tau...“ . ,1x.....;.. . .5 . .. . , . . rap. “5.1:, 111‘: 1.5011! .3. lust-int... 5‘ t 3":- flvwé?! . . 31.3.1145. 5t... .3. . » 2.22.... I :ii THESlS / This is to certify that the dissertation entitled Shame, Pride, and Substance Abuse Proclivity In Gay Men: A Theory-Based Investigation presented by Daniel W. Socall has been accepted towards fulfillment of the requirements for Ph. 1). degree in Counseling Psychology VIE—“'7‘ M, g! {33* Major professor Date June 13, 1995 MS U is an Affirmative Action/Equal Opportunity Institution 0-12771 3 1293 O‘Q‘WW 351 lCHlGAN STATE YNNERSITY BRARI .1 Milli\lili‘lli Willi Milli LIBRARY Michigan State University PLACE It RETURN BOX to romovothio checkout from your record. AVOID FINES Mum on or baton data duo. DATE DUE DATE DUE DATE DUE MSU is An Affirm-tin AotionlEqnl Opportunity trunnion WM! SHNE. PRlD Di SHAME, PRIDE, AND SUBSTANCE ABUSE PROCLIVI‘I'Y IN GAY MEN: A THEORY-BASED INVESTIGATION By Daniel W. Socall A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Counseling, Educational Psychology & Special Education 1995 SHAVE. PR1: The cons recently in psycl homosexuality Identity-hm 5 have high levels mfllmatoq fa: Sham is discerr ABSTRACT SHAME, PRIDE, AND SUBSTANCE ABUSE PROCLIVITY IN GAY MEN: A THEORY—BASED INVESTTGATTON By Daniel W. Socall The construct of shame and shame theory have been receiving attention recently in psychology. This study explored the construct of shame, its links with homosexuality, and how shame specific to gay identity affects gay men. The Gay Identity-based Shame and Pride Scales developed for this study were found to have high levels of internal consistency. Through the use of exploratory and confirmatory factor analysis, the present findings revealed that identity-based shame is discernible from self-esteem, identity-based pride, and positive and negative afibctivity. Although a majority of shame theorists posit that shame and pride are opposite ends of a single continuum, this study found that identity shame and identity pride, although moderately related, are different constructs. An original issue of this study based on a review of the literature was, “what accounts for the relatively high rate of substance abuse in gay men?”, with the supposition that gay shame may play a key role. Contrary to expectations, gay men in this study reported about the same rate of substance abuse proclivity as do men in *he gene? save as a signif: beam gar pri» measurement pr. proclivity. Imp? panic: are disc men in the general population. However, gay identity-based shame did appear to serve as a significant predictor of substance abuse proclivity. The relationship between gay pride and substance abuse is still unclear, and this may be due to measurement problems with the MacAndrew scale as an index of substance abuse proclivity. Implications of these findings for further research and clinical practice are discussed. Copyright by DANIEL W. SOCALL 1995 This st. ndy and whc: AIDS and had _: those lives hax sincerely dedic; DEDICATION This study is dedicated to the twenty-one men who were selected for this study and whose families or friends wrote back to state that they were victims of AIDS and had passed away. To these men, and the thousands of other gay men whose lives have been, and will be, cut short by the AIDS virus, this work is sincerely dedicated. There l". the completion shall try to be i: First it. has lived Ihror;f gaduare stride:I from approach '. on my mm issr. 81" man. He h: there I’m sure ”P on me, His 5 page Oflhls we; sure I could do To Dr (0“ch Olshar: as he pain 0for. tobecoming a h} DY. LincflI About this projek: sh , v . “I andher, vi ACKNOWLEDGMENTS There have been so many pe0ple who have contributed to the study and to the completion of this project that I am fearful I may forget to thank everyone. I shall try to be inclusive. First, many, many special thanks go to my advisor, Dr. Robert Lent. He has lived through “the many faces of Dan” throughout my seven year tenure as a graduate student in the program. He has patiently watched me cycle as I moved from approach to avoidance to approach of both he and the program, as I worked on my own issues of shame and pride as a developing psychologist and as an adult gay man. He has seen me at my best and at my very worst, and yet he has hung in there. I’m sure at times he was perplexed and had his doubts, but he never gave up on me. His guiding fingerprints are evident in every paragraph, and on every page of this work. He helped me to see that I could accomplish things I wasn’t sure I could do. He helped me to see areas of strength and independence thatI couldn’t see before. To Dr. Gersh Kaufman, my heartfelt thanks for introducing me to the concept of shame. His writing has helped me to understand my own pain as well as the pain of others. It has allowed me to begin to heal and has pushed me closer to becoming a healer myself, a title that I hope to have earned one day. Dr. Linda Forrest has helped in ways that go way beyond her insights about this project. She was able to spot my “scientist-practitioner identity-based shame” and helped me name and articulate it. She gave me the opportunity to stop and really the way I had 0 mill me. My war phone call from foruking a cha reject I also w. University C our been borh menti times about my 1,500 men In th' encouraElement balance. collate, Chmllne Olson Bernardo). It w Thanks to Dr, pg ROS and DI. To Mal) SICImi I 1 Finally‘ I I me to offer Supp mthank her I I0 my Ann: Ma. gm them all m" vii stop and really think about my developing identity as a psychologist, and about the way I had come to see myself as small and ineffectual. Her lessons will go with me. My warmest regards go to Dr. Irv Lehmann, who got a long-distance phone call from me begging for help - psychometrically speaking! Thanks to him for taking a chance on me, and what a wonderful contribution he has made to this project. I also would like to thank all the efforts and support given to me by the University Counseling Center at Colorado State University. Dr. Patricia Vigil has been both mentor and friend, listening to me talk about shame for hours - often times about my own. Megan Coover, who helped me type address labels for the 1,500 men in the study - bless your fingers. And thanks for all the notes of encouragement and chocolate! Also thanks to those who helped me counter- balance, collate, staple stuff and mail the 1,500 survey packets (Patty Vigil, Christine Olson, Megan Coover, Janine D’Anniballe, Jeannette Vanhook and Ken Bernardo). It was a huge undertaking and I could have never done it alone. Thanks to Dr. Randy Swaim of the Tri-Ethnic Center at CSU for his help with the confirmatory factor analyses of the study. Also thanks go to Dr. Lisa Thomson- Ross and Dr. Tom Ross for all their encouragement and CF A hints. And to Dr. Mary Stelma, I thank her dearly for her support during my eleventh-hour changes. Finally, very special thanks go to my farniiy, who were always there for me to offer support. To my Mom, although she cannot completely accept who I am, I thank her for always trying her best to understand me and be there for me. To my Aunt Marilyn and Uncle Paul who are my “adopt ” mother and father, I give them all my love for unconditionally accepting me. Also, my gratitude for helping me to :- m of consider Bernardo. let r: you for the mes viii helping me to finance the bulk of the copying and mailing of this survey, which was of considerable expense. And to my life-partner and soul-mate, Ken Bernardo, let me say that I’m so proud to be gay and in love with you! Thank you for the most precious gift of all - spending your life with me. LIST OF T. LIST OF P] CHAPTER Introductior Statement CHAPTER Literature R Relations'r Shame a Shame a Shame a Shame Th: AlTect T ACtivato Innate Interpe Cultur; Intemaliz Sham Shame‘ Prir Ga." Ideal Gay Iden: Shame as at m Gay Men Shame. A Slimman. 0 Research. or Relations; RElationsl'l and PA\'. redlCIlVe TABLE OF CONTENTS LIST OF TABLES ............................................................................................ xii LIST OF FIGURES .......................................................................................... xiv CHAPTER ONE ................................................................................................... 1 Introduction .......................................................................................................... 1 Statement of Problem ........................................................................................ 1 CHAPTER TWO .................................................................................................. 5 Literature Review ................................................................................................. 5 Relationship of Shame to Other Psychological Constructs ............................... 6 Shame and Guilt ............................................................................................. 6 Shame and Pride ............................................................................................ 8 Shame and Self-Esteem ............................................................................... 10 Shame Theory ................................................................................................. 13 Affect Theory ............................................................................................... l4 Activators of Shame ..................................................................................... l6 Innate Activators of Shame ...................................................................... 16 Interpersonal Activators of Shame ........................................................... 17 Cultural Sources of Shame. ...................................................................... 18 Intemalization of Shame ............................................................................. l8 Shame Binds ........................................................................................... l9 Shame, Pride, and Homosexuality .................................................................. 20 Gay Identity-Based Shame .......................................................................... 21 Gay Identity-Based Pride ............................................................................. 23 Shame as an Etiologic Agent of Increased Levels of Substance Abuse in Gay Men ...................................................................................................... 25 Shame, Alcoholism, and Gay Men .............................................................. 27 Summary of Literature .................................................................................... 28 Research Questions ......................................................................................... 29 Relationship of Shame to Pride ................................................................... 29 Relationship of Gay Identity-based Shame and Pride to Self-Esteem and PANAS .................................................................................................. 3O Predictive Utility of Identity-Based Shame ................................................. 30 CHAPTER THREE ............................................................................................ 31 METHOD ........................................................................................................... 31 Pilot Phase .......................................................................................................... 31 Respondents ........................................................................................................ 32 Design ................................................................................................................. 34 Procedure ............................................................................................................ 34 Measures , SUIC‘ICCI D: Marlowe-l Cray Idem: Rosenberg hlacAndre Positive ar ANALYSIS Pilot Phas: Main Sm:- Relation Prepar Furthe Relatior.- Relation: Explor Confr Comp} Utility 0: Cross l'ali.‘ CHAPTER r Results ...... Pilot Phase Rfiponsd Reiiabili: Main SIUdy Repose R Return SaIT‘ DemOgra Smdl’ Vane DllleienCQS EXPIOI'RIOrt It: ........ HYPOIIICSCS Conelalic Relatio, Relatio; "_ _ _ —— TABLE OF CONTENTS - Continued Measures ............................................................................................................. 36 Subject Demographics and Assessment of Sexual Orientation ...................... 36 Marlowe-Crowne Social Desirability Scale - Form C. ................................... 37 Gay Identity-based Shame-Pride Scales ......................................................... 38 Rosenberg Self-Esteem Scale .......................................................................... 4O MacAndrew Alcoholism Scale (MAC) ........................................................... 41 Positive and Negative Affectivity Scale- (PAN AS) ....................................... 42 ANALYSIS ........................................................................................................ 45 Pilot Phase ....................................................................................................... 45 Main Study ...................................................................................................... 45 Relationship of Shame to Pride ................................................................... 46 Preparation of Gay Identity-based Shame and Pride Scales for Further Analysis. ...................................................................................... 47 Relationship of Gay Identity-based Shame to Gay Identity-based Pride ....48 Relationship of Identity-based Shame-Pride to Self Esteem and PANAS .49 Exploratory Factor Analysis Method. ...................................................... 49 Confirrnatory Factor Analysis .................................................................. 50 Comparisons of Fit for Several CFA Models ........................................... 52 Utility of Identity Based Shame and Pride .................................................. 53 Cross Validation of the Study ......................................................................... 54 CHAPTER FOUR .............................................................................................. 56. Results ................................................................................................................ 56 Pilot Phase ....................................................................................................... 56 Response Rate .............................................................................................. 56 Reliability Analysis of the Gay Identity-based Shame-Pride Scales ........... 57 Main Study ......................................................................................................... 58 Response Rate ................................................................................................. 58 Retum Sample Characteristics ........................................................................ 59 Demographics .............................................................................................. 59 Study Variables ............................................................................................... 60 Differences Among Demographic Groups ...................................................... 62 Exploratory Factor Analysis of the Gay Identity-based Shame and Pride Scales ............................................................................................................... 67 Hypotheses and Analyses ................................................................................ 7O Correlation Analysis And Exploratory Factor Analysis .............................. 70 Relationship of gay identity-based shame to gay identity-based pride. ...70 Relationship of identity-based pride to self esteem and PANAS ............. 71 Confirmat Relation based or Utility of i Additional R Cross Valid; Split-Sam CHAPTER PI DISCUSSION Response Rate Differences Cay Identity The Relatior Relation of 1 her Psych Gay Identity-t I’mcliyity Implications Directions 1'. Conclusion... Principle Pi Implication TABLES FIGLRES ..... APPENDICE APP‘mdix A 5 Appendix B 3 Appendix c, < AI’llendiir D- 1 Appendix E ) Appendix 1:, r Appendix H I TABLE OF CONTENTS - Continued Confirmatory Factor Analysis ..................................................................... 77 Relationship of gay identity-based shame to gay identity- based pride, self-esteem and PANAS ....................................................... 77 Utility of identity-based shame and pride .................................................... 83 Additional Regression Analysis ...................................................................... 87 Cross Validation Analysis of the Study .......................................................... 88 Split-Sample Regression Analyses .............................................................. 89 CHAPTER FIVE ................................................................................................ 91 DISCUSSION ..................................................................................................... 91 Response Rate, Generalizability of Findings, and Sample Demographics ........ 91 Differences Among Demographic Groups ...................................................... 96 Gay Identity-based Shame Pride Scales .......................................................... 98 The Relation of Gay Shame to Gay Pride ..................................................... 100 Relation of Gay Identity-based Shame and Pride to Other Psychological Constructs .................................................................... 101 Gay Identity-based Shame and Pride as Predictors of Substance Abuse Proclivity ......................................................................................................... 104 Implications for Counseling .......................................................................... 107 Directions for Future Research ..................................................................... l 10 Conclusion ........................................................................................................ 1 l3 Principle Findings ......................................................................................... 1 l3 Implications for the Field ............................................................................. 114 TABLES ........................................................................................................... 1 16 FIGURES ......................................................................................................... 157 APPENDICES ........................................................................................................ Appendix A: Subject Demographics ................................................................ 165 Appendix B: Marlowe-Crowne Social Desirability Scale - Short Form C ...... 167 Appendix C: Gay Identity-based Shame - Pride Scales ................................... 168 Appendix D: Rosenberg Self-Esteem Scale ..................................................... 172 Appendix E: MacAndrew Alcoholism Scale (MAC) ....................................... 173 Appendix F: Positive and Negative Affectivity Scales (PANAS) ................... 175 Appendix G: Cover Letter ................................................................................ 176 Appendix H: Pre-Survey Postcard .................................................................... 177 Appendix I: Post-Survey Postcard .................................................................... 178 Appendix I: Results Requested/Not Interested Postcard .................................. 179 BIBLIOGRAPHY ................................................................................................ 180 .-J Lo) Positive and Prevalence I . Correlations and the Ho; Inventory (I . Response R2 . Interdtem C- . Inter-item C. . Gay Identity Variances xii LIST OF TABLES Table 1. Positive and Negative Affects with Corresponding Facial Behaviors ........... l 16 2. Prevalence Rates of Substance Abuse in Gay Men ...................................... 117 3. Correlations Between the Positive and Negative Affect (PANAS) Scales and the Hopkins Symptom Checklist (HSCL), Beck Depression Inventory (EDI), and STAI State Anxiety Scale (A-State) ........................... 118 4. Response Rate Breakdown for Pilot Phase .................................................... 119 5. Inter-item Correlations for the Gay Identity-based Shame Scale .................. 120 6. Inter-item Correlations for the Gay Identity-based Pride Scale ..................... 121 7. Gay Identity-based Shame Scale Item-Scale Correlations, Item Means, Variances and Communalities ...................................................................... 122 8. Gay Identity-based Pride Scale Item-Scale Correlations, Item Means, Variances and Communalities ...................................................................... 123 9. Response Rate Breakdown for Main Study ................................................... 124 10. Frequency and Percentages for Demographic Information Age through Relationship Status ........................................................................................ 125 11. Frequency and Percentages for Demographic Information Current Orientation through Comfort with Orientation .............................................. 126 12. Means and Standard Deviations for Psychological Measures of the Study... 127 13. Correlation Matrix for All Scale Scores and Internal Consistencies for All Scales ............................................................................................................ 128 14. T-Tests for Categorical Demographic Variables with Shame, Pride and MAC as Dependent Variables ................................................................................ 129 15. T-Tests for Categorical Variables with Self-Esteem, and PANAS as Dependent Variables ................................................................................... 