41 Audience Anxiety 36 39 37 ' 31 28 32 Rosenberg Self—Esteem ~96 ~92 ~89 ~92 ~85 ~73 Cheek & Buss Self-Esteem ~95 ~90 ~87 ~89 ~86 ~74 Cook Internallzed Shame 97 98 100 90 89 86 Note: 08 - Disappointment w Self, i - Feelings of Inferiorlty, D - Feelings of Defectiveness, W - Feelings of Worthlessness, U - Feelings of Unimportance, FS - Falling Short of Own Standards or Ideals 167 Table B-5 Summary of Conflrmatory Factor Analysis for Shame Profile Clusters: Correlations between items and Factors (Decimals Omitted) Correlations of Items (Grouped by Factor) with Factors CS RS DS 88 W 63 76 59 48 55 35 73 73 55 51 79 72 59 52 57 32 71 53 54 44 21 70 55 48 49 31 69 58 47 40 67 69 53 48 46 30 64 59 53 41 Belatipnsnip_§hame_1391 49 6O 78 46 43 12 51 71 36 37 37 50 70 4O 41 29 6O 66 46 39 25 61 63 46 43 27 54 63 39 33 Dependencx_9nama_L991 39 45 36 73 23 15 39 34 67 19 17 50 40 67 29 5 48 42 66 24 26 48 46 60 27 11 40 35 46 23 W 82 52 44 31 84 24 52 47 34 79 61 56 51 31 78 90 44 35 24 77 78 61 54 36 73 2 46 35 22 7O 86 33 26 19 67 72 55 49 30 64 Summary of Exploratory Factor Analysis on Shame Profile 168 Table B-6 Factor Loadings of Shame Profile Items: Items (Decimals Omitted) MaLimax_EaclpLs 1 2 3 4 119m_1 82 80* 17 13 16 90 79* 12 11 8 24 74* 17 17 2O 86 74* 1 11 5 61 65* 35 5 23 2 64* 29 2 9 78 59* 4O 9 26 72 51* 4O 3 24 63 3O 63* 22 24 21 24 61* 22 23 79 34 58* 27 21 67 22 56* 26 21 31 15 55* 26 28 32 2O 54* 35 19 3O 16 48* 32 33 25 19 43* 24 4O 39 7 17 70* 8 15 6 6 69* 13 5 8 16 63* 19 17 1O 29 61* 8 26 9 22 53* 23 11 8 21 38* 17 49 19 24 21 70* 37 21 13 18 68* 12 16 21 13 66* 27 9 28 17 58* 35 25 45 31 46* 29 15 39 25 46* Proportion of Variance' 1 2 3 4 0.16 0.14 0.11 169 Table B~7 Correlations of Shame Profile Items with Outside Variables (Decimals Omitted) Qu19L99_¥aL199199 (see note) 1 2 3 4 5 6 7 8 9 10 Campelen£9_9name 33 47 51 18 54 33 42 62 7 54 ~51 35 45 48 8 47 31 48 66 12 61 ~42 19 52 53 23 52 35 44 69 14 56 ~54 32 43 48 15 42 33 34 59 14 42 ~40 21 47 50 13 56 3O 4O 61 6 55 ~55 31 40 42 10 42 32 37 59 19 46 ~37 31 43 45 8 44 34 42 56 13 49 ~44 3Q 38 40 7 43 25 41 58 7 48 ~39 Relationship_sname $9 36 42 5 4O 29 51 58 13 58 ~34 12 26 32 3 34 15 37 52 13 56 ~28 31 29 37 4 34 23 42 51 12 53 ~29 23 40 47 11 47 28 37 6O 22 71 ~41 25 43 49 11 52 31 4O 6O 17 76 ~45 21 28 38 1 34 22 35 50 10 48 ~32 Qanendencx_§name 39 27 32 1O 29 35 34 47 15 36 ~28 15 22 26 8 22 33 26 39 18 29 ~23 11 33 35 18 35 34 32 47 9 39 ~39 5 29 33 13 32 33 33 49 16 40 ~31 25 32 38 15 34 37 29 49 11 40 ~34 11, 24 28 9 27 26 27 36 11 30 ~26 3991.391m9 32 38 35 19 34 27 31 40 9 30 ~36 2A 37 36 20 31 28 3O 42 14 32 ~32 31 43 41 20 39 30 33 46 11 39 ~39 SQ 37 33 28 27 28 26 37 11 24 ~29 13 43 41 16 43 34 38 49 9 42 ~42 2 36 35 19 36 21 25 37 6 32 ~35 33 30 24 20 21 23 21 27 11 17 ~23 Z2 39 39 7 41 28 32 43 ~2 44 ~47 Note: 1 - Embarrassment, 2 - Fear of Exposure, 3 - Public Self- Consclousness, 4 - Shyness, 5 - Guilt, 6 - Depression, 7 - Anxiety, 8 - Anger, 9 - Loneliness, 10 - Assertiveness Table B-7 (cont'd) 170 Correlations of Shame Profile items with Outside Variables (Decimals Omitted) 9919199_291139199 (see note) 11 12 13 14 15 16 17 18 19 20 Cam2919n99_933m9 33 ~55 30 26 37 ~39 ~41 ~28 -9 43 38 33 ~53 40 29 34 ~33 ~30 ~32 5 37 25 13 ~58 33 33 41 ~33 ~41 ~24 ~10 42 35 32 ~43 33 33 37 ~28 ~30 ~16 ~1 40 29 21 ~56 29 19 36 ~46 ~49 ~26 ~7 43 39 31 ~43 28 21 31 ~26 ~29 ~20 3 36 20 31 ~48 30 28 35 ~31 ~33 ~18 -8 38 31 33 ~45 23 18 33 ~38 ~37 ~26 2 40 26 891911999313_§nam9 59 ~49 32 20 28 ~28 ~19 ~38 12 29 16 12 ~43 25 15 19 ~25 ~24 ~38 9 22 15 31 ~45 24 19 25 ~20 ~15 ~35 12 22 10 29 ~46 34 30 35 ~30 ~34 ~35 5 31 21 23 ~51 34 30 36 ~31 ~34 ~36 ~3 26 22 21 ~42 25 21 21 ~32 ~27 ~38 8 25 17 3999nfi9n£l_finim9 33 ~31 29 25 27 ~15 ~18 ~12 4 34 15 13. ~20 24 28 24 ~11 ~10 -5 0 28 16 11 ~31 23 30 35 ~20 ~25 ~13 ~4 38 22 3 ~30 35 31 25 ~22 ~19 ~17 0 36 22 23 ~33 34 27 32‘ ~22 ~21 ~22 8 35 19 11 ~26 16 26 21 ~11 ~20 ~15 5 29 12 3931.339m9 32 ~36 25 19 36 ~19 ~23 ~16 ~7 26 21 25 ~36 28 23 38 ~17 ~20 ~18 ~2 25 17 31 ~43 28 25 42 ~19 ~25 ~20 ~2 27 22 33 ~37 25 24 39 ~16 ~21 ~15 ~5 23 21 13 ~45 29 23 38 ~29 ~31 ~27 ~5 31 26 2 ~41 27 21 31 ~23 ~25 ~18 -3 29 21 33 ~28 25 16 32 ~11 ~11 ~9 -4 20 11 12 ~46 22 18 28 ~37 ~43 ~30 ~13 28 29 Note: 11 - Satisfaction with Life, 12 - Resentment, 13 - Conformity, 14 - Egocentrism, 15 - Exhibitionism, 16 - Dominance, 17 - Self-Revelation, 19 - Need for Approval, 18 - Competitiveness, 2O - Audience Anxiety 171 Table 8-8 Correlations between Shame Profile and Outside Variables Competence Body Relationship Dependency Shame Shame Shame Shame Embarrassment .62 .51 .49 .44 Fear of Exposure .66 .48 .60 .51 Public Self—Consciousness .18 .25 .09 .20 Shyness .67 .46 .59 .47 Guilt (1) .55 .47 .45 .54 Guilt (2) .44 .37 .36 .53 Self-Esteem (1) ~.91 ~.68 ~.80 .61 Self-Esteem (2) ~.91 ~.67 ~.82 .62 Depression .58 .40 .59 .48 Trait Anxiety .86 .54 .81 .71 Trait Anger .16 .12 .21 .21 Loneliness .72 .44 .88 .57 Assertiveness ~.63 ~.48 ~.51 .48 Satisfaction with Life ~.71 ~.53 ~.67 .45 Resentment .43 .35 .43 .43 Conformity .36 .28 .33 .44 Egocentrism .50 .48 .40 .43 Exhibitionism ~.48 ~.29 ~.41 .27 Dominance ~.51 ~.34 ~.37 .30 Self-Revelation ~.34 ~.26 ~.54 .22 Competitiveness ~.04 ~.07 .11 .03 Need for Approval .56 .35 .38 .53 Audience Anxiety .43 .28 .24 .28 172 Table B-9 Error Analysis for a Hypothetical Hierarchical Model Subsuming Shame, Embarrassment, Fear of Exposure, Public Self-Consciousness, Shyness, and Guilt Higher Order Factor /..// \2 \.