130 16 AVOH fc hI-lC. Sitar 17 Rotated Fa; Shame-Fae 18 Rotated Fa: I PrideoIactg 19 Comparison 2!. Rotated f ac Gay Identity 2 Rotated F Affocnvity 23. Rotated Fa based Pride 24 Rotated Fa Affectiviry 25 ROIaICd F 3C1 Rosenberg s I Conclations Stud! Vanar ”WU FaCtt 28‘ CFAL Fiye f Efior of the Indicaje‘ Po xiii LIST OF TABLES - Continued 16. ANOVA for Demographic Variables: Sexual Relationship Status with MAC, Shame, Pride, Self-Esteem, and PANAS as Dependent Variables ..... 131 17. Rotated Factor Loadings for Gay Identity-based Shame and Pride Scales: Shame-Factor ................................................................................................ 132 18. Rotated Factor Loadings for Gay Identity-based Shame and Pride Scales: Pride-Factor .................................................................................................. 133 19. Comparison of Gay Identity-based Shame and Pride Scales as Parent Scales with Shame-F actor and Pride-Factor ............................................................. 134 20. Factor Loadings of All Scale Scores .............................................................. 135 21. Rotated Factor Loadings of all Scale Items for Factor One: Gay Identity-based Shame ............................................................................. 136 22. Rotated Factor Loadings of all Scale Items for Factor Two: Negative Affectivity ..................................................................................................... 137 23. Rotated Factor Loadings of all Scale Items for Factor Three: Gay Identity- based Pride .................................................................................................... 138 24. Rotated Factor Loadings of all Scale Items for Factor Four: Positive Affectivity ..................................................................................................... 139 25. Rotated Factor Loadings of all Scale Items for Factor Five: Rosenberg SES ............................................................................................... 140 26. Correlations of Shame and Pride Factors from EF A and CF A with All Other Study Variables ............................................................................................... 141 27. Overall Goodness-of-Fit Indices for a Five Factor Model Versus a Two Factor Model ...................................................................................... 142 28. CFA: Five Factor Model - Maximum Likelihood Estimates, Standard Error of the Estimate, T Value and Factor That Each Item was to Indicate, For All Survey Items ...................................................................... 143 29. CFA: Five Factor Model - Factor Loadings of all Scale Items: Factors 1,2, and 3 .............................................................................................................. 144 30. CPA. Five Factors 4 31. CPA Corr: 32w CPA Two Error of ti- Indicate. P 33. CPA. Two Seifestee: 34. CPA; Two Shame and 35. CPA Cone 36 Chi-square the Two P a 37. Multiple Hi Proclivity S Identity-b a EPA and C] 39' Muklple Hie Pfoclivigr 5‘. CPA Five P; 40‘ Overall Goo. PiVe FaCIOr Model Split- AIOOIIOIISm p] xiv LIST OF TABLES - Continued 30. CFA: Five Factor Model - Factor Loadings of all Scale Items: Factors 4 and 5 .............................................................................................. 145 31. CF A Correlation Matrix for Five Factor Model ............................................ 146 32. CFA: Two Factor Model - Maximum Likelihood Estimates, Standard Error of the Estimate, T Value and Factor That Each Item was to Indicate, For All Survey Items ...................................................................... 147 33. CFA: Two Factor Model - Factor Loadings of all Scale Items for Self-esteem, Positive Affectivity and Pride: Factors 1 and 2 ....................... 148 34. CFA: Two Factor Model - Factor Loadings of Composite Scale Items for Shame and Pride: Factors 4 and 5 ................................................................. 149 35. CFA Correlation Matrix for Two Factor Model ............................................ 150 36. Chi-Square Difference Test of the Five Factor Model with the Two Factor Model ................................................................................... 151 37. Multiple Hierarchical Regression Analysis With MacAndrew Alcoholism Proclivity Scale as Dependent Measure ......................................................... 152 38. Comparisons of Correlations of All Scale Scores Comparing Full Gay Identity-based Shame and Pride with Shame-F actor and Pride-Factor From EPA and CF A ................................................................................................ 153 39. Multiple Hierarchical Regression Analysis With MacAndrew Alcoholism Proclivity Scale as Dependent Measure Using Items from CFA Five Factor Model ................................................................................. 154 40. Overall Goodness-of-Fit Indices for a Five Factor Model Compared to the Five Factor Model Split-Samples ...................................................................................... 155 41. Split-Sample Multiple Hierarchical Regression Analysis With MacAndrew Alcoholism Proclivity Scale as Dependent Measure ....................................... 156 Figure 1. Reg: :1 Figure 2. Scree of the Figure 3. Scree of all .I Figure 4. Five 5. Figure 5. HistogL Figure 6_ Scatte the re; Figure 7. Norm; LIST OF FIGURES Figure 1. Regression model for the study ..................................................... 158 Figure 2. Scree plot of eigenvalues from principle components EFA of the Gay Identity-based Shame and Pride Scales ....................... 159 Figure 3. Scree plot of eigenvalues from principle components BE A of all items from all scales ............................................................. 160 Figure 4. Five factor confirmatory factor analysis model .............................. 161 Figure 5. Histogram of the residuals for regression analysis ......................... 162 Figure 6. Scatterplot depicting homoscedasticity of residuals of the regression equation .................................................................. 163 Figure 7. Normal probability plot of standardized residual ........................... 164 "A pen fundamentally E being" (Possum Pew SOC 511011eg usocia of the post-m: 41. our-”Cl“? (Kus PlaFETOUnd can 'faES’W Famil Samegendered . CHAPTER ONE Introduction ". . . when friends by shame are undefiled how can I keep from singing?" (Enya, 1991) "A pervasive sense of shame is that ongoing premise that one is fundamentally bad, inadequate, defective, unworthy, or not firlly valid as a human being” (Fossum & Mason, 1986, p. 5). W Few socially stigmatized attributes and conditions exist today that are as strongly associated with shame as is homosexuality (Nathanson, 1992). Examples of the powerful, negative shaming force of homosexuality are broadly evident in our society (Kus, 1988; Weinberg & Williams, 1974). Children playing on the playground can think of no put-down more painful than to call one another "faggot." Families still disown sons and daughters when they discover that their same—gendered "roommate" is a lover (Hammersmith, 1987; Hetrick & Martin, 1987). Sex between consenting adults of the same gender is still illegal in many states (Hetrick & Martin, 1987). Recently, the people of the state of Colorado voted to change that state's constitution, making it the first in United States history to allow legal discrimination of people on the basis of sexual orientation (Coloradoan, November 6, 1992). In most states, same gendered mates cannot legally marry, adopt children, touch in public, or carry insurance benefits on their partners (Hetrick & Martin, 1987). They may also have no legal claims to share property, and may not be able to plan for, or attend, their partner’s funeral (Hetrick & Martin. 198' state of C olora. is to feel that :3: that Others Wiii Althoug‘ toll on gay mer usually as dest' outside is now I this gay identirj Numerous auth. pain of internal. Nathanson, 19c estimated that n abuse issues—p,t AICOIIOI Abuse men may be the Important cIUIESI: “a“ “ml-gays? l0 mtmaIIZed S. tapioca”). line: I Cmenrlx me in their ide nihi & Martin, 1987). Society's general message (as very poignantly depicted by the state of Colorado) is clear: to be gay is shameful. To feel shame about being gay is to feel that there is something inherently wrong with oneself and to live in fear that others will discover this. Although society's negative views of homosexuality exact an enormous toll on gay men, the effects of the internalized shame that gay men carry are equally as destructive (Nathanson, 1992). The pain that originally came from outside is now the pain that comes from within (Kaufman, 1989, 1992). Because this gay identity-based shame is from within, its effects are often covert. Numerous authors contend addictive behaviors are a means of coping with the pain of internalized shame (Fossum & Mason, 1986; Kaufman, 1989, 1992; Nathanson, 1992). Researchers (see, for example, Weinberg, et a1., 1974) have estimated that nearly thirty percent of openly gay men struggle with substance abuse issues—nearly three times the average of non-gay men (National Institute on Alcohol Abuse and Alcoholism, 1978). Whereas the causes of alcoholism in gay men may be the same as the complex variables found in non-gay men, an important question is, why are gay men using alcohol and drugs at a higher rate than non-gays? One possible answer is that this pain-numbing behavior is related to internalized shame that is culturally-induced. Although some gay men may be utilizing substances to control their internalized shame, this link has not been empirically investigated. Currently, many gay men are empowering themselves with a sense of pride in their identities as gay and male. Initiatives to end the ban on gays in the military and to expand gay rights are examples of gay pride (Shilts, 1993). To have the pride mugh their is 1992). This in associated will" less public. int: shame has bee: effects of the ir In sum. Kaufman. 1% identity, shame to the self (Kat homosexual is I to shame theorj. difficulties, incl [WWII 198s through Pride, i (Kmfmm, 193...; Pride can d€\'eI.| Itsolved. Thls Stuc | and idendn'dlicr ‘daUOHShip of r. \ have the pride to confront societal oppression, gay men must first begin to work through their internalized sense of shame about being homosexual (Nathanson, 1992). This internalized shame is gay identity-based shame and is the pain that is associated with being homosexual. Along with a demand for political action, a less public, internal sense of pride may develop as the pain of gay identity-based shame has been addressed. This sense of pride about being gay can ameliorate the effects of the intemalized shame associated with the culture (Nathanson, 1992). In sum, in our culture homosexuality is strongly associated with shame (Kaufman, 1989; Nathanson, 1992). Because homosexuality is an aspect of identity, shame becomes bound (i.e., psychically connected), to various degrees, to the self (Kaufman, 1989; Nathanson, 1992). When shame about being homosexual is bound to the self it is called gay identity-based shame. According to shame theory, the results of these identity based shame binds lead to various difficulties, including substance abuse (Bradshaw, 1988; Possum & Mason; 1986; Kaufman, 1989, 1992 ). The negative effects of shame may be ameliorated through pride, i.e., developing a positive, healthy self-image as a gay male (Kaufman, 1989, 1992; H. B. Lewis, 1987; M. Lewis, 1992; Nathanson, 1992). Pride can develop as shame-binds to gay identity are addressed and become resolved. This study provides an opportunity for exploration of the newly researched construct of shame, and the potentially powerful relation of identity-based shame and identity-based pride issues to adjustment in gay men. Further, the relationship of identity-based shame and pride to other theoretically related cansuucts wil'. clarifying die ; the study may insights for pc constructs will be clarified as will the predictive utility of these constructs in clarifying the question of higher rates of substance abuse in gay men. In addition, the study may have implications for clinical interventions and may provide insights for potential prevention of substance abuse in gay men. If beir. how does it do a larger focus shame? How - are the effects Conce; oneself are not new conceptua Western societ| oiluires and, t} Vocalized and . aWording to SI CHAPTER TWO Literature Review Ifbeing gay in our culture is associated with shame, what is ”shame" and how does it develop? Ifshame exists and is so powerful, why has shame not had a larger focus in psychology? Conversely, what is pride and how does it relate to Shame? How is shame different from other negative aspects of being gay? What are the effects of this shame when it is directed at the identity of gay men? Concepts such as inferiority, worthlessness, and feeling poorly about oneself are not new to psychology (Kaufman, 1989), however, "shame” provides a new conceptualization of these experiences (H. B. Lewis, 1987). Until recently, Western society had not addressed the concept of shame as readily as other cultures and, therefore, shame is not in the "cultural consciousness" to be vocalized and examined (Kaufman, 1989; M. Lewis, 1992). Furthermore, according to shame theory, the very experience of shame is non-verbal, first learned at a pro-verbal stage. Consequently, it defies articulation and exploration (Kaufman, 1989,1992; M. Lewis, 1992; Nathanson, 1992). In addition, professional psychology has paid little attention to shame as an affect. Freud was more concerned with guilt than shame (Kaufman, 1989). Several of Freud’s followers such as Eric Erikson, Karen Homey, and Alfred Adler each addressed the concept of shame to varying degrees (Kaufman, 1989). None, however, made shame a central construct. The wave of behavioral psychologists, such as Skinner, focused exclusively on behavior and had little use for affects (H. B. Lewis, 1987). Finally, recent developments in social cognition, albeit deepening the underStan; of human at? e R. The ch: with any new . psychological shame and ide associated wit} shame, are also In addition, cc- (Kaufman. 193 clarification or Shaman. Althouy GORSWCLS. sha: and the public . however, have 19864K3Ufmapl Wmmser. 1981' and sham, is at I Can be “Peri Cr, ' — —-_ the understanding of thought and the social forces, do not address the complexity of human affect, of which shame is a powerful example (H. B. Lewis, 1987). EI°I'ESI DIEII°IC The theoretical and empirical investigation of shame has just begun. As with any new or revitalized construct, the relationship between shame and existing psychological constructs must be addressed so that the validity of identity-based shame and identity-based pride can be examined. Several constructs are often associated with shame. Negative beliefs about the self, that are the hallmark of shame, are also associated with the construct of low self-esteem (Chang, 1988). In addition, constructs such as guilt and pride are also closely linked to shame (Kaufman, 1989, 1992; Nathanson, 1987, 1992). Therefore, atheoretical clarification of shame in relationship to these other constructs is warranted. Shameandfiuilt Although conceptualized by shame theorists as distinctly different constructs, shame and guilt are often used interchangeably by both psychologists and the public (Kaufman, 1992; H. B. Lewis, 1987). The majority of theorists, however, have differentiated guilt from shame (Buss, 1980; Possum & Mason, 1986; Kaufman, 1989; H. B. Lewis, 1987; Lynd, 1958; Piers & Singer, 1953; Wurrnser, 1981). Unlike popular conceptualizations that guilt is about behavior and shame is about identity (Bradshaw, 1988), shame theorists (Kaufman 1974, 1989, 1992; H. B. Lewis, 1987) posit that both shame and guilt can be about behavior or identity. In addition, these authors point out that both shame and guilt can be experienced either publicly or privately. Individuals can feel guilty or shamefirl about an action or can feel guilty or shameful about who they are. The cl: affect theory. shame is one c mediated by tr I I963) are “per patterns and acl from each othe unique, Specif; guilt is not con fundamental, p possible core a 11?” Lhem)" 5: mm"? affects 3$€C0ndary‘ m u“dfi'SIand Its as a Way of 00' broad range of focused‘ Prime magical hone 1:“T‘d‘msntziiiy i“ do Will m, The er vi : CWI“g them The clearest theoretical distinction between shame and guilt is set forth by affect theory. According to affect theory (Kaufman, 1989; Tomkins, 1963), shame is one of the nine core affects, i.e., it is a basic, hard-wired emotion mediated by the limbic system of the brain. Core affects, according to Tomkins (1963) are “pure” emotional states that have concordant specific neuronal firing patterns and autogenic facial expressions that are both disciminant and unique from each other. Guilt, according to Tomkins (1963), does not have its own unique, specific neuronal firing or facial patterns, whereas shame does. Therefore guilt is not conceptualized as a core affect. Consequently, shame is a more fundamental, pristine affect, whereas guilt is an amalgamation of several other possible core affects (see Table l for a list of the nine core affects) (for more on affect theory see below). Because there are several possible combinations of primary affects that can comprise guilt (e.g., shame + fear; shame + distress), it is a secondary, more auxiliary state that must be addressed phenomenologicafly to understand its components (Kaufman, 1989; M. Lewis, 1992). Guilt can be used as a way of controlling the self and others (M. Lewis, 1992) and may reflect a broad range of affective states (Kaufman, 1989). Shame, however, is a more focused, primary affect (Kaufman, 1989). Shame, unlike guilt, " . . . is not a magical hope nor an attempt to control others. Shame is the experience of being fundamentally bad as a person. Nothing you have done is wrong, and nothing you can do will make up for it " (Kaufman, 1974, p. 569). The empirical relationship between guilt and shame provides support for viewing them as conceptually different. Chang (1988) reviewed and summarized the psychom CIT .40 between th identical cons: Lewis. 1987; E Perlman, 1953. shame is a prirr emotions, these diiie'ing constr a‘i'ecr and share to this study. l2. ,4 .~ Just as s 1993)- Pride is detemem affec lolful feeling a . that the 'me" iny ’ch self is 300i 35m “WOT cxc mmphshmenn in Self, or in There cxi relatimship bent I992) h-‘T’Othmz the psychometric literature on shame and guilt and found an average correlation of .40 between them, lending empirical support that shame and guilt are not identical constructs (e.g., Binder, 1970; Cook, 1991; Crouppen, 1976; Harder & Lewis, 1987; Hoblitzeile, 1987; Korpi, 1977; Minnan, 1984; Negri, 1978; Perlman, 1953; Smith 1972). Although this does not provide direct evidence that shame is a primary affect and guilt a secondary mixture of other more primal emotions, these studies are congruent with affect theorists claim that the two are differing constructs. Because of shames theoretical primacy in the literatures of affect and shame theories, it is the affect of shame, not guilt, that is of importance to this study. Shmandhide Just as shame is rooted in the primary affects, so also is pride (Nathanson, 1992). Pride is the experience of the primary enjoyment affect or the primary excitement affect directed at the self (Nathanson, 1987). Pride is the positive, joyful feeling a child experiences when she or he performs a task well and realizes that the "me" involved with the task did well. Pride is the sense of knowing that the self is good, efficacious, and healthy. Pride is the experience of enjoyment affect and/or excitement affect focused directly on the self or on the selt‘s actual accomplishments (Nathanson, 1987). Pride is the enjoyment/excitement invested in self, or in accomplishment of self (Kaufman, 1989). There exists a discrepancy, however, in the way theorists conceptualize the relationship between shame and pride. Several (Lynd, 1958; Nathanson, 1987, 1992) hypothesize that shame and pride are opposite ends of a continuum. Nathanson ( l C confluence of ' by its nature is Further develc I to conclude, "SI us to afiiliate VI capable of beinl measure our ex Lynd (1958) w tsense of . , , “her (In Kauf Kaufma '5 i“dependent meow dimi Nathanson (1987) wrote, "Our name for the emotion resulting from the happy confluence of the affect joy and the experience of personal efficacy is pride, that by its nature is always linked to the emotion shame in reciprocal fashion" (p.186). Further developing his work on shame and pride, Nathanson (1992) has gone on to conclude, "Shame, of course, is the polar opposite of pride. Where pride allows us to affiliate with others, shame makes us isolate from them. All our actions are capable of being viewed along a shame/pride axis, a yardstick along which we measure our every action." (p. 86). Other authors agree with a shame-pride axis. Lynd (1958) wrote, ”If shame lies at one end of the process of identity formation, a sense of . . . pride in oneself and humbleness in relation to the world lies at the other" (in Kaufman, 1974, p. 572). Kaufman and Raphael (1991), however, believe that shame and pride exist as independent constructs. Conceptualizing shame and pride as independent, orthogonal dimensions allows an individual to possess a more complex constellation of shame and pride issues. Kaufman and Raphael point out that although an individual may possess a deep sense of shame about himself or herself with regard to several needs, drives, or affects, that individual may also possess a developing sense of pride about other aspects of the self. In this regard, shame and pride are not seen as unipolar opposites, but rather as complex, independent constructs that affect individuals on multiple levels. Applied to gay men, for example, an individual may begin to feel a developing sense of pride about his relationship with other gay men, but may still have unresolved shame issues about his homosexuality in relation to his religious convictions. In this stay, borh a St operating sirr. has yet exami based pride in| m , el Perhapl clarification 01" most theorists are emive (Br (1933, 1937), ; does not menti constructs (see Shame Piers & Singer‘ f ' _ 10 way, both a sense of pride and a sense of shame about homosexuality can be operating simultaneously but independently. Unfortunately, no empirical research has yet examined the relationship between identity-based shame and identity- based pride in gay men. ShameanifielL-Esteem Perhaps the most thorny issue confronting shame theorists is the clarification of shame versus self-esteem. A review of the literature reveals that most theorists are either silent about the relationship of shame to self-esteem or are evasive (Branden, 1983, 1987; Campbell, 1984; Jackson, 1984). Branden (1983, 1987), like many authors who have written a great deal about self-esteem, does not mention shame, as if there is no important relationship between the two constructs (see also Campbell, 1984; Jackson, 1984). Shame theorists (e.g., Buss, 1980; Kaufman, 1989, 1992; Lewis, 1971; Piers & Singer, 1953; Wursmer, 1981) describe shame as feelings of inferiority, defectiveness, worthlessness, unimportance, and falling short of one’s own standards or ideals. Low self-esteem has been defined by Jacobson (1964) as falling short of one’s standards or ideals, or as the judgment of the worthiness that a person holds about him or herself (Coopersmith, 1967). In addition, self- esteem has been described as the discrepancy between ideal and actual self (Campbell, 1984; Jacobson, 1964), as self-worth (Branden, 1983; Check & Buss, 1981; Coopersmith, 1967), or as a positive or negative attitude toward oneself (Rosenberg, 1965, 1979). Branden (1983) defines self-esteem as follows: Self-es of eff" com pe esreerr. and in of part: particu standai (3312) However, tha use of the ten UOminon ther “be identicc z11d self-ate, Clan' is a Primary 1984) is tha self, This d "amnion Primary 01 Ch mismatch saii‘tsteei ll Self-esteem is a concept pertaining to a fundamental sense of efficacy and a fundamental sense of worth, to competence and worthiness in principle . . . In sum, self- esteem is an 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 . . . Living up to my own standards is an essential condition of high self-esteem. (p. 12) However, that shame theorists could substitute the word “shame” for Branden’s use of the term “self-esteem.” Negative views or beliefs about the self are a common theme for both shame and low self-esteem. At first glance they appear to be identically defined terms. So what then is the difference between shame and self-esteem? Clarity is again provided by affect theory. Affect theory posits that shame is a primary, innate, afl'ect. A more current definition of self-esteem (Campbell, 1984) is that self-esteem is the degree that the actual self image matches the ideal self. This definition indicates that self-esteem is a complex, organized cognitive evaluation of the self. Affect theory, then, would suggest that shame is the primary afl'ect associated with the cognitive construct of low self-esteem. Chang (1988) proposed that when there is a cognitive discrepancy or mismatch between the actual self and the ideal self (the current definition of low self-esteem) the resulting affect is shame. Lewis (1987) similarly stated that ’ l shame is 'he 3. correlation be identical cons shame and se. cognition vets low self-esteel primary affect cognitive belie This si' different pans they Were invt Sim thing~ S mmmon cons WeSllons abo himself-e519 measuring se‘ “Insuring ju nw Perfect c lheory more i inStIllrnents t Unlik Shame theory dmigned to e 12 shame is the affective state of low self-esteem. Chang (1988) found an empirical correlation between shame and self-esteem of -.95 suggesting that the two are identical constructs. Chang suggested that the empirical relationship between shame and self-esteem is very strong because they tap differing aspects (i.e., cognition versus affect) of the same dimension (Chang, 1988). However, when low self-esteem is assessed empirically, the instrument may be measuring the primary affect of shame because the items on traditional measures confound cognitive beliefs and affective experiences. This situation is not unlike the three blind people who, afier touching different parts of the same elephant, came to differing conclusions about what they were investigating. In fact, they were all touching differing aspects of the same thing. Shame and self-esteem may be differing conceptualizations of a common construct. Further, instruments that tap low self-esteem appear to ask questions about both cognition and affect, thereby blurring shame affect items with self-esteem items. Consequently, many extant empirical scales appear to be measuring several aspects of the “elephant”, although purporting they are measuring just the “trunk.” This may, in fact, explain why Chang (1988) found a near perfect correlation of shame and self-esteem. As theorists develop shame theory more extensively, researchers may need to further refine and homogenize instruments to ensure specificity and discriminant validity. Unlike self-esteem definitions, shame is borne out of a developed theory, shame theory (see below). This study, utilizing shame theory as an anchor, is designed to explore more fully the affect (i.e., the shame component) piece of the construct cor: investigations relationship b: that is specific gay (see later : To uni miew the de'. 'Phenomenolc sense. ...In‘r. of exposure. \ “TORI Shamei insuffici humiliat- l3 construct comprised of both self-esteem and shame. Although preliminary investigations (e.g. Cook, 1991; Chang, 1988) have demonstrated a significant relationship between shame and self-esteem, this study will examine the shame that is specific to a particular aspect of identity, identity-based shame about being gay (see later section on Gay Identity-Based Shame). Shamelhm To understand how shame is currently conceptualized and theorized I will review the developing field of shame theory. Kaufinan (1992) argued that "Phenomenologically, to feel shame is to feel seen in a painfully diminished sense. . . . Inherent to this experience of shame is this sudden, unexpected sense of exposure. We stand revealed as lesser" (p. 1). Possum and Mason (1986) wrote: Shame is an inner sense of being completely diminished or insufi'rcient as a person. A moment of shame may be humiliation so painful or an indignity so profound that one feels one has been robbed of her or his dignity or exposed as basically inadequate, bad or worthy of rejection. A pervasive sense of shame is the ongoing premise that one is fundamentally bad, inadequate, defective, unworthy, or not firlly valid as a human being. (p5) Shame is a powerful emotion that when experienced in a timely, discrete manner can facilitate the development of identity, conscience, humility, and respect. When shame is internalized and directed at the self, however, the results can be psychologist. isinponant 1 human txper. l99l,.\'atha: hemid-l950. langmage for 'l as predominai I see a MOllVI R 1985; Ze l4 psychologically damaging (Kaufman, 1992). To understand the affect of shame it is important to examine a field of psychology that views affect as primary to the human experience-affect theory (Kaufman, 1989, 1992; Kaufman & Raphael, 1991; Nathanson, 1992). Afiectlhm Shame theory has its roots in affect theory, which was first developed in the mid-1950s by Silvan Tomkins. Affect theory provides a model and a language for the exploration of affect (Kaufman, 1989). Tomkins viewed affect as predominant over cognition and behavior. I see affect or feeling as the primary innate biological motivating mechanism, more urgent than drive deprivation and pleasure, and more urgent even than physical pain. Without its amplification, nothing else matters, and with its amplification anything can matter. (1987, p.137) Tomkins distinguished nine different primary or innate affects categorized into two groups: positive and negative affects (see Table l). Insert Table 1 about here Recent research on positive and negative affectivity (Watson & Tellegen, 1985; Zevon & Tellegen, 1982) has confirmed Tomkins' (1963) belief that there are two basic self-reported afiectivity at. lellegen's sca enthusiastic. 3 characterized researchers ha major human lhwry that p05 work has cxm 0f Positive and and Pride. on t and Pride are c The fre 1989), As a re Public display: PTOhibired). A although it is c 15 are two basic dimensions of human mood. Emerging consistently in studies of self-reported affect are two dimensions that have been identified as positive affectivity and negative affectivity (commonly abbreviated as PANAS in Tellegen's scale). Positive affectivity reflects the extent that a person feels enthusiastic, active, and alert, whereas negative affectivity is a dimension characterized by sadness and lethargy (Watson, Clark, & Tellegen, 1988). These researchers have concluded that positive and negative affectivity represent two major human trait dimensions. The results seem consistent with Tomkins' affect theory that posits two categories of affect: positive and negative. No empirical work has examined the relationship between these two broad human affect traits of positive and negative affectivity, on the one hand, and the constructs of shame and pride, on the other. Afi‘ect theory (T omkins, 1963) would predict that shame and pride are components of negative and positive affectivity, respectively. The free expression of affect is powerful and highly contagious (Kaufman, 1989). As a result, societies have developed rules to govern its expression (e.g., public displays of rage, tears, and expressions of affection between men are prohibited). According to Tomkins (1963), shame occupies a unique role in that although it is one of the primary affects, it can also serve as an auxiliary affect that is activated by any perceived barrier to the free expression of the primary positive affec otpressing or. according to l society's proh inboth stress . i‘fvt‘wzQ—‘E ch Tomki. and were 'pmé bOIh innate anc mtmally, In contributing to some: of Shan brain (b) inter; Shame. 1111.113; the “Pefience “Mm“ ( e o(timed as “To 16 positive affects. That is, any imposed barrier that prevents a person from expressing one of the innate positive affects will illicit the shame affect. Stress, according to Tomkins (1963), is the result of suppression of affect. Therefore, society's prohibition of the full expression of affect, positive or negative, results in both stress and shame. But there are other ways to illicit the affect of shame. Actixatoranffihame Tomkins (1987) postulated that affects or feelings were biologically based and were "programs" located deeply within the brain. He thought that affects had both innate and learned components and could be stimulated both internally and externally. In addition, an individual's contextual experience is seen as contributing to the learned components of affect. Kaufman (1989) outlined three sources of shame activation involving (a) either innate mechanisms within the brain, (b) interpersonally transmitted shame activators, or (c) cultural sources of shame. W. Shame can occur when there is a reduction in the experience of positive affect such as when a person's fundamental expectations (e.g., that they are good and efficacious at a task) are suddenly exposed as wrong (they fail) (Kaufman, 1989). Shame is activated whenever a positive affect is expected and denied expression by experience. An example of this could be others and the manner. it ser Imerr . individuals is experience po llhen this doe relationships r relationships 1 When and and disappoi, mleehavior Consistent“. 1 this exDerien that the pain h“ broken or bridgg} BY nr broke", and. 17 this could be when a man anticipates he has the confidence to ”come out" to others and then is unable to find the courage. When shame is activated in this manner, it serves as an inhibitor of positive affect. W; A second realm that shame occurs for individuals is within an interpersonal context. According to Kaufman (1992), to experience positive affect children must have their needs met by caregivers. When this does not occur, shame is activated. Children build interpersonal relationships with significant others, such as parents, and rely on these relationships to meet their needs. When these needs are occasionally not met, as is typical and healthy, the child learns an appropriate set of parameters for shame and disappointment and, in the case of parental disapproval of the child's misbehavior, the beginnings of conscience. If, however, the child repeatedly and consistently fails to have his or her needs met, the child will search for meaning in this experience. Because adults appear as big and powerful, the child concludes that the pain in the relationship is her or his fault. It is the adult, however, who has broken what Kaufman (1974, 1989, 1990, 1992) calls the "interpersonal bridge.” By not meeting the child's needs, the interpersonal bridge is temporarily broken, and the child experiences shame. This initiates the development of feelings of inl his or her nee becomes cons” Cuh‘u" nature The c importance to context that tlr communicated l11100. Adoles themselves is r exPel’ltnce of link We,“ , 18 feelings of inherent unworthiness that increase as the child continues not to have his or her needs met. The end result is the internalization of shame-the child becomes convinced that there is something fundamentally wrong with him or her. W. A third set of shame sources are contextual in nature. The cultural shame learned later, during adolescence, is of particular importance for homosexuality (Kaufman, 1989, 1992). It is within the cultural context that the inseparable link between shame and homosexuality is communicated (Kaufman, 1989). Touching among males is shamed and made taboo. Adolescent males learn that the way to create a strong male image of themselves is to accuse others of being a ”sissy" or ”queer.” The resulting experience of shame in their targets quickly elucidates the powerful effects of the link between homosexuality and shame. I l' . E 51 Internal representation develops as people experience shame through innate activators, interpersonal sources, and developmental sources. These representations or ”shame scenes” (Kaufman, 1989,1992) develop when reoccurring internal or external patterns of interactions have been repeatedly paired with shame. Once shame is internalized, extemal agents are no longer 363 Old 6 a Chili sha' sens bind “he Split Con: lot-h “1hr 19 necessary to activate shame. The individual need only experience a need, drive, or affect that has been associated with shame to experience shame. Wins, The association of shame with affects, drives, and needs is called a shame-bind (Kaufman, 1989,1992). Once an individual is shamed with enough force, duration, or consistency, shame is bound to the event, behavior or issue in question. This process of shame binding is analogous to the conditioned emotional response (CER) concept in classical or Pavlovian conditioning (Schwartz, 1984). For example, shame binds can develop around an affect such as anger. In a particular family or culture, the display of anger may be taboo. The child is shamed for displaying this emotion so much that the child learns to bind shame with anger until the child no longer feels anger without an overwhelming sense of shame. In this manner, individuals in our culture learn strong shame binds surrounding any affects, needs, or drives associated with homosexuality. When these shame binds become painful enough, an individual may disown or split off the various aspects of the self that have been bound by shame. Consequently, men who find emerging aspects of themselves bound to shame about homosexuality will be forced to deal with the pain of their own shame. Due to their shame binds, they may not be able to experience homoerotic feelings without also feeling shame. Their shame can be addressed by engaging in one or may {B 1'3! horns horn: their feelir “PO there 20 more of the following processes: they may face their pain and shame about being homosexual; they may disown or split off these aspects of themselves; or they may develop ways of modulating their pain with addictive/numbing behaviors (Bradshaw, 1988; H.B. Lewis, 1987; Kaufman, 1989). 51 E . I I H l' Shame and pride are intimately involved in homosexuality and homosexual identity (Lynd, 1958; Nathanson, 1992). Although shame, in general, has received little attention until recently, the relationship between shame and homosexuality remains virtually unexplored. Those who have struggled with their developing sense of self as homosexual can relate to the concept of shame as feeling seen (i.e., open to public ridicule), and fearing an unexpected sense of exposure. For many, the suspicion by others that one is homosexual is to feel that there is something inherently wrong with the self, and to live in fear that one's sexual orientation, and thus one's inherent defectiveness, will be exposed (Nathanson, 1992). The high degree of homOphobia in our culture contributes to the relationship between shame and homosexual identity. Innate activators of shame occur when a child discovers that the expected enjoyment of touching another boy is forbidden by others. Nathanson (1992) explained that young boys are taught that the culture views homosexuality with disgust, contempt, and shame. This disgust and contempt invites a break in the "interpersonal bridge.” The shame surrounding homosexuality reinforces that it is wrong, defective, and disgusting. Kaufman (1990) noted that any touching or affection displayed between two men 21 is taboo and may elicit disgust and contempt from others. The consequences for gay men are revulsion, hatred, and contempt, either directed outward at others or turned inward against the self (Kaufman, 1990). Nathanson (1992) wrote: Each and every one of them [identified homosexuals] will be subjected to taunts, shaming assaults, physical attack and abuse, neglect and outright abuse by the legal system, ostracism, and rejection on a massive scale. Merely to act tenderly toward a beloved companion sets up so much disgust and dissmell in the average heterosexual onlooker that the homosexual world has always been one of secrecy and isolation. (p. 299) Wasted—Shams A ”shame bind” (Kaufman, 1989, 1992) develops when the affect of shame is paired either through intensity or repetition to some form of experience (e.g., other affects, needs, or drives). Shame, in general, can be bound to any number of affects, drives, or needs. Applying shame binds to homosexuality, it is possible to describe the affective development of a young boy, who is destined to be homosexual. First, developmentally, he learns to bind the topic of homosexuality, and undoubtedly his own budding homoerotic feelings, with shame. This shame surrounding homosexuality is pervasive at all levels of our society and is learned from the culture, family, and friends. Homoerotic topics, feelings, and thoughts are permanently linked to shame. Homophobia makes the binding of shame and homosexuality inevitable. The result is that fixture their dad 22 activation of homoerotic thoughts will activate shame, which inhibits expression of these homoerotic feelings. Just to experience homosexual thoughts or feelings produces overwhelming shame. Thus, shame is bound to affects, needs, and drives. This is a less significant issue for boys who develop heterosexual identities, as their future identity (i.e., sense of self) will only occasionally be threatened by homosexual shame binds (e.g., if they experience a situation where their actions or motivation are falsely interpreted as homosexual). As gay males develop, however, they must face powerful feelings of shame that will occur over and over as they are forced to confront their emerging orientation. At the point when the boy recognizes himself as gay (i.e., comes out to himself) shame is bound not only to several homoerotic affects but also directly to the self. In this study, the binding of shame to homosexuality will be referred to as gay identity- based shame. Gay identity—based shame is the most "toxic” kind of shame bind (Bradshaw, 1988; Nathanson, 1992). Unlike the shame bind associated with a circumscribed affect, drive, or need, shame-bound homosexuality is directly about identity because homosexuality is a part of a person’s identity. Although negative consequences may exist for men unable to express or feel anger without feeling shame, according to shame theorists, shame bound to a part of one's identity may lead to serious psychological disorders (Bradshaw, 1988; Kaufman, 1989, 1992; HB. Lewis, 1987; M. Lewis, 1992; Nathanson 1992). Shame binds attached to the identity are a source of intense pain. Shame theorists posit that people will try almost anything to avoid this pain. As mentioned earlier, there are several courses of: the say the a 'h less off l Only litter 23 of action that a person can engage in to numb the pain of shame: they may face their pain and the shame of being homosexual, they may disown or split off these aspects of themselves and not be able to come out to self, or they may develop ways of modulating their pain with addictive/numbing behaviors. Until recently, the conceptualization of shame as an etiologic agent of psychopathology has been a theoretical view that has not received empirical investigation. Several recent studies have begun to test the link of shame to eating disorders (e.g., see Cook, 1991). The results confirm shame theorists' tenet that shame-based identity is so powerful that it may lead to the development of serious "pain numbing” pathology (e.g., alcoholism and other substance abuse, depression, eating disorders, compulsive disorders), and the consequences of these serious disorders may be death (Kaufman, 1989). Ci I I . -E I E . I Theoretically, for gay men to feel positively about themselves and their identity, they must work through their shame issues to develop a sense of pride about their homosexuality (Nathanson, 1992). This ”working through" process takes time, and as an individual moves along this journey they become less and less bound by shame. As Nathanson (1992) stated, "gay pride” is shame warded off (p. 299). A sense of pride about being homosexual is possible for gay men only after they have begun to work through the painfirl shame they have internalized from their experience in our culture with family, friends, and peers. For gay men, and for those who have experienced a great deal of shame 24 throughout their lives, a sense of pride develops only as the individual realizes that there is not something fundamentally wrong with who they are. This process of healing and insight is difficult, at best, when gay men feel that parents, friends, associates, and the culture sees them as having something basically wrong with them. Pride occurs when gay men look deeply at the self and discover that there really is nothing wrong (Nathanson, 1987). For the gay man, this means coming to the realization that despite what he has experienced and internalized from others, his sense of self as a person and as a homosexual is positive, healthy, and valuable. This working through, according to Kaufman (1974, 1989, 1992), can occur through psychotherapy, deep friendships, and ‘ relationships with colleagues. The important rule is that the person must find someone with whom he has a significant relationship who can break his shame binds and let him undo the sources of shame activation. This includes lessening the innate activators of shame related to being gay; re-establishment of interpersonal bridges; and reliving/rethinking the developmental and cultural sources of shame activation. As noted above, there are, theoretically, several possible pathological outgrowths of painful internalized shame affects. Of these, substance abuse has been repeatedly cited by shame theorists as a negative consequence of internalized shame (Bradshaw, 1988; Kaufman, 1992; Nathanson, 1992). For example, alcohol can be used to medicate the pain of shame (Bradshaw, 1988; Possum & Mason, 1986; Nathanson, 1992). Ifshame is associated with homosexuality, and subsranc may be t ham r‘ died fror alcohols available (Nardi, 1 literature that as or problem 1 (National Statistics . abuse pro As fairly gens difficulties (see, {Or 6, Williams, ] dependenC) 25 substance abuse is a means of dealing with internalized shame, then homosexuals may be at higher risk for abusing substances. .11.