\ .. Fear of Embarrass- I Shyness Guilt Shame Ii Public ' Exposure ment 1L S-C j Elgure 12. A hypothetical hierarchical model subsuming shame-related traits, Including shame. WWW: M2113 231133135 FE E SHY G S PSC We: Fear of Exposure 1.00 .87 .82 .65 .65 .48 Embarrassment .87 1.00 .74 .61 .58 .46 Shyness .82 .74 1.00 .46 .62 .24 Guilt .65 .61 .46 1.00 .52 .47 Shame .65 .58 .62 .52 1.00 .18 Public Self-Consciousness .48 .46 .24 .47 .18 1.00 W Fear of Exposure 1.00 .90 .76 .71 .66 .45 Embarrassment .90 1.00 .70 .66 .61 .41 Shyness .76 .70 1.00 .55 .52 .35 Guilt .71 .66 .55 1.00 .48 .32 Shame .66 .61 .52 .48 1.00 .30 Public Self-Consciousness .45 .41 .35 .32 .30 1.00 £LL9L9 Fear of Exposure ~.O3 .06 -.O6 ~.01 .03 Embarrassment ~.03 .04 ~.05 ~.03 .05 Shyness .06 .04 ~.09 .10 ~.J] Guilt ~.06 ~.05 ~.09 .04 .15 Shame ~.O1 ~.O3 10 O4 ~.12 Public Self-Consciousness .03 .05 :111 115 :112 173 Table B~1O Summary of Results of Higher Order Conflrmatory Factor Analysis on Hobiltzelle (1987) Study Data (Decimals Omitted) 1 2 3 4 5 6 7 8 9 10 11 1. ASGS Guilt 100 ~24 35 ~10 14 3O 16 100 ~24 18 27 2. BSGT Guilt ~24 100 16 16 16 02 ~02 ~24 100 23 00 3. ASGS Shame 35 16 55 46 56 57 44 35 16 74 60 4. BSGT Shame ~10 16 46 39 46 30 29 ~10 16 62 35 5. FOAI 14 16 56 46 55 43 44 14 16 74 51 6. Zung Depression 30 02 57 3O 43 72 71 3O 02 61 84 7. Beck Depression 16 ~02 44 29 44 71 72 16 ~02 55 84 8. 501 100 ~24 35 ~10 14 3O 16 100 ~24 18 27 9. 502 ~24 100 16 16 16 02 ~02 ~24 100 23 OO 10. 503 18 23 74 62 74 61 55 18 23 100 69 11. 504 27 00 60 35 51 84 84 27 OO 69 100 uglg; ASGS - Adapted Shame/Guilt Scale (Glolella, 1981; Hobiltzelle 1982); BSGT - Beall Shame-Guilt Test (Smith, 1972); FOAI - Fear of Appearing Incompetent (Good & Good, 1973); Zung Depression - Self- ratlng Depression Scale (Zung, 1965). 501 Is ASGS Guilt; 502 Is BSGT Guilt; 503 Is higher order factor for three shame scales; 504 Is higher order factor for two depression scales. 174 Table B~11 Summary of Results of Higher Order Conflrmatory Factor Analysis on Harder & Lewis (1987) Study Data (Decimals Omitted) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 1. PFQ Shame 52 55 41 35 47 18 11 20 43 50 72 60 11 20 58 2. Self-Derogatlon 55 60 26 56 41 14 14 21 68 51 78 51 14 21 74 3. Self-Esteem 41 26 19 24 35 20 16 08 23 23 43 51 16 08 29 4. Self-instability 35 56 24 36 32 07 04 14 41 42 59 36 04 14 51 5. PFQ GuIIt 47 41 35 32 29 27 13 10 44 40 61 52 13 10 52 6. SUS GuIIt 18 14 20 07 27 29 08 27 14 04 23 52 08 27 11 7. Mosher Gullt 11 14 16 04 13 08100-42 11 11 18 19100-42 14 8. Sex 20 21 08 14 10 27-42100 12~09 25 34-42100 02 9. MAACL Depression 43 68 23 41 44 14 11 12 65 64 69 54 11 12 80 10. Beck Depression 50 51 23 42 40 04 11~09 64 65 66 41 11~09 80 11. 501 72 78 43 59 61 23 18 25 69 66100 78 18 25 83 12. 502 60 51 51 36 52 52 19 34 54 41 78100 19 34 58 13. 503 11 14 16 04 13 08100-42 11 11 18 19100-42 14 14. 504 20 21 08 14 10 27-42100 12-09 25 34-42100 02 15. 