-..4 00' 6'10 I'd“.-"vr' " 0 0.-.t‘6' ”t r3 U‘l According to the Pride Institute (1989), more gay men and lesbians have died from chemical dependency than from AIDS, however, the topic of alcohol/substance abuse among gays is discussed in less than 1/10 of 1% of all available references in the alcohol literature during the 30 year period 1951-1981 (Nardi, 1982), and this oversight continues. Ironically, although the proportion of literature devoted to substance abuse in homosexuals is small, estimates predict that as much as 30% of the openly homosexual population have a drinking problem (see Table 2), compared to an estimated 10% of the general population (National Institute on Alcohol Abuse and Alcoholism, 1978). If the above statistics are accurate, then a large proportion of gay men have drinking/substance abuse problems, yet this relationship is not being adequately investigated. Insert Table 2 about here As Table 2 shows, the prevalence rate of substance abuse in gay men are fairly consistent across studies. Almost all of these studies have methodological difficulties due to utilizing convenience samples in gay bars and gay “ghettos” (see, for example, Fifield, De Crescenzo, & Latham, 1975; Weinberg & Williams, 1974), or utilize questionable methods for determining substance dependency (Fifield, De Crescenzo, & Latham, 1975; Saghir, & Robins, 1973; Weinberg . Fifteld et 2 Communir estimates c adequately dependenc: Hos that gay me One questic alcoholism,” The in psychoan l0 explain a Iamrm to criticism of 26 Weinberg & Williams, 1974). For example, the most quoted figure is from Fifield et al.’s (1975) study conducted in Los Angeles County by the Gay Community Services Center. They relied on gay bar patrons and bartenders’ estimates of substance dependency rates in the gay community, and did not adequately define substance dependency or how their respondents estimated dependency. However, the consistency of these findings lends support to the conclusion that gay men have a higher rate of substance abuse than the general population. One question for this study is, what factors contribute to a higher rate of alcoholism/substance abuse in homosexuals? The historical link between alcoholism and homosexuality has been rooted in psychoanalytic theory. For decades, followers of Freudian thought have tried to explain alcoholism simply in terms of latent homosexuality (Israelstam & Lambert, 1983; Nardi, 1982). Although Freud provided both support for and criticism of the psychological "health” of homosexuality (Lewes, 1988), later psychoanalytic followers were convinced that the origins of both alcoholism and homosexuality were intrapsychic, intertwined, and pathological (Israelstarn & Lambert, 1983). The neo-Freudians attempted to conceptualize the relationship between homosexuality and alcoholism as possibly demonstrating a bi-directional relationship, positing that alcoholism could precipitate latent homosexuality, or that homosexuality could precipitate latent alcoholism (Nardi, 1982). More recently, researchers have explored other explanations for the relationship between alcoholism and homosexuality (Beaton & Guild, 1976; Gay con ra- our. inn. COl'l 27 Council on Drinking Behavior, 1982; Kus, 1988). These explanations posit societal and cultural as well as intrapsychic causes. In examining many of these intrapsychic and sociocultural forces that presumably increase homosexuals' proclivity to abuse substances, the common theme of oppression (i.e., being in a ”less than acceptable" role in society) may be at the heart of the etiology of increased substance abuse (Kuss, 1988). The core of this oppression of homosexuals in our society may be internalized by gay men as shame about being homosexual. ShameAlccthismmdfiaLMen The specific issue here is not "what is the etiology of alcoholism in homosexuals," because the "causes" of alcoholism in homosexuals may be the same complex, etiologic variables that lead non-homosexuals to abuse substances. The question addressed here is "what might explain the higher level of substance abuse in gay men than in the general population?" Several authors (Bradshaw, 1988; Fossum & Mason, 1986; Kaufman, 1992; M. Lewis, 1992; & Nathanson, 1992) have discussed the relationship between shame and addictive disorders, concluding that shame is at the root of these addictions. Kaufman (1989), in particular, addresses the shame-addiction cycle. An individual may use alcohol to numb the pain of his shame. The addiction, however, also produces a sense of innate shame for the person, that then has to be numbed by alcohol, and the cycle continues (Bradshaw, 1988; F ossum & Mason, 1986). Therefore, although shame may lead to a proclivity to abuse substances in all humans, shame may be an ape: of she leuis in eff: lKaut may 1 lesis PM hlaso 1992; 28 especially potent etiologic factor in the alcoholism of gay men. The relationship of shame to alcoholism in gay men will be examined in this study. Summmeitetamre Shame has been an underexplored issue in modern psychology (H. B. Lewis, 1987; Kaufman, 1989; Nathanson, 1992). Shame occupies a unique place in affect theory (Tomkins, 1963 ), serving both positive and negative functions (Kaufman, 1974, 1989, 1992; Tomkins, 1963). When evoked repeatedly, shame may become bound to needs, drives, or other affects (Kaufman, 1989, 1992; M. Lewis, 1992; Nathanson, 1992). Ifshame becomes bound to the self, it can have powerful, negative effects on psychological health (Cook, 1991; Possum & Mason, 1986; Kaufman, 1989, 1992; Lewis, 1989; M. Lewis, 1992; Nathanson, 1992). Homosexuality, in our culture, is strongly associated with shame (Kaufman, 1989; Nathanson, 1992). Through innate, interpersonal, and developmental activators, shame becomes bound to many drives, affects, and needs in gay men (Kaufman, 1989). Because homosexuality is an aspect of identity, shame becomes bound, to varying degrees, to the self (Kaufman, 1989; Nathanson, 1992). When shame about being homosexual is bound to the self it is called gay identity-based shame. According to shame theory, the results of these identity based shame binds lead to various difficulties such as substance abuse (Bradshaw, 1988; Possum & Mason, 1986; Kaufman, 1989, 1992 ). The negative effects of shame can presumably be ameliorated (Kaufman, 1989, 1992; H. B. Lewis, 1987; M. Lewis, 1992; Nathanson, 1992) through pride. At this point, shame study r aliecti l. W} has of ; ort 511; ha the rela ori 29 shame theory as applied to homosexuality had not been tested empirically. This study explored the relation of shame binds to pride, self-esteem, positive affectivity, negative affectivity, and substance abuse proclivity in gay men. Rmhfluestinns This study focused on two broad areas: (a) the relation of gay identity- based shame and gay identity-based pride to one another, positive and negative affectivity, and self-esteem; (b) the utility of gay identity-based shame in predicting vulnerability for substance abuse. Before subject data could be utilized to address these two areas, the Marlowe-Crowne Social Desirability Scale was used to eliminate subjects who may have biased their responses in a socially desirable manner. Specifically, this study will examine the following questions: B l . l . [S] E . l 1. What is the relationship between gay identity-based shame and gay identity- based pride? Several authors have described shame and pride as opposite ends of a continuum (Nathanson, 1992), but others predict that shame and pride are orthogonal constructs (Kaufman, 1992). The orthogonal relationship is supported by current research on positive and negative affectivity, suggesting that an individual could have concurrent negative and positive feelings about the self (Watson, Clark, & Tellegen, 1988). This study will examine the relationship between shame and pride to see if they lie on the same continuum or if they represent independent constructs. Exploration of this issue will 30 address the possibility of a gay man having both a sense of pride and a sense of shame about being gay. 1' .ror‘r- o r- 0‘.“ .\ u-“vn turn-10 'IO' 0 '-,. "H -..r0 'sua 2. What is the relationship between gay identity-based shame and gay identity- based pride to self-esteem? Although other studies have found that shame in general and low self-esteem may be highly related, the relationship between identity-based shame about being homosexual and self-esteem is unknown. Is the relationship between gay identity-based shame and self—esteem identical to the relationship reported by Chang (1988) between global shame and self- esteem? Are gay identity-based shame and gay identity-based pride orthogonal to self-esteem? For example, is it possible for an individual to have fairly high self-esteem but still have high gay identity-based shame? 3. How are gay identity-based shame and gay identity-based pride related to positive and negative affectivity? Can gay identity-based shame and pride be subsumed under positive and negative affectivity? This exploration would provide information regarding the discriminant and convergent validity of shame and pride to the global personality traits of PANAS. E 1° . 11']. [II °-E IS] 4. What is the utility of gay identity-based shame and gay identity-based pride in predicting substance abuse? The final question is to assess the utility of these constructs in explaining a very important social issue: high levels of substance abuse in gay male populations. The : shame and p measures in: Shame Scale marchers t 1993). HOW one's sexual based shame' to being gay items 0f the l fitted with g; cosine-e1). w, based upon 5 modified 155 Gttshen Kai An ir hlimited Sub CHAPTER THREE METHOD Pilot Phase The first task in this study was to develop an instrument that assessed shame and pride about being homosexual. The Intemalized Shame Scale measures internalized shame and pride issues in general (see the Intemalized Shame Scale, ISS; Cook, 1987, 1991) and is becoming more widely used by researchers (see, for example, Chang, 1988; Reynolds, 1991; Wang, 1992; Wong, 1992). However, there were no existing measures of shame or pride specific to one's sexual identity. Therefore, the ISS was modified to measure gay identity- based shame and gay identity-based pride. Insertion of sentence stem references to being gay or to homosexuality were added to the existing twenty-five shame items of the ISS. In addition, the positive self-esteem items of the ISS were also fitted with gay reference sentence stems. Because the original ISS had only six positively worded items, an additional eighteen gay identity-based pride items, based upon shame and affect theories, were developed. These items, as well as the modified ISS items, were reviewed by an expert in the area of shame theory, Gershen Kaufman, for their construct validity. An initial sample of gay men were used to pilot the measures. Five hundred subject names from The National Community Masterfile (NCM), which contains the names of over 350,000 direct-mail responsive gay men, were asked to take the full complement of measures, and were asked to comment on their experience of being a participant. Their feedback was used in eliminating 31 possible by cabled cla. identity-hast reliability. no correctio sample of g: Acco (discussed in addressed sh. .95) between (1980) suggf ofmuincom ‘fieCllVe wa Salnple Of g; 32 possible barriers to participation for future subjects. In addition, this procedure enabled clarification of item wording and directions on the newly developed gay identity-based shame-pride scales and provided initial data on internal consistency reliability. Because of the positive reactions and feedback of these respondents, no corrections were necessary, and the next phase was conducted with a larger sample of gay men. Respondents According to Dawis (1987) construction of scales by the Likert method (discussed below) requires an N of at least 150 subjects. Researchers who have addressed shame and self-esteem have found a high correlation (r; = -.74 to r = - .95) between these two constructs (Chang, 1988; Cook, 1991). Lewis-Beck (1980) suggests that when dealing with constructs that may possess a high degree of multicollinearity, increasing the sample size is often the most desirable and effective way of more clearly partitioning the variance. Because access to a large sample of gay men was possible, 1000 gay men were recruited. Researchers have often obtained convenience samples in bars or bathhouses or assessed large numbers of the general population to find ”latent" homosexuals. The former approach may bias results for substance abuse and the latter is an extremely inefficient use of resources. The compromise appears to be conducting research with self-identified gay men in homophile organizations (see for example, Pifield, DeCrescenzo & Latham, 1975; Lohrenz, Connelly, Coyne & Spare, 1978; Saghir & Robins, 1973; Weinberg & Williams, 1974). This survey obtained subjects from The National Community Masterfile (NCM). NCM acquires and compiles the names of gay men from a variety of sources, and then sells these it researchers on any yaria stratified (by I country frorr. One I still greatly d rocnrited soit men differ fir people in this| Pm of this ch Product or se hlihly closet “Wmmmc homes, order scold also bl has the pom, me“ a‘v’ailabl homosexual m not ya a. w Howe Elmore, th Wed to l 33 sells these names to a variety of groups including businesses and market researchers. Unfortunately, NCM is unable to provide a sample that is stratified on any variable other than geographic location. Accordingly, NCM supplied a stratified (by geographic location) random sample of gay men from across the country from their database. One limitation of any study is that the external validity or generalizability will greatly depend on sampling procedures (Dawis, 1987). Ifgay men are recruited solely from the NCM database, one may ask how results based on these men differ from gays who are not on their list. It can be assumed however that the people in this database represent a diverse group. In order for a name to become part of this database, the person would have had to purchase some gay related product or service, with their name being subsequently sold to NCM. Possibly, highly closeted men or men who are early in the coming out process and who may not venture out into the " gay community" could, from the privacy of their own homes, order magazines, newspapers, and so on. Likewise, very openly gay men could also buy products and end up in the database. Therefore, the NCM database has the potential of representing the largest, most representative group of gay men available, at this time, in the country. In sum, sampling the entire range of homosexual individuals, from those who are totally out of the closet to those who are not yet aware that their future development will be homosexual, would be ideal. However, the latter are impossible, at this time, to identify and survey. Therefore, the present sampling methodology, although admittedly less than ideal, appeared to be a reasonable strategy. Thi. design is a; experiment being home Ace. do not give To aid in thi for oonducei utilized in th Initia selected fror afirming re about a Weel Wk later, 5 included a c lie had be. “Ethology r“earth Wa Examine m e regarding .h 34 Design This study used a correlational field study design (Gelso, 1980). This design is appropriate for correlational research when there is little or no experimental control over, or manipulation of, the independent variables such as being homosexual, degree of pride, shame, self-esteem, and so on. Procedure According to Dillman (1978) many researchers who conduct mail surveys do not give adequate thought to the design and implementation of their surveys. To aid in this process, Dillman (1978) has developed W for conducting mail and telephone surveys, and it is this survey protocol that was utilized in this study. Initially, subjects received a postcard informing them that they were selected from the NCM database encouraging them to participate in a gay- affirming research study, and stating that they would be receiving a survey in about a week (see Appendix I for a copy of the pre-survey postcard). About one week later, subjects received a survey packet through first class mail. The packet included a cover letter and a set of questionnaires. The cover letter explained that they had been selected to participate in a study being conducted by a gay male psychology student as part of his doctoral dissertation, that the intent of the research was to be gay-positive, and that the purpose of the survey was to examine the subjects' self-concept as well as some attitudes and behaviors regarding their homosexuality. The subject was assured that his responses were to be kept anonymous, and that if he had questions at anytime, he could contact the experiment; be reported I materials (5 it would tairl questionnarrl Nextl dernographi I shortened vel (Coleman, 1 I (Robinette, 1 Gay Identity (Resenberg Tellegen. 19s 1965). To ori order for We. After mmpleted QU “amped me sepil‘eate 00v ”DY of the su the Project ’ECeived a £0“ 35 experimenter by phone. In addition, subjects were assured that their scores would be reported anonymously and that only group data would appear in published materials (see Appendix G for a copy of the cover letter). Subjects were told that it would take approximately twenty to thirty minutes to complete the questionnaires. Next, subjects were asked to complete the enclosed questionnaires: a demographics questionnaire (taken from the US. Bureau of the Census), a shortened version of the Assessment of Sexual Orientation questionnaire (Coleman, 1987), the Marlowe-Crowne Social Desirability Scale - Form C (Robinette, 1991), the internalized Gay Identity-based Shame and internalized Gay Identity-based Pride Scales (modified ISS), the Rosenberg Self-Esteem Scale (Rosenberg, 1979), the Positive and Negative Affectivity Scale (Zevon & Tellegen, 1982), and the MacAndrew Alcoholism Scale (MAC) (MacAndrew, 1965). To offset any bias involved in the order of administration, all testing packets were counterbalanced so that the questionnaires were not in the same order for every subject After completing the survey, subjects were instructed to return the completed questionnaire to the experimenter in the enclosed self-addressed stamped envelope. In addition, subjects were asked to complete, and mail under separate cover, a yellow postcard included in their packet if they would like a copy of the survey results and were provided this information upon completion of the project Within two weeks of receiving the questionnaire packet, subjects received a follow-up postcard that thanked them for their participation in the study and reminde responded l in case they = any. Finally second follos' and remindir. responded (St 36 and reminded them to complete and retum the survey if they had not yet responded. In addition, subjects were given the phone number of the investigator in case they did not receive the survey or to express concerns should they have any. Finally, within three weeks of receiving the survey, subjects were sent a second follow-up postcard, again thanking them for their participation in the study and reminding them to complete and retum the survey if they had not yet responded (see Appendix I for a copy of the follow-up postcard). Measures All subjects were asked to supply basic information about themselves, such as age, sex, ethnic or racial identification, socioeconomic status (SES), and education level (see Appendix A). This form and questions designed to assess this information were modeled after the 1990 United States Census (U .8. Bureau of the Census, 1991). The Assessment of Sexual Orientation Scale (Coleman, 1987) asked subjects to describe their sexual orientation in a multi-dimensional manner. Coleman has augmented the "classic" Kinsey homosexual-heterosexual continuum. A shortened version of his measure (see Appendix A) was used to simply ask subjects about their current sexual orientation, future desired sexual orientation, and current comfort level with their sexual orientation. By using this instrument, a richer picture of subjects’ sexuality and gender identity could be obtained than simply using a one question Kinsey scale. In addition, the placement of these questions within the general demographic questions was designed It I. Other demo Social 0851? I972). For: Croone Soc measures the as reflected ": traits. Subjec Marlowe—Cr: M'erse item rInge of 13.: PTBSent a mo ‘ 93) With n “991), subje this measure anah’Sis. and a Six We SW (1985; 37 designed to encourage subjects to report these data as honestly as they would other demographic information. Ill _: S'lll'l'l'Sl-E C Description, Subjects responded to a short form of the Marlowe-Crowne Social Desirability Scale (Reynolds, 1982; Robinette, 1991; Strahan & Gerbasi, 1972). Form C is composed of 13 items from the original 33 item Marlowe- Crowne Social Desirability Scale (Crowne & Marlowe, 1960). The short form measures the extent that a subject is motivated to present a favorable impression as reflected by a willingness to admit to culturally desirable and undesirable traits. Subjects responded using a True (1) or False (2) to all 13 items. The Marlowe-Crowne includes item reversals, that is, negatively worded items that are reverse item scored. The resulting appropriate item transposals yields a scale range of 13-26 with higher scores representing a higher degree of motivation to present a most favorable impression. Form C appears to have high correlation (r = .93) with the original scale (Robinette, 1991) As suggested by Paulhouse (1991), subjects who scored two standard deviations above the mean or higher for this measure (i.e., are biasing their answers) were not included in further data analysis. Reliability. Cronbach's alpha (.76) has been reported by Reynolds (1982), and a six week test-retest correlation of .74 has been documented by look and Sipps (1985). Mammy, The short form has been found to correlate significantly with the L, F, and K scales from the MMPI (L = .50, F = -.52, K = .54) which are also reported to . Marlowe-C .' B). lies; I based on the original ISS have internal point scale (l shame. A 5a more. like t items of the ] Shame by the homosexuali LBGCause l a ”mailing b. Beta and “We, POSitive feel worded Shar Score 0f the gay identity lemme 38 reported to measure social desirability (Robinette, 1991). (A copy of the Marlowe-Crowne Social Desirability Scale - Form C can be found in Appendix B). 5 H ._I IS] $15] Won, The Gay Identity-based Shame-Pride Scales (GIBS-PS) are based on the Intemalized Shame Scale (ISS) developed by Cook (1991). The original ISS is a thirty-item scale designed to measure the extent that subjects have internalized painful levels of shame emotions. The ISS is a Likert-type five point scale (I = Never, 5 = Almost always) with higher scores representing more shame. A sample item from the ISS is "I feel as if I am somehow defective as a person, like there is something basically wrong with me.” The twenty-four shame items of the ISS were modified by the present author to tap gay identity-based shame by the insertion of sentence stems that reference being gay or homosexuality. An example of a modified gay identity-based shame item is “Because I am gay, I feel as if I am somehow defective as a person, like there is something basically wrong with me.” Because a large proportion of the original ISS asks about negative feelings and experiences, this scale includes six positively worded items concerning positive feelings about the self to correct for the response set of the negatively worded shame items. These items are not used in computing the shame scale score of the ISS. These six positively worded items were modified to measure gay identity-based pride. Eighteen additional items were added to complete the items needed for the gay pride scale. All the modified ISS shame items as well as the newly C | by an expedl and to ensur' | of shame-prl Re: and was oh;I I Additional a group of our researchers. and non-clir. .94 for a col; correlated his resulted in th other items fr PM were .: SCI therefom, a n imbies tssr 0fthe 153 w The Rosenbe Mme self. 1988) With 1h 39 the newly developed eighteen new gay identity-based pride items were reviewed by an expert of shame theory, Gershen Kaufman, to determine construct validity, and to ensure that they were clear and concise, and that they had not lefi any area of shame-pride untapped (DeVellis, 1991). Reliability, The alpha reliability of the original ISS was reported as .95, and was obtained from a non-clinical group. An alpha of .96 was reported for a group of out-patient chemical dependency clients (Cook, 1987, 1991, 1993). Additional alpha reliability coefficients for the ISS have been reported by other researchers. Rybak (1991) reported an alpha of .97 for a mixed group of clinical and non-clinical subjects. McFarland (1992) reported a reliability coefficient of .94 for a college student sample. Although the positively worded item set correlated highly with the total shame score (1 = -.69), a factor analysis of the ISS resulted in the positive items clustering in a unique factor that did not include any other items from the ISS. The test-retest reliability coefficients over a nine week period were .84 for the total shame score and .71 for the positively worded item set. 3mm Shame theorists posit that shame is a negative affect, and therefore, a measure of shame would be expected to show convergence with variables associated with emotion and psychopathology. The convergent validity of the ISS with low self-esteem varies depending on the self-esteem measure. The Rosenberg Self Esteem Scale, which is usually scored in the direction of positive self-esteem, correlates at r = -.74 (Cook, 1991) and r = -.95 (Chang, 1988) with the total shame score. The ISS and the Beck Depression Inventory have been c Spielberger correlated 1993). A Ct ISS. can be DICE; that measurj | optimize ens salidity (Bl Gum-op Point scale m i: Tomaka 1 item: that an )ields a scalr M 1979) to .88 M iilit-re m With measur psb'chiam'su confines lT 40 have been correlated at .79 (Cook, 1993) and .78 (Waite-O’Brien, 1991). The Spielberger State and Trait Anxiety Scales (Spielberger, 1983) have also correlated with the ISS, yielding correlations of .83 and .91, respectively (Cook, 1993). A copy of the Gay Identity-based Shame-Pride Scale, modified from the ISS, can be found in Appendix C. W Description, The Rosenberg Self Esteem Scale (SES) is a ten-item scale that measures global self-esteem (Rosenberg, 1979). The SES was designed to optimize ease of administration, economy of time, unidimensionality, and face validity (Blascovitch & Tomaka, 1991). Although originally designed as a Gunman-type scale (Rosenberg, 1979), the SES is typically scored using a four- point scale ranging from strongly agree (1) to strongly disagree (4) (Blascovitch & Tomaka, 1991). The SES includes item reversals, that is, negatively worded items that are reverse item scored. The resulting appropriate item transposals yields a scale range of 10-40 with higher scores representing higher self-esteem. Reliability. Cronbach alphas of .77 (Dobson, Goudy, Keith & Powers, 1979) to .88 (Fleming & Courtney, 1984) have been reported. The SES has a two week test-retest reliability of r_=.85 (Rosenberg, 1979). Wu, Convergent validity: The SES correlates highly to moderately with measures of self-ideal discrepancy score, [ =.67, self-image, r = .83, and psychiatrists' ratings of self-esteem, _[ =.56. Divergent validity: The SES correlates moderately with measures of depressed affect scales, r = -.30, and psychophys Rosenberg : De; Scale (MA Inventory, t lhe scale is . consistent 0v lime (Greene although she 41 psychophysiological anxiety measures, 1: = -.48 (Rosenberg, 1979). A copy of the Rosenberg Self-Esteem Measure is in Appendix D. MacAndmmAlmhnfismSealflMAC) Description, MacAndrew (1965) developed the MacAndrew Alcoholism Scale (MAC) from items taken from the Minnesota Multiphasic Personality Inventory, that differentiated alcoholic outpatients from nonalcoholic outpatients. The scale is composed of fifty-one items that significantly separated these two g'oups at the .01 level. Only No of the items directly ask about alcohol use, and the rest are unobtrusive. Reliabjflu. Several investigators report that subjects’ scores remain consistent over time (Huber & Danahy, 1975; Rohan, Tatro, & Rotrnan, 1969) indicating that the measure is tapping a dimension of behavior that is stable over time (Greene, 1980). Test-retest reliability is reported as high (Duckworth, 1983), although she reports not actual test-retest data. M A score of 24 or more was able to detect alcoholism in 81.8% of an outpatient population (MacAndrew, 1965) and in cross-validation samples the percentage of classificatory accuracy was found to be 81.5% (MacAndrew, 1965). Other researchers have found that the MAC could correctly classify 61.5% to 76%, and that cutting scores of 26 to 28 resulted in the highest percentage of correct classifications (Apfeldorf & Hunley, 1975; Rhodes, 1969; Rich & Davis, 1969; Uecker, 1970; Vega, 1971; Whisler & Canor, 1966). Kranitz (1972) and Lachar, Berman, Grisell, & Schoofi‘ (1976), have shown that the MAC cannot differentiate between alcohol and other substance abuse, suggesting that it is a measure of genera (1979) found that context of the enti validity. This has of the MACS vali- general substance isin Appendix E.) E; 12mm consists of two 10 dimension (Warsc blisters. 1982). : (llto Memely (: CDmputed by addi w“ r"‘sPeCtivel} The PAN; ms and then fag dimmiOns. The POSitive and Regan has“. Items lb or the Other Were Utilized“) r Educe sCale. 42 measure of general substance abuse, rather than alcoholism only. MacAndrew (1979) found that the MAC can be given as an independent scale apart from the context of the entire MMPI with only negligible differences in discriminant validity. This has been substantiated by Duckwlo (1986). Therefore, because of the MAC's validity, the unobtrusiveness of its items, and its ability to tap general substance abuse, it was chosen for use in this study. (A copy of the MAC is in Appendix E.) B .. III . 3m .. SI Willis: Description, The Positive and Negative Affectivity Scale (PANAS) consists of two 10 item mood scales that comprise a positive and a negative dimension (Watson, Clark & Tellegen, 1988; Watson & Tellegen, 1985; Zevon & Tellegen, 1982). Subjects rate each item ranging from very slightly or not at all (1) to extremely (5). Scores for both the positive and negative dimensions are computed by adding the ratings for the 10 items for the positive and negative scales respectively. The PANAS was developed by compiling a large number of affective terms and then factor analyzing these terms into positive and negative affective dimensions. The number of items needed to adequately measure the factors of positive and negative affectivity was shortened to provide a concise, yet valid, measure. Items that had substantial loading on one factor but a near-zero loading on the other were selected. Again factor analytic and reliability analyses were utilized to reduce the number of items from over 60 to a final 10 items for each scale . UKPANAS‘ feel that day, or hovs assess how the subje instructions of Other Reliability. \ Clark, & Tellegen, 1‘ internal consistency r Affectivity, .87. Big and t = .71 for Negar only slightly, 1: o. 17. 43 The PANAS asks respondents to rate how they generally feel, or how they feel that day, or how they have felt for specific periods of time. This study will assess how the subject feels in general, so that it can be comparable to the instructions of other measures. Reliability, With subjects responding how they feel in general (Watson, Clark, & Tellegen, 1988), alpha coefficients of the PANAS are as follows: internal consistency reliability for Positive Affectivity is .88, and for Negative Affectivity, .87. Eight week test-retest reliability is; = .68 for Positive Afi‘ectivity and r = .71 for Negative Affectivity. The two scales have been found to correlate only slightly, [ = -.17. Insert Table 3 about here Will The external validity of the PANAS has been demonstrated by it’s good c0nvergent correlations with other indices of positive and negative affectivity. Watson, Clark, and Tellegen (1988) report correlations of the PANAS with three commonly used measures of psychological distress and psychology: (1) The Hopkins Symptom Checklist (HSCL), which is a measure of general distress and dysfunction, (2) The Beck Depression Inventory (BDI), which is a self-report measure of depression, and (3) The A-State, which is a scale that asks subjects to rate their current mood or affect. Watson, Clark and Tellegen examined the correlations between the PANAS and these extant measures and reported correlati mdqulw reported correlations that indicate that the PAN AS provides good convergent validity (see Table 3) (A copy of the PANAS is in Appendix F). The first maj Identity-based Shari based on the guideli desmbed below: Reliability e based Pride Scales, correlations, item-s: multiple correlation an alpha of 3.70 co the scales, such as r deemed necessary. Pride Scales mm W demonstrated, “gay men. SUbjCCts' SC Social Desirability ANALYSIS 1&1:thth The first major goal was to construct a measure of internalized Gay Identity-based Shame-Pride. The construction and analysis of this measure were based on the guidelines developed by Dawis (1987) and DeVellis ( 1991) and are described below: Reliability estimates of the Gay Identity-based Shame and Gay Identity- based Pride Scales, completed by the pilot sample, included inter-item correlations, item-scale correlations, communalities (estimated by squared multiple correlations) and the alpha coefficient. It was hoped, and realized, that an alpha of 2.70 could be obtained (Nunnally, 1978). Therefore, adjustment of the scales, such as removing items with poor inter-item correlations, was not deemed necessary. After the construction phase of the Gay Identity-based Shame- Pride Scales measure was completed and the internal consistency of the scales was demonstrated, the scales were used in the main study with the larger sample of gay men. Mainimdx Subjects' social desirability data, as measured by the Marlowe-Crowne Social Desirability Scale - From C, were analyzed, and all subjects scoring above the cutting score were eliminated from further analysis. This was done to ameliorate potential social desirability response bias effects. In addition, the Assessment of Sexual Orientation measure was examined, and any subjects indicating that they were heterosexual or bisexual-sexual were eliminated from 45 the subject p00 study: The rel: based Pride hat relationship bet and PANAS. i strictly theoretir relationship of t data called for a consictions beg A combi analysis for hi g} Stfixtures are “n Pmposed by Ch fm‘” aliaiysis ( Personality Cons be hYPOIheSizec‘ 1983) argue tha- factor SOiutiOns‘ WWW an (HWY test the and 300an of that the comma 46 the subject pool. The following analyses paralleled the research questions of the study: E l . l . E S] E . l The relationship between Gay Identity-based Shame and Gay Identity- based Pride had never before been tested empirically, and neither had the relationship between each of these two variables and the variables of self esteem and PANAS. However, shame and affect theories provided a cogent, albeit strictly theoretical, basis for a priori predictions about the nature of the relationship of these constructs to one another. Consequently, lack of empirical data called for a more exploratory model of analysis whereas strong theoretical convictions begged for a confirmatory model. A combination strategy of exploratory and confirmatory models of factor analysis for higher-order personality structures (i.e., higher order personality structures are “meta” constructs such as Self-esteem and PANAS) has been proposed by Church and Burke (1994). These authors argue that exploratory factor analysis (EF A) has been utilized almost exclusively for the study of personality constructs to date and are advantageous when no a priori structure can be hypothesized. Advocates of confirmatory factor analysis (CF A)(e.g., Long, 1983) argue that exploratory techniques have limitations including unidentified factor solutions, indeterminate factor solutions, and lesser flexibility than confirmatory analysis (Church & Burke, 1994). In addition, confirmatory models directly test theoretically driven hypotheses a priori and offer significance tests and goodness of fit indices (Church & Burke, 1994). Finally, these authors argue that the confirmatory factor analysis requires a substantially geater proportion of the variance to be 50-60% of variant An obviou specificity needed study, provides a r psychological con and cons of both t fora combination and better defined well n the nature PANAS could be . an i“depth discus {ml analysis for (1994). W Before a c relationships of Si analyses were cm Along with the re and Pride items “ analysis With Obli oflhase “ales, E Uti . llled to Com/en 47 the variance to be accounted for, unlike exploratory models where accounting for 50-60% of variance is acceptable (Church & Burke, 1994). An obvious problem with the CFA approach lies with the high degree of specificity needed by the CFA model to obtain significance. Shame theory, in this study, provides a theoretical model for the uniqueness of shame as a psychological construct, which could be tested by CFA. Due to the various pros and cons of both the CF A and BE A approaches, Church and Burke (1994) argue for a combination of these two strategies. Utilizing both, a clearer, more concise, and better defined assessment of the factor structure of gay shame and pride, as well as the nature of the relationships between shame, pride, self-esteem and PANAS could be obtained than by utilizing either of these approaches solely. For an in-depth discussion of the pros and cons of confirmatory versus exploratory factor analysis for higher-ordered psychological constructs, see Church and Burke (1994). Ines-«r- .. . r2 --. 'a .. as ....- .u 'h- .. a . “a an Before a confirmatory factor analysis was conducted to examine the relationships of shame and pride to other psychological measures, additional analyses were conducted on the Gay Identity-based Shame and Pride Scales. Along with the reliability analyses previously described, the entire set of shame and pride items were analyzed by an exploratory principle components factor analysis with oblimin oblique rotation to examine the underlying factor structure of these scales. Eigenvalues, scree plots and measures of communality were utilized to converge a set of factors that represented the essence of Gay Identity- based Shame and ' pride scales was d subsequent confir i l . l . i: a For resean of the Gay ldentit was utilized. In ad using a confirmatc 1. What is the rel bucd pride? 5 0ft continuum Oflh0gona] cor the Possibility about being g3 based shame 3 Based Shame. Pride were ind m“ tw0 scal assumed that r mnfinuum. Ir confil’lltatory 1 48 based Shame and Pride. Once the underlying factor structure of the shame and pride scales was determined, the resulting shame and pride factors were utilized in subsequent confirmatory factor analysis. For research question 1 (see below), an intercorrelation matrix composed of the Gay Identity-Based Shame-Pride Scales (with both shame and pride scores) was utilized. In addition, the correlation between Shame and Pride was calculated using a confirmatory factor analysis (see below). 1. What is the relationship between gay identity-based shame and gay identity- based pride? Several authors have described shame and pride as opposite ends of a continuum (Nathanson, 1992), but others predict that shame and pride are orthogonal constructs (Kaufman, 1992). Exploration of this issue addressed the possibility of a gay man having both a sense of pride and a sense of shame about being gay. The intercorrelation was examined between gay identity- based shame and gay identity-based pride, as measured by the Gay Identity- Based Shame-Pride Scales, to determine whether identity shame and identity pride were independent, orthogonal constructs. If the intercorrelation between these two scales exceeded their internal consistency values, then it was assumed that they are not orthogonal constructs and they may lie on one continuum. In addition, this correlation was estimated utilizing a confirmatory factor analysis. 2. W has get ide Wa hit 3. H0' sub ‘l’lll‘ txplora scores . Scale, a latter a Rosent factor a tempo1 firm live Sca Shame 49 :- .... . - . .-. s -: m loH'-'I0‘ .- - ."n .u 'u; 2. What is the relationship between gay identity-based shame and gay identity- based pride to self-esteem? Although other studies have found that shame in general and low self-esteem may be highly related, it was unknown how identity-based shame about being homosexual and self-esteem are related. Was it possible for an individual to have fairly high self-esteem but still have high gay identity-based shame? 