505 58 74 29 51 52 11 14 02 80 80 83 58 14 02100 uglg1 PFO Shame - Personal Feelings Questionnaire (Harder & Lewis, 1987); Self-Derogation - the Kapian Self-Derogatlon Scale (Kaplan & Pokorny, 1969); Self-Esteem (reflected) - Janis-Field Self—Esteem Scale (Eagly, 1967); Self-Instability - Rosenberg Instability of Self—Image Scale (1965); PFO Guilt - Personal Feelings Questionnaire (Harder & Lewis, 1987); SUS Guilt - Situational Upset Scale Guilt (Beall, 1973); Mosher Guilt - Mosher Total Guilt Scale (1968); Sex; 1 - male, 2 - female; MAACL Depression - Multiple Affect Adjective Check List Depression (Zuckerman & Lubin, 1965); 501 Is the higher order factor for PFO Shame, Self-Derogatlon, Self-Esteem (reflected), and Instability of Self-image; 502 Is the higher order factor for PFO Guilt and SUS Guilt; 503 is the Mosher Guilt; 504 Is the sex variable; 505 Is the higher order factor for MAACL Depression and Beck Depression. BIBLIOGRAPHY BIBLIOGRAPHY Abraham, K. (1911). Notes on the psychoanalytic Investigation and treatment of manic-depressive Insanity and allied conditions. In se1nsted_napeLs_9n_nsxsnp:anaixsis. pp. 137-156. London: Hosarth Press, 1927. Abraham, K. (1916). The first pregenltal stage Of the Ilbldo. in 9e1ected_naneLs_anp§xch:anaixsis. pp. 248-279- London: Hogarth Press, 1927. Abraham, K. (1924). A short study of the development of the libido, viewed In the light of mental disorders. In 3313£1§Q_2333L§_Qfl paygng;an11y§1§, pp. 418-502. London: Hogarth Press, 1927. Adler. A. (1929). Ine_science_pl_1111nn- New York: Groonberg- Adler, A. (1931). Wna1_111§_§ngu1g_mgan_19_1gu. Boston: Little, Brown. Alexander, F. (1938). Remarks about the relation of Inferiorlty feelings to gullt feelings. intennatipnai_quLna1.91.2915392AnaLsts. 19. 41-49. American Psychiatric Association. (1980). Q1agnQ§11g_ang_§1111511331 manuai_91_mentai_disQLdeLs (3rd ed.). washington. DC: Author. American Psychiatric Association. (1987). Q1agnQ§119_gng_§1311§11331 manu91_91_m9n191_3199L39L9 (3rd ed. rev.). Washington, DC: Author. Beall, L. (1972). Ing_§nam§;fiu111_lg§1. Unpublished manuscript, The Wright Institute, Berkeley, CA. Beck, A. T., 8 Beck, R. (1972). Screening depressed patients In family practice: A rapid technique. £9919L939919_M93191n9. 52(6), 81-85. Beck, A. T., Ward, C. H., Mendelson, M., Mock, J. E., & Erbaugh, J. K. (1961). An Inventory for measuring depression. Argnlyes 91 General EsxcnialLl. A. 561-571. Blbrlng, E. (1953). The mechanism of depression. In P. Greenacre (Ed.), A113§111§_fi1§QLflQL§. New York: International Universities Press. Blnder. J. L. (1970). Ine_Le1a11xe_nLcneness_te_shame_pL_nuiit_as_a a1mgn319n_91_gnaLag19L_§111§. Unpublished doctoral dissertation, 175 176 University of Michigan. Branden. N. (1969). 139W. Los AngeloS: Nash. Branden, N. (1983). figng;1ng_1ng_§gij. Los Angeles: J. P. Tarcher. Branden. N- (1987). HQH_1Q_L3133_XQHL_3311:33133M- New York: Bantam BOOKS . Briggs, S. R. , 8 Smith, T. G. (1986). The measurement of shyness. In w. H. Jones, J. M. Cheek. , & s. R. Briggs (Eds. ). §n139992 E9L9p9911x99_9n_L9999Lsn_ann_1L991m9n1 New York: Plenum Press Broucek, F. (1982). Shame and Its relationship to early narcissistic dOVBIODmentS- 1D13LD311Qfl31_JQHLD31_QI_Efil£flQ:AD31131§o 33. 369- 378. BUSS. A. H. (1980)- 3311:3Qfl3313H§fl3§3_iflfi_§33131_3fl11311- 53” Francisco: Freeman. Buss, A. H., & Durkee, A. (1957). An Inventory for assessing different kinds of hostility. J9uLna1_91_ansu111n9_fisxnn91991. 21. 343-349. Campbell, R. N. (1984). In9_n9u_9919n991.3911299199m_9919991991- Lanham, MD: University Press of America. Chang, S. S. (1988) Snam9_1119LA1uL9.L9119!1_Ihfinlflllsii_iflfi mgasnLgmgnl_a§pggLs. Unpublished manuscript, Michigan State University. Cheek, J. M., 8 Buss, A. H. (1981). Shyness and sociability. Journal of E9LsQna1111_and_9991a1_fisxsnninnx. 41. 330-339. Coen, S. J. (1986). The sense of defect. JQuLnai_91_1ng_AmgL1ggn Esxcngana1111£_Assgslation. 34. 47-67. Cook. D. R. (1985). Ine_inteLna11199_§name_sgale. Unpublished manuscript, University of Wisconsin, Menomonle. Cook, 0. R. (1988a). Development and validation of the Internallzed Shame Scale. Submitted for publication. Cook, 0. R. (1988b). The measurement of shame: the Internallzed Shame Scale. Paper presented at 1988 APA Convention, Atlanta, Georgia. Coopersmlth, S. (1967). Ing_Anjaggggn1§_91_5311;gfilg§m. San Francisco: Freeman. Crouppen. G. A. (1976). E1313_33D3D33D33:1n3333333D£3_1fl_333L3§333_333 . Unpublished doctoral dissertation, California School of Professional Psychology. 177 Diener, E., Emmons, R. A., Larsen, R. J., 8 Griffin, S. (1985). The Satisfaction Wlth Life Scale. J9uLna1_91_EeLsgna1111_Assessment. 53(1), 71-75. Fenlchel. O. (1945). Ine_ps19nQana11119.1399L1_91_neuL9313. New York: Norton. Fenlgstein, A., Scheler, M. F., 8 Buss, A. H. (1975). Private and public self-consciousness: Assessment and theory. W103 ann_911n1ca1_flsxcnoienx. 43. 522-527. Freud, S. (1953). Mourning and melancholla. In J. Strachey (Ed. and Trans.). In9_91an9aL3_931119n_91_1n9_99mn1919.9919991991991_99LK9 Qj_§1gmung_£Lgun (Vol. 14). London: Hogarth Press. (Original work published 1917) Friedman, M. (1985). Toward a reconceptuallzatlon of guilt. ggnlgmngLaLy Esxcnoanaixsis. 21(4). 501-547. Gaiassi, M. 0., 8 Gaiassi, J. P. (1980). Similarities and differences between two assertion measures: Factor analyses of the College Self—Expression Scale and the Rathus Assertiveness Schedule. 89n9119L91_999999m9n1. 2. 43-57. Harder, D. W., 8 Lewis, S. J. (1987). The assessment of shame and guilt. In J. N. Butcher 8 C. D. Splelberger (Eds.), Anxangg§_1n Qanggn11111_33§g§§mgn1. Vol. 6. Hlllsdale, NJ: Lawrence Eribaum Associates. Harrow, M., 8 Amdur, M. J. (1971). Guilt and depressive disorders. ALcn1xas_QL_GenaLai_Esxnniath. 29. 240—246. Hays, R. 0., 8 DIMatteo, M. R. (1987). A short-form measure of loneliness. JQuLna1_91_EaLsQna1111_Assessment. 51(1). 