3. How are gay identity-based shame and gay identity-based pride related to ' positive and negative affectivity? Can gay identity.based shame and pride be subsumed under positive and negative affectivity? W To address research questions 2 and 3 (above), two principle-components exploratory factor analysis were conducted. The first EFA utilized total scale scores of the Gay Identity-Based Shame-Pride Scales, the Rosenberg Self-Esteem Scale, and the PANAS scales. The second EFA was a principal-components factor analysis using the items of the Gay Identity-Based Shame-Pride Scales, The Rosenberg Self-Esteem Scale, and the PANAS scales. Consequently, total scale scores and all scale items fi'om each scale were factor analyzed separately. This strategy has been modeled after the principal- components EF A study conducted by Tinsley, Bowman and York (1989). In the first EFA, principal-components analysis was conducted on the scores from the five scales of the study (the Rosenberg Self-esteem scale, the Gay Identity-based Shame scale, the Gay Identity-Based Pride scale, the Positive Affectivity scale, 50 and the Negative Affectivity scale) to determine if gay identity-based shame, gay identity-based pride, self-esteem, and positive and negative affectivity are separate factors or if these scales comprised less than five independent factors. Possibly, several of the scales could be subsumed under a common factor. The second EF A utilized a principle components analysis of all items that comprise the Gay Identity-based Shame-Pride Scale, the Rosenberg Self-esteem scale and the PANAS measures to see if the scales’ items comprised several discernible latent structures. Because the nature of the relationship between shame and the other constructs of the study had not been empirically tested, items from one measure (e.g., such as Gay Identity—based Shame) may also have loaded on other measures (e.g., such as self-esteem). Therefore, a separate analysis of the individual scale items was also conducted. Resulting factors were examined to determine what each factor may be tapping. This analysis provided exploratory factor structures that provide discriminant validity for latent variables of shame versus low self- esteem. CmfinnatemfiactQLAnalxsiuCEA) Research questions #2 and #3 below were also addressed using a confirmatory factor analysis. This CF A used the covariance structure analysis program CALIS, in the SAS statistical program. Maximum likelihood estimation was used and all analyses were performed on covariance matrices to estimate the parameters hypothesized by theory. Indices quantified how well the hypothesized model reproduces or fits the observed covariance data. theorl Gay 1 hie: to be facto: coast items Gay 1 hit I progi with i 51 The model tested by CFA consisted of the most parsimonious model with theoretical underpinnings that also included the five major constructs in question: Gay Identity-based Shame, Gay Identity-based Pride, Self-Esteem, Positive Affectivity, and Negative Affectivity. Each item from all scales was only allowed to be an indicator of only one of the five higher-order constructs. All secondary factor loadings were fixed at 0.0. To test this five factor model, an 80 x 80 covariance matrix was constructed utilizing data from all the items of the Positive Affectivity Scale (10 items), Negative Affectivity Scale (10 items), Rosenberg Self Esteem (10 items), Gay Identity-based Shame (25 items) and Gay Identity-based Pride (25 items). Due to the large number of items comprising each of the five factors, data transformation was necessary to reduce the size of the matrix to allow the program to run successfully. As the GIB Shame and Pride Scales were the largest with 25 items each, they were targeted for distillation. The method for reducing the number of shame and pride items is as follows. (a) The 50 items of the Gay Identity.based Shame and Pride Scales were factor analyzed. This EFA, with oblimin oblique rotation, examined the underlying factor structure of these scales. Eigenvalues, scree plots and measures of communality were utilized to converge on a two factor solution. These factors appear to represent the essence of Gay Identity-based Shame and Pride. (b) Composite factors, based on the summation of the items gleaned from the EPA of shame and pride, were calculated to represent the gay shame and gay pride factors and fitrther reduce the size of the covariance matrix. Half of the shame items were randomly selected and summed sea P0: 52 to compose the first shame factor and the other half became the second shame composite factor. The same procedure was utilized with the items of the pride scale, producing two pride composite factors. In sum, the original 50 items of the Gay Identity-based Shame and Pride Scales were reduced to 4 composite items, two shame and two pride. The resulting covariance matrix was a 34 x 34, and was within the constraints of the statistical package. This CF A was utilized to test and confirm a five factor model of the survey items that accounted for a large amount of the variance and maintained a separate nature of these five constructs: Gay Shame, Gay Pride, Self-Esteem, Positive Affectivity and Negative Affectivity. WW5. One advantage to utilizing a CFA approach to test this five factor model is the ability to compare the goodness of fit for this model with other plausible models. That is, although shame theory would predict that shame and pride are distinct from one another and from self- esteem and PANAS, arguments could be made for testing additional a priori models utilizing differing numbers of latent variables. One such possible competing model would argue for greater parsimony and predict that a two factor solution utilizing self-esteem, positive affectivity and pride as one factor and negative affectivity and shame as the other, may converge with higher goodness of fit indices than the five factor model. This model, which is not based on any theoretical position, would simply argue that the first latent factor would tap positive psychological well-being, whereas the second would measure a negative dimension. Accordingly, an additional CFA was run testing 53 this alternative two factor model with the five factor model to determine which model best fit the data. Several goodness-of-fit indices and a chi-square difference test were utilized to make this determination. HT Ell . E lSl IE'I 4. What is the utility of gay identity-based shame and gay identity-based pride in predicting substance abuse? The final question was to assess the utility of these constructs in explaining a very important social issue: high levels of substance abuse in gay male populations. A multiple hierarchical regression analysis was used to determine whether high gay identity-based shame scores predict high alcoholism proclivity scores, as measured by the MAC, and also whether high scores on the gay identity-based pride scale predict low alcoholism proclivity scores, as measured by the MAC. The theory, design, and measures of this study made this question well suited for a hierarchical multiple regression analysis (Kerlinger & Pedhazur, 1973) The MAC is a continuous dependent variable that was regressed upon several continuous independent variable measures (subject demographics, Positive and Negativity Affectivity Scales, Rosenberg Self-Esteem Scale, and the Gay Identity-based Shame-Pride Scales). A hierarchical regression analysis was appropriate because there were theory driven reasons for the ordering of the independent variables. Specifically, this analysis explored whether Identity-based Shame-Pride issues account for unique variance in the prediction of substance abuse proclivity, over and above the other constructs. The regression equation may be represented conceptually in Figure l. iali tit 54 Insert Figure 1 about here C lll'l' El 5' Due to the large size of the sample, it was possible to repeat several of the preceding analyses utilizing a split-sample technique (DeVellis, 1991). The rationale for the use of a cross-validation analysis was to test the stability of the factor structure of the Gay Identity-based Shame and Pride Scales (Kim & Mueller, 1978), and the predictive validity of the regression analyses. Cross validation analyses consisted of repeating the CFA and regression analyses above, with the split-samples. The sample of 971 gay men (n = 971 to correct for social desirability response bias from the Marlowe-Crowne) was divided in half to produce two samples of 486 subjects and 485 subjects each. Next, two identical confirmatory factor analyses were conducted on the two samples. The model tested by CFA was the five factor model identical to the one tested previously with the entire data matrix. It included the five major constructs in question: Gay Identity-based Shame, Gay Identity-based Pride, Self-Esteem, Positive Affectivity, and Negative Affectivity. As before, each item from all scales was allowed to be an indicator of 55 only one of the five higher-order constructs. All secondary factor loadings were fixed at 0.0. Again, due to the large number of questions for each of the factors in the model, composite factors based on the summation of multiple items, had to be calculated for the Gay Identity-based Shame and Pride Factors. Two composite items were computed for both the Gay Shame and Gay Pride factors respectively. Maximum likelihood estimation was used and all analyses were performed on covariance matrices. Indices quantified how well the hypothesized model reproduces or fits the observed covariance data. Finally, two identical regression analyses were conducted on the two samples. As before, a hierarchical multiple regression analysis was utilized. The MAC was regressed upon the independent variable measures (subject demographics, Positive and Negativity Affectivity Scales, Rosenberg Self-Esteem Scale, and the Gay Identity-based Shame-Pride Scales). Specifically, this analysis explored whether the two halves of the sample would produce results that were consistent with one another. --—-—-‘...U Va CHAPTER FOUR RESULTS EiletPhase W Respondents for the pilot phase of this study were taken from the National Community Masterfile. NCM supplied the names of 500 direct mail-responsive gay men. These names were stratified by geographic location. Of these 500 questionnaires, 200 completed and returned surveys were useable, providing a 40% return rate. Twenty-two were returned to the researcher incomplete for a variety of reasons (see Table 4). Six were received by men who related that they were not gay and that the inclusion of their name on a list of gay men was inappropriate. No response was elicited for 278 (56 %) of the surveys from either the men to whom they were mailed, or by any other source. Non- respondents had been asked to complete the demographic information only, if they were unwilling to complete the rest of the survey. No subjects elected this option. Insert Table 4 about here 56 In 57 El'l'l'!l'ElGIl'-l IS] 215] Reliability estimates for the Gay Identity-based Shame and Gay Identity- based Pride Scales were calculated for the 200 gay male respondents and consisted of inter-item correlations, item-scale correlations, squared multiple correlations as an estimate of communality, and the alpha coefficients for each scale. The inter-item correlations for the Gay Identity-based Shame Scale are reported in Table 5, and for Gay Identity-based Pride are reported in Table 6. The item scale correlations, estimates of communalities, and item means and variances for the Gay Shame Scale are displayed in Table 7, and these statistics for the Gay Pride Scale are displayed in Table 8. Insert Tables 5, 6, 7 and 8 about here The alpha’s for the Gay Identity-based Shame Scale and the Gay Identity- based Pride Scale were or = .94 and or = .93 respectively. Given the strength of alpha for both scales, and the unlikelihood that removing individual scale items would significantly increase either alpha, it was decided that no items be deleted from either scale. Also, subjects reported that the survey directions were clear and most were able to complete the survey in the projected time frame of 20 to 30 58 minutes. Therefore, no changes were made in the survey packet. Consequently, subjects for the main study received identical survey packets as the pilot subjects. Main Study ResmnseRate Respondents for the main study were also taken from the National Community Masterfile. NCM supplied the names of an additional 1000 direct mail-responsive gay men. These names were stratified by geographic location. Of these 1000 questionnaires, 783 completed and returned surveys were useable, providing a 78% return rate. Sixty-eight were returned to the experimenter for a variety of reasons (see Table 9). An additional 18 were received by men who related that they were not gay and that the inclusion of their name on a list of gay men was inappropriate. Only thirteen percent or 131 of surveys did not elicit a response from either the men to whom they were mailed, or by any other source. Insert Table 9 about here Because all respondents (i.e., both pilot phase and main study) received identical questionnaires, all were treated as one sample of 983 subjects for the following descriptions and analyses. 59 Dsmmnhiss The combined sample was composed of 983 males who were: 95% Caucasian, 1% black, 2% Hispanic, 1% Asian, and 1% Native American. The average age of the respondents was 47.0 years (SD = 11.25). The majority of the sample (68%) was between 32 and 56 years of age. The average education of the sample was at the Bachelor’s degree level and the respondents had a mean individual income of $40,000 to $59,000. The majority of the subjects (82%) were employed. The average relationship status indicated by respondents was “single, multiple partners,” although the modal category was “single, no sexual partners”. Using a modified version of Coleman’s Assessment of Sexual Orientation Questionnaire (1987), the majority (88%) indicated that they identified themselves as “exclusively homosexual” and would like to identify themselves as the same in the future. Insert Tables 10 and 11 about here 60 Finally, 72% indicated that they were very comfortable with their current sexual orientation. A numerical breakdown for all subject demographics can be found in Tables 10 and 11. Wills: Descriptive statistics for all scales utilized in the study can be found in Table 12. Internal consistencies for the study variables can be found on the diagonal of the correlation matrix of Table 13. Insert Tables 12 and 13 about here The Marlowe-Crowne Social Desirability Scale was used to eliminate subjects who may present themselves in a more favorable fashion than might be truthful. The mean on the Marlowe-Crowne was 19.1 with a standard deviation of 3.2. Ten subjects scored higher than two standard deviations above the mean and were eliminated from further analysis. The internal consistency for the Marlowe- Crowne for this study was .77. The MacAndrew Substance Abuse Proclivity Scale bears closer descriptive examination for two reasons: (1) it serves as the dependent measure for the regression analysis of this study, and (2) previous studies reported high 61 rates of substance abuse in gay men (see Table 2). Scores of this sample ranged from 8 to 34, with a mean of 18.9 and a standard deviation of 4.09. The internal consistency of the MAC for this sample was .52, which is the lowest indice of reliability of the study. Clinically, a cutting score of 24 or higher is able to identify substance abuse with over 80% accuracy (Greene, 1980). Using this criterion, 123 subjects or 13% of the sample scored 24 or higher on the MacAndrew. This would indicate that this sample of gay men have a proclivity to abuse substances that is much closer to the national average for non-gay persons (approximately 10%; NIAAA, 1978) than any other studies have previously reported. A correlation matrix was constructed to examine the relationships between the various scales of the study (see Table 13). As predicted, Gay Identity-based Pride has moderate, yet significant correlations with the other variables of the study, and all but one of these relationships are in the directions predicted. Gay Identity-based Pride correlates positively with self-esteem and positive affectivity, and negatively with negative affectivity. This lends support to the construct validity of identity-based pride, that is, it measures a positive view of identity distinct from self-esteem and PANAS. However, the Gay Identity-based Pride 62 Scale has a positive relationship, [ =.22, with the MacAndrew Substance Abuse Proclivity Scale, which would not be predicted theoretically. The Gay Identity-based Shame Scale also behaved as predicted with self- esteem, positive affectivity and negative affectivity yielding negative, negative and positive correlations respectively. However, like the Pride Scale, the Shame Scale does not correlate with the MAC in the direction predicted by theory. A slightly negative, non-significant relationship exists between Gay Identity-based Shame and the MacAndrew, r = -.03. In firrther analysis, in which this relationship is explored within the context of other constructs, this direction of relationship changes to a positive association (see regression analysis and Discussion section below). 11 . m l D l . Ci Differences were found between study respondents on the measures of gay shame, gay pride, self-esteem, positive affectivity, and negative affectivity when compared by their membership in various groups defined by the demographic variables of the study (see Tables 14, 15, and 16). Significant differences were found for education level. Subjects who had higher levels of education (college and graduate level degrees) had significantly lower levels of substance abuse proclivity (M = 18.40, SD = 4.1) than those with less education (M = 20.43, SD = 63 3.9), t = 6.97, p < .05. Significant differences were found for subjects with higher income levels ($25,000 a year or more). Subjects with greater income scored significantly higher on self-esteem (M = 34.18, SD = 5.3), and positive affectivity (M = 37.12, SD = 6.1) and significantly lower on gay shame (M = 36.10, SD = 12.4) than those subjects with lower income (less than $25,000) (self esteem: M = 31.84, SD = 6.6; positive affectivity: M = 34.51, SD = 7.4; gay shame: M = 39.74, SD = 14.6; 1s, respectively, were 1; = - 4.71, n <.05, t= - 4.59, p <.05; t= 2.61, n <05). Insert Tables 14 and 15 about here The way in which subjects identified their current sexual orientation, comfort with their sexual orientation, and preferred sexual orientation for the future was utilized to compare their scores on gay shame, gay pride, self-esteem, positive affectivity, negative affectivity, and substance abuse proclivity. Significant differences were found on these variables by sexual identification and comfort with sexual identification categories. Subjects who identified as exclusively or predominately gay scored significantly higher on gay pride (M = 95.19, SD = 16.8), self-esteem (M = 33.90, SD = 5.4) and positive affectivity (M = 36l their l 7.48, exclu sham p<.0: oner (M = who prid 6.2 <0." = 36.82, SD = 6.2) than did those that indicated they were bisexual or unsure of their orientation status (gay pride: M = 70.61, SD = 26.3; self-esteem M = 30.61, SD = 8.7; positive affectivity M = 33.32, SD = 10.4; ts, respectively, were 1 = 7.48, p<.05, t = 3.11, p<.05, t = 2.88, p < .05). In addition, those that identified as exclusively or predominately gay scored lower (M = 36.52, SD = 12.5) on gay shame than bisexuals or those that were unsure (M = 42.14, SD = 23.8) t = - 2.27, p<.05. Likewise, those indicating that they were comfortable with their current orientation had higher levels of gay pride (M = 97.10, SD = 15.56), self-esteem (M = 34.38, SD = 5.17), and positive affectivity M = 37.20, SD = 6.1) than those who indicated that they were uncomfortable with their present orientation (gay pride: M = 70.38, 512 = 17.50, t = 15.71, p <05; self-esteem: M = 28.53, 512 = 6.2, t= 10.27, p <.05; positive affectivity: M = 32.27, SD = 7.0, t= - 8.83, p <.05). Subjects indicating comfort also had lower gay shame scores M = 34.50, SD = 9.65) and lower negative affectivity M = 17.70, SD = 6.2) than those who were uncomfortable with their current orientation (gay shame: M = 57.02, SD = 20.16, t= - 18.79, p <.05; negative affectivity: M = 23.78, SD = 8.1, t = - 8.83, p <.05). Subjects who would like to continue to identify as gay in the future also had significantly higher levels of gay pride M = 95.79, SD = 16.3), self-esteem M = 34.03, SD = 5.3) and positive affectivity M = 36.89, SD = 6.1), and lower 65 scores on gay shame M = 36.06, SD = 11.79) and negative affectivity M = 18.10, SD = 6.4), compared to those not wishing to identify as predominately gay in the future (gay pride: M = 74.12, SD = 23.2, t = 9.44, p <.05; self-esteem: M = 30.16, SD = 7.75, t = , p <.05; positive affectivity: M = 33.91, SD = 8.9, t= 3.45, p <.05; gay shame: M = 46.23, SD = 22.91, t = - 5.86, p <.05; negative afi‘ectivity: M= 21.35, 512 = 9.4,;= - 3.63, p <.05). Insert Table 15 about here Subjects differed on levels of gay shame, gay pride, self-esteem, positive affectivity, and negative affectivity by their relationship status (see Table 15). Respondents were grouped into three categories: (1) sexually monogamous, (2) sexually active but not exclusive, and (3) no sexual partners. Analysis of variance between these three groups yielded significant main effect difi'erences for gay shame (E (2, 974) = 23.70, MSE = 160.24), gay pride (E (2, 974) = 32.00, MSE = 292.51), self-esteem (E (2, 974) = 9.67, MSE = 30.34), positive affectivity (E (2, 974) = 15.74, MSB = 39.38), and negative affectivity (E (2, 974) = 4.04, MSE = 43.65). Post hoc contrasts utilized scheffe tests to clarify differences among the three relationship groups. In all cases, subjects indicating that they had no sex part1 seru (M = aliet 15.9 111011 34.4 stud self- neg; 18,8 partners scored significantly different than those with either one or multiple sexual partners. Those without partners scored significantly lower on gay pride M = 87.50, SD = 19.04), self-esteem M = 32.59, SD = 5.55), and positive affectivity M = 34.92, SD = 6.68) and higher on gay shame M = 41.01, SD = 15.90) and negative affectivity M = 19.10, SD = 6.99) than those with monogamous sex partners (gay pride: M = 97.67, SD = 17.26; self-esteem: M = 34.45, SD = 5.74; positive affectivity: M = 37.39, SD = 7.07; gay shame: M = 33.79, SD = 10.40; negative afi‘ectivity: M = 17.40, SD = 6.18) or multiple sex partners (gay pride: M = 97.00, SD = 15.82; self-esteem: M = 34.23, SD = 5.38; positive affectivity: M = 37.45, SD = 5.62; gay shame: M = 35.55, SD = 11.46; negative affectivity: M = 18.24, SD = 6.56). No significant differences were found between groups for substance abuse proclivity. Finally, no significant differences were found for ethnicity on any of the study variables of shame M = 36.67, SD = 12.9), pride M = 94.61, SD = 17.4), self-esteem M = 33.87, SD = 5.4), positive affectivity M = 36.71, SD = 6.29), negative affectivity M = 18.29, SD = 6.6) or substance abuse proclivity M = 18.88, SD = 4.0) for anglo versus respondents of color (shame: M = 36.60, SD = 12.8, t = .04, p = .47; pride: M = 93.10, SD = 21.3, t = .61, p = .54; self-esteem: M = 32.92, 512 = 7.3, ;= 1.2, p = .23; positive affectivity: M = 37.38, 512 = 7.5, 1 cont lactc stall iden blul fact lror. mil Pn'c 67 = - .75, n = .45; negative affectivity: M = 18.21, SD = 6.9, t = .08, p = .93; substance abuse proclivity: M = 19.86, SD = 5.3, t = - 1.69, p = .10). .oovos ,.,. at: 0 .1‘ ro 0‘18-1-‘0 1.11'110'00‘ ..'