69-81. Hobiltzelie, W. (1987). Differentiating and measuring shame and guilt: the relation between shame and depression. In H. B. Lewis (Ed.), Ine_Lo19.91_sname_in_axmptem_19Lmation. Hlllsdale. NJ: Lawrence Eribaum Associates. Hunter, J. E. (1973). Methods of reordering the correlation matrix to facilitate visual inspection and preliminary cluster analysis. JQuLna1_91_Educatiena1_MeasuLement. 19. 51-61. Hunter, J. E. (1986). Mu111n19_99n9nn9n1_11L193199_1n_9xn9L1m99111 agaign. Workshop monograph presented at University of Iowa, iowa City, Iowa, October, 1986. Hunter, J. E. (1987). Multiple dependent variables In program evaluation. In M. M. Mark, 8 R. L. Shotland (Eds.), Multiple m91nQg§_1n_pngLam_§xa1uai1Qn. New Directions for Program Evalua- tion, no. 35. San Francisco: Jossey-Bass. 178 Hunter. J. E. (1988). EACKA9E_pn_1BM_malnlLam91_¥eL919n_21Q. Michigan State University, March 29, 1988. Hunter, J. E., 8 Gerbing, D. W. (1982). Unldlmenslonal measurement, second order factor analysis and causal models. in B.W. Staw, 8 L. L. Cummings (Eds.). WW (Vol. 4). Greenwich, CN: JAI Press. Hunter, J. E., Gerbing, D. W., Cohen, S. H., 8 Nicol, T. (1980). EAQEAQE 1933;_A_9x919m_91_E9L1Lan_L9u11n99_19L_1n9_nn911919_91 QQLL31311Qfl31_3313. User's manual prepared by Academic Computing Services, Baylor University, Waco, Texas. HUDtGr. R. F. (1984). ID3_1flfl3L_131331_Ihfi_33¥3133m3nl_3n3_131133113fl_31 3n_Q31331113_1331_13_331331_1h3_3L333fl33.3£.33§3fl33_31_1nlfiLflni g1aiggng. Unpublished doctoral dissertation, Michigan State University. Izard, C. E. (1971). Ing_Lagg_QL_ng119n. New York: Appleton-Century- Crofts. izard, C. E. (1972). EaLIeLns_91_emQ11Qn3;_A_new_anaixsis_91_anxietx_an9 ggpLg§§1gn. New York: Academic Press. Izard, C. E. (1977). Human_gm911gn§. New York: Plenum Press. Izard, C. E., 8 Hyson, M. c. (1986). Shyness as a discrete emotion. In- W. H. Jones, J. M. Cheek., 8 S. R. Briggs (Eds.), $n1n999; W. New York: Plenum Press. Jackson, M. R. (1984). fig1jzgglggm_nng_mggn1ng. Albany: State University Of New York Press. Jacobson, E. (1953). Contribution to the metapsychology of cyclothymlc depression. in P. Greenacre (Ed.). A11991119_3199L39L9. New York: International Universities Press. Jacobson. E. (1964). In9_s911_and_1n9_991991_wnLid. New York: international Universities Press. Jacobson, E. (1971). ngnLg§§1Qn. New York: international Universities Press. Jacobson, E. (1975). The regulation of self-esteem. In E. J. Anthony 8 T. Benedek (Eds.), DapLessinn_ann_numan_existence. Boston: Little, Brown. Joffe, P. (1984, Spring). Shana. Paper presented at the annual meeting of the Student Services of the Big Ten Conference, Ann Arbor, Michigan. Jones, W. H., Cheek, J. M., & Briggs, S. R. (Eds.). (1986). 3fl133§§1 179 E9Lsnact1xes_Qn_LeseaLch_and_1Leatment. New York: Plenum Press. Kaufman. G. (1985). 3nam91_In9_3399L_31.39L1fl9_1293_9311. Cambridge. MA: Schenkman. Kaufman. G. (In press). Ih9_3913h31331_3I_3ham91_Ih93L1_aD3_1L931m9n1_91 shame:based_sxndemes. New York: Springer- Klnston, W. (1987). The shame of narcissism. In D. L. Nathanson (Ed.), Ine_manx_1acas.91_sname. New York: Gullford. Klein, M. (1935). A contribution to the psychogenesls of manic- depressive states. in L3131_fln111_An3_L331L1113n_&_3139L_fl3LK§1 1321:1255. pp. 262-289. Delacorte Press, 1975. Klein, M. (1937). Love, guilt and reparation. In Lgxg1_gu111_ang L333La113n_&_3133L_!9LK§1_1321:1393- pp. 306-343. Delacorte Prose. 1975. Kohut, H. (1971). Ina analysis 91 tn: salt. New York: International Universities Press. Kohut, H. (1972). Thoughts on narcissism and narcissistic rage. E919n9ana11119_filunx_91_lne_flh113. 21. 360-400. Kohut, H. (1977). Ing_133191311gn_91_1ng_3311. New York: international Universities Press. Kohut, H. (1984). ng_ngg§_nn311§1§_guLg? Chicago: University of Chicago Press. Korpl, D. R. (1977). E919n9n191QL1sa1_a9n9919.91.9nam9_ann_nu111_99 1ung11gn§_g1_pg1111331_1ggg1gg1. Unpublished doctoral dissertation, California School of Professional Psychology. San Francisco. Korzybskl, A. (1933). 9c1en9a_and_sanitx;_An_1ntLQdu9119n_19 n9nzAL1st9te11an_sxst9ms_and_ganeLa1_semantiss. New York. Lancaster Press. Lehmann, H. E. (1985). Affective disorders: Clinical features. in H. I. Kaplan 8 B. J. Sadock (Eds. ). 3QmnL9h9n9119.1911333K_31_3§19_191L1/ 11. Vol. 1. Baltimore: Williams 8 Wilkins. Levin, S. (1967). Some metapsychologlcal considerations on the differentiation between shame and guilt. 1n13Ln3113fl31_JQnLfl31_Q1 EsxcnezAnaixsis. 93. 267-276. LGWIS. H. B. (1971). sname_and_gu111_1n_neuLpsLs. New York: International Universities Press. Lewis, H. B. (1979). Shame In depression and hysteria. In C. E. lzard (Ed.). Em911Qn9_1n_n9L9Qna1111_anfl_99x9n9991n9199x. New York: 180 Plenum Press. Lewis, H. B. (1981). Shame and guilt In human nature. In S. Tuttman, C. Kaye. & M. Zimmerman (Eds.). WWW finnLgflgh. New York: International Universities Press. LBWIS. H. 8. (Ed.). (1987a). In9_LQ19191.9nam9_1n_91m919m_19Lmilieu. Hlllsdale, NJ: Lawrence Erlbaum Associates. Lewis, H. B. (1987b). Shame and the narcissistic personality. In D. L. Nathanson (Ed.), Ing_man1_1ngn§_91_§nam§. New York: Guilford. Lynd, H. M. (1958). Qn_9nam9_an9Llhe_9eaL£n_19L_13991111. New York: Harcourt, Brace. Miller, s. B. (1985). In§_§namg_gxngg1gngg. Hlllsdale, NJ: The Analytic Press. Mirman, M. C. (1984) 3h3m3_3fl3_931111_A311¥319£31.3333313133 unc9nsc19us_dangeLs1_and_delenses Unpublished doctoral dissertation, Michigan State University. Mollon, P. (1984). Shame In relation to narcissistic disturbance. W. 51. 207-214. Mollon, P., 8 Parry, G. (1984). The fragile self: Narcissistic disturbance and the protective function of depression. BEIIISD J9uLna1_QI_Medica1_£axcnoingx. 31. 137-145. Morrison, A. P. (1983). Shame, Ideal self, and narcissism. anjganLaLy Esxcnganaixsis. 19(2). 295-318. Morrison, A. P. (1984). Working with shame in psychoanalytic treatment. JQuLna1_Q1_Am9LL9an_E9xennana1111£_Assgcialign..92. 