-.~ A principal components factor analysis with oblimin oblique rotation was conducted on all items of the Gay Identity-based Shame and Pride Scales. The factor analysis was used to examine the collective factor structure of these two scales to distill and concentrate the items that best reflected the constructs of identity-based pride and shame. Insert Figure 2 about here Multiple criteria were used to determine the number and composition of the factors from this analysis. First, the number of factors to be justifiably extracted fiom the shame and pride scales was determined as follows: (a) eigenvalues for all emerging factors had to be equal to or greater than one; and (b) a scree plot was examined to eliminate any factors with eigenvalues greater than one that were located in the bottom, or scree of the plot (see Figure 2). Two factors met these criteria and were extracted fiom this EFA of all the gay identity.based shame and pride items. Next, the membership of the individual items comprising these two ex: eitl 0f ' its] acc fat Ide dc: of the the 68 factors was determined as follows: (a) principally, all items required a factor loading magnitude of .40 or higher on either of the two factors; and (b) secondarily, percentage of variance criteria, or communality, for each item was examined. A sizable amount of the item’s variance had to be accounted for by either of the factors. Ifthese two criteria were met, the item was included on one of the two factors. This EFA of the gay identity-based shame and pride scale items, utilizing the preceding strategy, converged on a two factor solution, accounting for 43.7 % of the variance. The composition items of each of these factors are displayed in separate tables (see Tables 17 to 18). Insert Table 17 and 18 about here Factors 1 and 2 from this analysis represent the latent factors of Gay Identity-based Shame and Gay Identity-based Pride, respectively. The first factor deals with the negative aspects of gay identity. Factor 1 (Table 17), is composed of 21 items from the Gay Identity-based Shame Scale and one item from the Gay Identity-based Pride Scale and accounts for 33.1% of the variance. The item from the Pride Scale is correlated negatively with this factor making it consistent with the content of the shame items. Factor 1 taps several aspects of gay shame and neg: “Be: som [den posi with theii C0111 The item Shar lden 69 negative gay identity-based affectivity. An example of an item from this factor is, “Because I’m gay, I feel as if I am somehow defective as a person, like there is something basically wrong with me.” Factor two (Table 18) deals with Gay Identity-based Pride. It accounts for 10.6 % of the variance, and appears to tap a positive gay affectivity. An example item of this factor is, “Being gay fills me with joy.” (see Table 18). The validity and efficacy of these factors to serve as shorter versions of their parent scales is apparent in a corelational analysis (see Table 19). Table 19 compares the factors of shame and pride with their correspondent parent scales. These comparisons indicate that the Factor 1 shame items and the Factor 2 Pride items are very indicative of the parent scales fiom which they were drawn. Factor 1 correlates nearly perfectly (r = .98, p < .001) with the parent Gay Identity-based Shame Scale. Insert Table 19 about here Factor 2 also correlates substantially ([ = .90, p < .001) with it’s parent Gay Identity-based Pride Scale. The correlation of Factor 1 with Factor 2 is 1: = -.43 , 3C3 lde she 01" 70 p < .001, consistent with the relationship between the parent shame and pride scales, [ == -.52, p < .001. Factors 1 and 2 represent the underlying factor structure of the Gay Identity-based Shame Pride Scales, and appear to represent the constructs of gay shame and pride in a concentrated and parsimonious fashion. Therefore, these two factors will be used in the subsequent confirmatory factor analysis. HmtheseurndAnalm The following three research questions of this study were analyzed by a combination of exploratory and confirmatory factor analyses. The results of each of these factor strategies will be discussed below. 1. What is the relationship between gay identity-based shame and gay identity-based pride? The intercorrelation between gay identity-based shame and gay identity-based pride, as measured by the Gay Identity-Based Shame-Pride Scales, was found to be r = -.53 (p < .001). The intercorrelation between these two scales does not exceed the internal consistency of either the Gay Identity- Based Shame Scale, which has an internal consistency of a = .94, or the Gay Ida and dicl 2.11 has 3.1-. pos sub Sha and 0% 1111 71 Identity-Based Pride Scale on = .93. This suggests that gay identity-based shame and gay identity-based pride are related, though distinct constructs, not dichotomous ends of a singular continuum. El' l' ['1 .4 1'1 1E 1111115 2. What is the relationship between gay identity-based shame and gay identity- based pride to self-esteem? And: 3. How are gay identity-based shame and gay identity-based pride related to positive and negative affectivity? Can gay identity-based shame and pride be subsumed under positive and negative affectivity? A principal-components factor analysis was conducted using the five scale m from the survey (the Rosenberg Self-Esteem Scale, the Gay Identity—based Shame Scale, the Gay Identity-based Pride Scale, the Positive Affectivity Scale, and the Negative Affectivity Scale) to determine the number of latent factors these overall scales comprise (see Table 20). Only one factor emerged from the principal components analysis with an Eigenvalue greater than or equal to one. Due to the emergence of only one factor, rotation was not possible. sh fa: Sb lac 72 Insert Table 20 about here This single factor accounted for 56.6 % of the variance. This analysis suggests that all five of these scales, when analyzed as scale scores, rather than scale items, are measuring a “positive self-vievnf’ latent variable with which all of these scales share a strong degree of communality. In addition, as expected, the shame and negative affectivity scales correlate in the negative direction to the factor whereas the other scales correlate in a positive direction. Next, a principal components factor analysis was conducted on all m that comprise each of the instruments of this study, the Gay Identity-based Shame-Pride Scales, the Rosenberg Self-esteem Scale and the PANAS. The factor analysis was used to determine any discernible latent constructs being tapped and the scale items comprising these factors. Insert Figure 3 about here of 73 Eigenvalue, scree (see Figure 3) and percentage of variance criteria converged on a five factor solution, accounting for 48.8 % of the variance. Each of the five factors is displayed in a separate table (see Tables 21 to 25). Insert Table 21 to Table 25 about here The five factors that emerged from this analysis can be organized into factors tapping: gay shame, gay pride, self-esteem, positive afi‘ectivity and negative affectivity. The first factor from this analysis deals with the negative aspects of gay identity. Factor 1 (Table 21), which accounts for 28.1% of the variance, is composed of 21 items from the Gay Identity-based Shame Scale, 1 item from the Gay Identity-based Pride Scale, and 1 item from the PANAS negative activity scale. The 1 item from the Pride Scale is correlated negatively with this factor making it consistent with the content of the shame items. The negative affect from the PANAS scale is the word “Ashamed” which appears to sum up the tone of this factor, it taps several aspects of gay shame and negative gay identity-based affectivity. An example of an item from this factor is, “Because I’m gay, I feel as 74 if I am somehow defective as a person, like there is something basically wrong with me.” Factor 1 from this analysis is 151mm to the Factor 1 that emerged from the exploratory factor analysis of the items of the Gay Identity-based Shame and Pride Scales, except for the inclusion of the one PANAS item in the current EF A. This provides firrther empirical evidence of the reliability and stability of the factor structure of gay identity-based shame. Factor two (Table 22) emerged as the Negative Affectivity Scale of the PANAS. All ten items from the Negative Affectivity Scale are present in this factor, accounting for 7.9 % of the variance. The third factor deals with Gay Identity-based Pride. It accounts for 6.1 % of the variance, and appears to be the antithesis of factor one; it taps a positive gay affectivity. An example item of this factor is, “Being gay fills me with joy.” (see Table 23). Like the shame Factor from this analysis, Factor 3 is closely approximated by the earlier EFA of all items from the Gay Identity-based Shame and Pride Scales. Two pride items, however, appear on the original EF A of the shame and pride scales that are not captured by this analysis, otherwise these factors are identical. Again, this supports the reliability and stability of the shame and pride factors. 75 Factor 4 from this principle components factor analysis of all items from all scales is simply the reconstituted scale of positive affectivity. The Positive Affectivity Scale of the PANAS, extracted as Factor 4 of this analysis (see Table 24), accounts for 3.9 % of the variance. Finally, all ten items of the Rosenberg Self-Esteem Scale comprise Factor 5, and account for 2.8 % of the variance (see Table 25). The results of this EFA solidify and substantiate the legitimacy of the factor structure of the Gay Identity-based Shame and Pride Scales as represented by shame Factor 1 and Pride Factor 2 from the EFA of all items of the Gay Identity-based Shame and Pride Scales. After including the two pride items from this second EF A that were not captured in the first analysis, the resulting Gay Identity-based Shame and Pride Factors appear to capture the essence of gay identity shame and pride, and were used in subsequent analysis. They will be referred to as Shame-F actor and Pride-Factor in further analyses and discussion, to differentiate them from their parent scales, Gay Identity-based Shame and Pride. The relationship of the Shame-F actor and Pride-Factor with other measures of the study and with their parent scales, Gay Identity-based Shame and Gay Identity-based Pride can be found in Table 26. As with Table 19, Table 26 76 shows that these factors are very highly correlated with their parent scales, and they behave in a similar fashion to these parent scales in their strength and direction of relationship with the other variables of the study. Insert Table 26 about here In sum, this exploratory factor analysis of all of the items from the scales in this study indicate tlnat the constructs of Gay Identity-based Shame and Pride are distinct from, but related to, the other psychological constructs of the study. Botln the PANAS and the Rosenberg Self-Esteem Scales emerged as separate from each other and from the items of the Shame and Pride Factors. Consistent with earlier EF A analyses of the Gay Identity-based Shame and Pride Scales, the larger Shame and Pride Scales appear to be comprised of several sub-factors, as not all The Gay Identity-based Shame and Pride items loaded with a sizable factor loading magnitude on factors 1 and 3 respectively. The items that did factor out on Factors 1 and 3, however, appear to be relatively good indices of their larger, parent scales, as they are nearly identical to the results of the separate EFA on just the shame and prides scales. 77 An 3 mm model for the CFA was derived from the theoretical tenets of shame theory, that would predict that shame, pride, self-esteem and the PANAS are separate, but related psychological constructs. This five factor model has initially been substantiated by the EF A, which produced a factor structure of five latent variables. The five factors from the EFA, however, accounted for only 48.8% of the total variance. Therefore, a CFA was conducted on the items of the Rosenberg Self- Esteem Scale, the PANAS and two composite items representing the Gay Identity-based Shame-Factor and two composite items representing the Gay Identity-based Pride-Factor to test the fit of this model to the data. All items of the Shame-F actor and the Pride-Factor were utilized, but in the form of composite items, constructed as follows. Half of the Shame-F actor items were summed to create one Shame composite item and the other half were summed to become Shame composite item two. The same process was used to develop Pride composite items 1 and 2. The CFA treated these five factors as primary and did 78 not allow items to load on multiple factors (i.e., no secondary factor loadings were allowed). Insert Table 27 to 31 about here The five factor model provided an excellent fit to the data (see Table 27), and was parsimonious because none of the scales were allowed to load on multiple factors (see Tables 28, 29 and 30). This model of (511, N = 963) = 2026.320,p< .001) produced the following goodness of fit indices: NFI = .90 and CF I = .91, xz/df = 3.97. Conventional standards for a good fit for a CFA model are NFI (Norrned Fit Index) and GFI (Goodness of Fit Index) from .80’s to .90’s. Conventional standards for an acceptable chi-square/degees of freedom ratio, ledf, is < 5.0 (Bollen, 1989). Church and Burke (1994) concluded that current psychological theory generally is not sufficiently specific to obtain extremely high goodness of fit indices (GFI) (see also Borkenau & Ostendorf, 1990) of higher order personality variables and suggest interpretation guidelines for reasonable GFI’s in the range of mid .80’s and up. The goodness of fit (GFI) statistic tests the degree of acceptance that a given model fits the observed data (Long, 1983). The fit of the model is assessed 79 by comparing the observed covariance matrix S with the covariance matrix estimated by the equation 2 = A (I) A + O. 2 will not perfectly reproduce S, due to constraints on the model’s parameters, but the chi-square goodness-of-fit test compares theses two matrices to compute the degree to which they “fit” or approximate one another (Long, 1983). There are multiple good-of-fit indices and, unfortunately, there is limited consensus about which indices are best, and available interpretive guidelines become subjective (McDonald & Marsh, 1990). Therefore, use of multiple fit indices is recommended (Bollen, 1990). An additional CFA was run testing an alternative two factor model with the five factor model to determine which model would produce a better fit to the data. A competing two factor model argued for greater parsimony and suggested that a two factor solution utilizing self-esteem, positive affectivity and pride as one factor and negative afi’ectivity and shame as the other, would produce higher goodness of fit indices than the five factor model. The two factor model did not perform as well as the five factor model (see Table 27). This model On“ (526, N = 973) = 6389.805, p < .001) produced the following: NFI = .66 and GFI = .65, x’ldf= 12.15. 80 Insert Tables 32 to 35 about here The two factor model did not provide an adequate fit to the data (see Table 27). However, it was parsimonious because only a two factor solution was specified (see Tables 32, 33, and 34). This model argued that a first latent factor would tap positive psychological well—being, whereas a second would measure a negative dimension. The correlation between these two factors are r = -.57 (see Table 35). This model, however, does not show goodness of fit indices that are as high, or as close to the conventional acceptable range, as does the original five factor model. In addition, the model fit of the five factor and the two factor models were compared directly, using a chi-square difference test. This was possible as the models can be considered as nested (Long, 1983). For any two models (call them M, and M2), M, is nested in M, if M! can be obtained from M, by constraining one or more of the parameters in M2 to be fixed or equal to other parameters (Long, 1993). In this case M, (2 factor model) is nested in M2 (5 factor model) by constraining the shame and negative affectivity items to load on one factor and the items from pride, positive affectivity and pride to the second factor, instead of loading on five factors. Therefore, a chi-square difference test was computed as a 81 test of the hypothesis that the covariance matrices generated under M1 and M, are equivalent; that is Ho: f, (9! ) = f; (9;) imply no loss of statistical information in using M1. (Bentler & Bonett, 1980). This computation (see Table 36) provides a test that rejects the null hypothesis and indicates that model M, (2 factor model) cannot adequately and significantly represent the covariance matrix as well as M, (5 factor model). Insert Table 36 about here In sum, the CFA findings support the notion that when all items were set to load on a five factor model of shame, pride, self-esteem, positive affectivity and negative affectivity in a specified manner, the resulting covariance matrix would closely approximate the covariance matrix of the observed data. This, in fact, is what occurred. Figure 4 depicts the five factor confirmatory factor model of the study. Insert Figure 4 about here 82 The CFA provides answers for the three research questions above concerning the relationship between these five factors of this model. 1. What is the relationship between gay identity-based shame and gay identity-based pride? And: 2. What is the relationship between gay identity-based shame and gay identity-based pride to self-esteem? And: 3. How are gay identity-based shame and gay identity—based pride related to positive and negative affectivity? Can gay identity-based shame and pride be subsumed under positive and negative affectivity? The five factors, although discernible from one another, are also related (see Table 31). Consistent with the correlation analysis above of Gay Identity- based Shame to Gay Identity-based Pride, a correlation of r = -. 54 (p <.001) was obtained from the CFA. This supports the previous conclusion that Gay Identity- based Shame and Pride are related but distinct constructs, rather than dichotomous ends of a singular continuum. Gay Identity-based Shame, although yielding moderate correlation with the constructs of self-esteem ([ = -.58, p < .001), positive affectivity (r = -.28, p < .001), and negative afi’ectivity (r = .55, p < .001) is distinct from these constructs. In addition, the direction of these correlations is consistent with the direction predicted by theory (i.e., shame is negatively 83 correlated with self-esteem and positive affectivity and positively correlated with negative affectivity). Likewise, Gay Identity-based Pride is also moderately correlated with self-esteem (r = .60, p < .001), positive affectivity (r = .61, p < .001), and negative affectivity ([ = -.34, p < .001) yet is distinct from these factors. Gay Identity-based Pride also correlates in the direction consistent with theory (i.e., pride is positively correlated with self-esteem and positive affectivity and negatively associated with negative affectivity). In sum, the correlation analysis, exploratory factor analysis, and confirnnatory factor analysis of this study are all consistent in tlneir support for the distinction of Gay Identity-based Shame from Gay Identity-based Pride fiom one another and from the other higher order personality constructs of this study: self- esteem, and positive and negative affectivity. Furthermore, the relationships of shame and pride to these three constructs is consistent and in the directions predicted by theory. 111’ ['1 . 1 ll 1 'l 4. What is the utility of gay identity-based shame and gay identity-based pride in predicting substance abuse? To test the utility of Gay Identity-based Shame and Pride and the theoretical prediction that high internalized shame should predict substance abuse proclivity, a multiple hierarchical regession was utilized. Using the MacAndrew Substance Abuse Proclivity Scale as the dependent measure, the following measures were used as independent variables in this hierarchical order. All Demogaphic data were added as a block (Age, Ethnicity, Education Level, Employment Status, Income, Relationship Status, Current Sexual Orientation, Future Sexual Orientation and Comfort with Sexual Orientation); the Rosenberg Self-Esteem Scale, Positive Affectivity Scale, Negative Affectivity Scale were added as a second block; the Gay Identity-based Pride Scale, Gay Identity-based Shame Scale were added as a third block; and the Shame x Pride interaction term was added as a final block. (Inclusion of the interaction term was based on the assumption that shame might predict substance abuse proclivity more strongly at lower levels of pride than at higher levels of pride.) The results of this analysis can be found in Table 37. Insert Table 37 about here All linear regession equations are based on several assumptions about the distribution and relationships between variables and error terms (Glass & Hopkins, 1984). Violations of the assumptions for regression analyses were conducted for the regession model of the study. 