479-505. Morrison, A. P. (1987). The eye turned Inward: Shame and the self. in D. L. Nathanson (Ed.), In§_many_jggg§_91_3namg. New York: Guilford. Nathanson, D. L. (1987). A timetable for shame. in D. L. Nathanson (Ed.). In9_manx_11999_91_9nam9. New York: Guilford. Negri. T. F. (1978). WWW ngggnal111. Unpublished doctoral dissertation, Michigan State University. Novak, W. (1986). Ine_maasuLemen1_and_cnnsiLu91_91_sname. Unpublished doctoral dissertation, Michigan State University. Perlman, M. (1953). W913: sjugx_QL_an11911. Unpublished doctoral dissertation, University of Chicago. 181 Piers, G., 8 Singer, M. (1953). Sham9_and_nu111. New York: Norton. Prosen, M., Clark, D. C., Harrow, M., 8 Fawcett, J. (1983). Guilt and conscience In major depressive disorders. AmgL1gan_JQuLna1_g1 Esxnnialnx. 199. 839-843. Rado, s. (1928). The problem of melancholla. 1n1gLn1119na1_JguLnn1_g1 W. 9. 420-438. Rathus, S. A. (1973). A 30~Item schedule for assessing assertive behavior. 923W. 4. 398-406. Reich, A. (1960) Pathological forms of self—esteem regulation. MW. 15. 215-32. Rosenberg. M. (1965). WWW- Princeton. NJ: Princeton University Press. Rosenberg, M. (1979). angg1y1ng_1ng_3311. New York: Basic Books. Russell, 0., Pepiau, L. A., 8 Cutrona, C. E. (1980). The revised UCLA loneliness scale: Concurrent and discriminant validity evidence. W. 39.. 472-480 . SChnarCh. D. M. (1976). ID3_LQ13_31_D3L33fl31111.1D1m19L31fl3133333113fl- Unpublished doctoral dissertation, Michigan State University. Schneider, C. D. (1977). snamg1_gxngauLg_ang_nL1xagx. Boston: Beacon Press. Schneider, C. D. (1987). A mature sense of shame. In D. L. Nathanson (Ed.). Ihe_manx_Laces_91_sname. New York: Guilford. Smith, R. L. (1972). Ina_Le1a11xa_nLcneness_to_sname_oL_guiit_as_an 1ng1ga1oL_91_ngjgn§1xg_§1y1g. Unpublished doctoral dissertation, Northwestern University. Splelberger, c. 0. (1983). W911 W181. Paio AItO. CA: Consulting Psychologists Press. Splelberger, C. 0., Gorsuch., 8 Lushene, R. E. (1970). ManuaL_1gL_1ng MWW. Palo Alto, CA: Consulting Psychologists Press. Splelberger, C. 0., Jacobs, G., Russell, 5., 8 Crane, R. S. (1983). Assessment of anger: The State-Trait Anger Scale. in J. N. Butcher 8. C. 0- Splelberger (Eds.). WM (Vol. 2). Hlllsdale, NJ: Lawrence Erlbaum Associates. TomkinS. s. s. (1963). Wm (Vol. 2). lbs mums. New York: Springer. 182 Tomklns, S. s. (1987). Shame. in D. L. Nathanson (Ed.), Ing_man1_1acgs 91_§nam§. New York: Guilford. Tryon, R. C. (1937). Q13313L_An31131fi. Ann Arbor, MI: Edwards Brothers. Tryon, R. C., 8 Bailey, 0. E. (1970). Q1u§1gL_an111§1§. New York: McGraw-Hlll. Wells, L. E., 8 Marwell, G. (1976). 5311:35133m, Beverly Hills: Sage. Wurmser, L. (1981). Ing_mask_91_§nam§. Baltimore: Johns Hopkins University Press. Wurmser, L. (1987). Shame: The veiled companion of narcissism. In D. L. Nathanson (Ed.). Ina_man1_1aces_91_sname. New York: Guilford. "‘ilimiilmB