85 Insert Figures 5, 6 and 7 about here One assumption tested was that the error or residuals were normally distributed at all points along the regression line (see Figure 5). SPSS plotted this histogram of the residuals, which follows a nearly perfect normal bell shape, indicating that this regession assumption was not violated. A second assumption tested was the concept of homoscedasticity, or that the variance of the residuals is homogeneous at all points along the regression line. Figure 6 is a graphic representation of predicted values plotted against residuals. A lack of any identifiable pattern suggests homoscedasticity. A final assumption is that there is a linear relationship between the Y’s and the standardized residuals of Y’s. This linear relationship is depicted in Figure 7. In sum, these tests show no violation of the assumptions of regession analysis, therefore interpretation of the regession results is permissible. As shown in Table 37, demogaphics (R2 change .09), self-esteem and PANAS (1?.2 change = .08) and Gay Identity-based shame and Pride (12.2 change = .04) each accounted for siglificant, unique increments in predictive variance. The 86 full equation accounted for 21% of the variance in the MacAndrew Substance Abuse Proclivity scale. The results indicate that several demogaphic variables significantly covary with substance abuse proclivity. Education level negatively covaries with substance abuse proclivity whereas age covaries positively. That is, higher levels of education are associated with lower rates of substance abuse proclivity, and increased age is associated with higher substance abuse proclivity. Of the established psychological constructs of the study, only the Positive Affectivity Scale siglificantly covaries with the MacAndrew, and this association is positive. That is, higher rates of experienced positive feelings are associated with higher substance abuse proclivity. The other constructs, self-esteem and negative activity, did not significantly predict substance abuse proclivity. Both the Gay Identityebased Shame and Gay Identity-based Pride Scales siglificantly predicted substance abuse proclivity on the MacAndrew. After accounting for the variance attributable to all the demogaphic variables and other psychological constructs, Gay Identity-based Shame scores positively, and significantly, covary with substance abuse proclivity, as predicted by shame theory. This relationship, which is significant, is in the opposite direction than that evidenced in the correlational analysis (see Table 13). These findings may be 87 the result of a suppressor effect and are considered in the Discussion section below. Gay Identity—based Pride scores also significantly covary in a positive direction with substance abuse proclivity, which is not predicted by tlneory. 111" 18 'll' Additional correlation and multiple hierarchical analyses that were similar to the previous analyses were conducted. However, instead of using the scale scores of the Gay Identity-based Shame and Pride Scales, these analyses used the Shame-Factor and Pride-Factor scores instead. The correlation of these factor scores with both the parent Shame and Pride Scales, as well as the other variables of this regression, can be found in Table 38. Insert Table 38 about here Because the Shame-Factor and Pride-Factor appear to be “refined” measures of the constructs of gay identity-based shame and pride, it was hypothesized that examining each of these EF A and CPA tested factors to see how they covaried Vvith the MacAndrew might provide a “clearer” picture of the relationship of Gay Pride. However, the results of this regression analysis (Table 39) simply replicate and do not differ substantially from the original regression analysis. 88 Insert Table 39 about here E lCl'l' El'ElSl Due to the study’s large sample, cross validation of tlne analyses of the study was possible. The same a priori five factor model for the CFA that was utilized previously was used for the split-sample. This model “confirms” the prediction that shame, pride, self-esteem and the PANAS are separate, but related psychological constructs. Two CFA’s were conducted on the items of the Rosenberg Self-Esteem Scale, the PANAS and two composite items representing the Gay Identity-based Shame-Factor and two composite items representing the Gay Identity-based Pride-Factor. All items of the Shame-F actor and the Pride-F actor were utilized, but in the form of composite items. The CFA treated these five factors as primary and did not allowed items to load on multiple factors (i.e., no secondary factor loadings were allowed). Insert Table 40 about here 89 The five factor model provided an excellent fit to the data (see Table 40). This model for the first halfoftlne data (12 (511, N = 491) = 1608.800, p < .001) had goodness of fit indices > .80 (NF I = .85 and GFI = .87). The second half of the data was similar to the first (x2 (511, N= 490) = 1680.629, p <.001) and had goodness of fit indices > .80 (NF I = .84 and GFI = .86). Conventional standards for a good fit for a CF A model are .80s to .908 and these statistics test the degee of acceptance that a given model fits the observed data (Long, 1983). In sum, these split-sample confirmatory factor analyses indicate support for the stability of the five-factor structure of the study’s variables; i.e., Gay Identity-based Shame and Gay Identity-based Pride are distinct from one another and from the constructs of self-esteem, and positive and negative affectivity. 5 1° -5 l B . l I Two identical regession analyses were conducted on the two samples. As before, a hierarchical multiple regession analysis was utilized. The MAC was regressed upon tlne following independent variables: subject demogaphies, positive and negativity affectivity, Rosenberg Self-Esteem Scale, and the Gay Identity-based Shame-Pride Scales. Specifically, this analysis explored whether the two halves of the sample would produce results that were consistent with one another, and with the regession analysis of the entire data set. Insert Table 41 about here As shown in Table 41, demographics (First sub-sample R2 change .13; Second sub-sample R2 change .07), self-esteem and PANAS (First sub-sample R2 change = .06; Second sub-sample R2 change .09) and Gay Identity-based Shame and Pride (First sub-sample R2 change = .03; Second sub-sample R2 change .04) each accounted for significant, unique increments in predictive variance in both halves of the sample. The first sub-sample’s equation accounted for 22% of the variance in the MacAndrew Substance Abuse Proclivity scale, and the second sub-sample accounted for 20%. The results are consistent across the two sub-samples, with two exceptions: (a) bisexual identification is associated with higher rates of substance abuse proclivity in the first but not in the second sub-sample; and (b) shame sigrificantly predicted MacAndrew scores only in the second sub-sample. Taken together, tlnese split-half regressions generally support the stability of the results of the original regression equation. CHAPTER FIVE DISCUSSION In various ways, the focus of this study has been on the construct of shame. How does shame associated with one’s identity relate to other psychological constructs and with one’s identity pride? And are these constructs of identity shame and identity pride useful? Do they tell us something unique that we did not know before? This study answered some of these questions, but raised new issues for investigation. The pilot study response rate was 40 % and the main study rate was 78% making the overall response rate for both 66 %. In their comprehensive review of mail studies and techniques to increase response rates, Kanuk and Berenson (1975) reported that many surveys fail to achieve a 50% response rate. This was confirmed by a meta-analysis of the effects of mail survey techniques on response rates (Fox, Crask & Kim, 1988). Therefore, a 66% response rate by the gay men of this study is quite high. The survey design and lay-out (see Appendices), as well as the pre and post-survey postcard methodology, may have aided in increasing survey response rate. Although this design was not inexpensive, the 91 92 metlnodology appeared to boost response rates. From sheer numbers and from additional comments written to the investigator, it appears that the gay men in this study were very interested in expressing their views. There is a vast difference in response rates for the two samples; subjects in the main study responded almost twice as often as those in the pilot study. This is unusual given that all subjects received the same packet. One possible factor for the discrepancies in response rate may have been the time of summer during which the mailings were conducted. The pilot subjects received their survey packets in early July, during the height of the summer, and may have not responded due to competing engagements and activities such as vacations and recreation. The main study subjects received their survey packets near the end of August, when many may have finished their summer-time diversions. This sample of gay men is primarily white and is somewhat older, more highly educated, and more wealthy than would theoretically be expected from a representative sample of gay men. Homosexuality cuts across all layers of age, ethnicity, and social status, and theoretically so should any good sampling strategy. Expanding the sources of gay men to include subjects located in gay community centers, youtln progams, gay religious goups, and gay bars may enable future research to assess the generalizability of these findings. 93 Therefore, although the sampling strategy of this study was arguably superior to previous studies of this specialized population, its generalizability is somewhat limited. These results generalize most readily to older, more affluent and educated gay men, whereas the generalizability to younger gay men, or those from lower socio-economic status, is uncertain. An attempt was made to gather demographic data from subjects who were unwilling to complete the entire questionnaire, but no subjects elected to choose this option. Therefore, comparisons between survey respondents and non-respondents was not possible. Respondents’ modal annual income was greater than $60,000, and their modal education level was a masters degree or beyond. This may be the result of selecting subjects from a vendor who has compiled names from companies providing goods and services to gay consumers. Those with higher income and education may be more likely consumers and, therefore, comprise the bulk of the subject pool. Two interesting demogaphic features of the sample are worth noting. One is that the largest group of men in this study rated themselves as single with no sexual partners, n=271. The second most common category (n=217) indicated that they were living with a single, monogamous partrner. Therefore, half of the respondents are currently either not engaging in sexual activity or are in exclusive 94 sexual relationships. This runs counter to the stereotype of gay men as extremely sexually promiscuous. Anotlner important finding regarding relationship status is that only 13% of respondents indicate that they are with one partrner the majority of the time but have the freedom to have sex outside of this relationship. This is in sharp contrast to McWhirter and Mattison’s (1984) survey of 156 gay male couples, that led the authors to report “. . . all couples with a relationship lasting longer than five years have incorporated some provision for outside sexual activity in their relationships “ (p. 252). The present findings indicate that chastity and monogamy are common options, and may reflect a change in the gay community from the early eighties to the current AIDS era. A final descriptor of the study bears special mention. The finding that only 13% of the respondents achieved a MacAndrew score of 24 or higher is in contrast to past studies that estimated the rate of substance abuse in the gay community as high as 30%. One explanation is that poor subject recruitment strategies and lack of precision in measurement of substance dependency in past research renders those findings unreliable. For example, Fifield, De Crescenzo, and Latham’s (1975) findings are cited most often in the literature as the definitive estimate of substance abuse rates in gay men. Their index of substance 95 abuse dependency, however, was based on a number of questionable sources, such as bartenders’ estimates of alcohol dependency. Further, their methods for selecting subjects were quite limited and biased, selecting gay men fiom convenience samples of gay bar patrons and alcohol recovery goups. The current findings, based on a non-bar sample, serves to emphasize the point that subjects gleaned from convenience samples may be skewed. It is possible that this study provides a more reliable methodology for data gathering and utilizes improved measurement of substance abuse than past studies. Another explanation of the lower rate of substance abuse proclivity for the current study, as compared to previous studies, is that the current sample is skewed in the direction of lower abuse proclivity. Given that the current sample is decidedly well educated and of higher SES, and that these attributes are significantly associated with lower substance abuse proclivity (in this sample), it is possible that this group of gay men is skewed in the direction of lower substance abuse proclivity. A final possibility is that gay men have altered their rates of substance usage since the earlier studies conducted during the 1970’s. The AIDS era may have forced the gay community to modify its behavior in a number of arenas. An emphasis on health and choices that enhance pride, although aimed at reducing 96 the spread of a deadly disease, may also be reducing negative behaviors, such as substance usage. E'Efl 1 11 l' 5 Differences were found between study respondents on the measures of gay shame, gay pride, self-esteem, positive affectivity, negative affectivity, and substance abuse proclivity, when compared by their membership in various groups defined by the demographic variables of the study. Levels of gay shame and gay pride differed between categories of subjects on several demogaphic variables including education level, income level, current sexual orientation identification, current comfort with identification, and future sexual orientation identification. Subjects with higher income had less shame than subjects with lower SES. Gay men who reported feeling comfortable with their identities as exclusively or predominately gay had higher levels of pride and lower levels of shame than those who were uncomfortable with their orientations. Along these lines, men who would like to identify as gay in the filture had lower levels of shame and higher levels of pride than those who would like to identify as bisexual or heterosexual in the future. Results for other psychological variables of the study are similar. Gay men with higher levels of income had higher levels of self-esteem, and positive 97 affectivity, than those with lower SES. Gay men with higher levels of education had lower levels of substance abuse proclivity than tlnose with lower levels of education. Subjects who were comfortable with current identification, and those that identified as predominately gay, had higher levels of self-esteem and positive affectivity than those who were uncomfortable with their sexual orientation, or who were not identifying as exclusively or predominately gay. These results indicate that comfort and clarity about career, income, and sexual identity are associated with higher degrees of positive self-evaluation and decreased substance abuse proclivity than those who are less clear about these issues. Subjects who identified that they are not currently having sexual relations appear to have less favorable scores on most measures than those who are currently sexually active, either monogamously or with multiple partners. Men who were not having sex scored lower on gay pride, self-esteem, and positive affectivity than those men who indicated that they were sexually active. Additionally, the celibate members had higher levels of shame and negative affectivity than the other men. Sexual relationship status was not associated with substance abuse proclivity; that is, men in the categories of sexual monogamy, celibacy, or sexual non-exclusivity were not significantly different in their proclivity to abuse substances. These findings have several plausible explanations. 98 First, it may be that gay men who are sexual have their needs for sex and companionship met. This may lead to positive evaluations of the self, such as higher gay pride, self-esteem, and positive affectivity, concurrently lowering gay shame and negative affectivity. An alternative explanation is that gay men witln higher gay shame, higher negative affectivity, lower gay pride, and lower levels of self-esteem may not be as successful in finding sexual partners, and are consequently not sexually active with others. Taken together, these differences among all demographic variables may indicate that gay men who recognize and are more accepting of their homosexuality have more gay pride, enjoy more self-esteem, feel more positive affect, and are more likely to be sexually active than those who do not clearly identify as gay, and who are not comfortable with their sexual identities. Respondents who were sexually active, either with one or multiple partners, did not differ from non-sexual gay men on substance abuse proclivity. Ii 11 .4 ISI E°ISI The reliability estimates for the Gay Identity-based Shame and Pride Scales, developed for this study, were quite high. Of all the existent psychology measures utilized in the study, the alpha coefficients for the newly developed gay shame and gay pride scales, or = .94 and or =.93, respectively, were the highest of 99 all the instruments. The gay shame scale is a modified version of its parent Intemalized Shame Scale. This original scale also possess a high degree of interrnal consistency. Therefore, the gay shame scale performs comparably to the original ISS. The gay pride scale, however, was derived from theory alone and modified versions of the six positively worded items from the ISS. The internal consistency of this measure is impressive given it is an initial attempt to measure a theory-based construct. The validity of both the Gay Identity-based Shame and Pride Scales, as provided by convergent validity with the other measures of the study, is also substantial. The individual items of these two scales both appeared to be indicative of the latent variables they were designed to predict This was demonstrated by the excellent goodness of fit indices provided by the confirmatory factor analysis. In addition, the correlations of these scales with each other, and to the other measures of the study, supported the hypothesis that the gay shame and gay pride scales were independent constructs and represented independent latent variables. Both scales performed in the directions predicted by tlneory. In sum, it appears that this study was successful in developing measures of gay identity-based shame and gay identity-based pride. This initial study has 100 provided encouraging data on their reliability and validity, and supports the possibility of their use clinically (see Implications for Counseling below). 1181' [351 CiE'l How do Gay Identity-based Shame and Gay Identity-based Pride relate to one another? Many shame theorists would have predicted that shame and pride are opposite ends of a single continuum. This does not, however, appear to be the case with Gay Identity-based Shame and Pride. The two significantly correlate, as predicted, in a negative direction, x = -.53, p < .001. Although this indicates that the two share variance, the strength of this association is not large enough to support a single continuum hypothesis. As predicted, subjects may have varying and independent levels of shame and pride about their homosexual identity. An interpretation of this finding is that these variables parallel the relation among the negative and positive affectivity scales on the PANAS. Subjects can possess various combinations of gay shame and gay pride (e.g., high levels of both shame and pride, low levels of both, high levels on one and low on the other). This independence allows for varying levels of botln shame and pride, and is supported by the present data. 101 't‘: no r 0‘15-0- 'er ltll'nl' ’0" 0011‘ 100' A major focus of this study was the question, “What is the difference between gay identity-based shame, pride, and self-esteem?” Unlike Chang (1988), who found a very high negative correlation between shame and self- esteem, this study utilizing identity shame, found only a moderate negative correlation (r = -.55, p < .001). As with shame and pride, this correlation suggests that tlnese two constructs are not the same. Although they share some commonality, both appear to be measuring some unique aspects of the individual. These findings suggest that it would be possible to have a rather positive global self-esteem, but still experience negative feelings of gay identity shame. Additionally, the same is true with gay identity pride and self-esteem. One could have a high global self-esteem, yet still have little gay pride. Perhaps global shame and self-esteem are truly differing aspects of the same “elephant” as was described earlier. One may be more cognitive and the other the affective experience of the cognition, as has been argued by Pelham and Swann (1989). But there appears to be something unique about gay identity- based shame and pride that set them apart from global feelings of self-worth. An explanation of the difference between global shame with respect to self-esteem is 102 provided by shame tlneory. Theory would predict that shame associated with identity would have significantly greater negative effects than global shame associated with a circumscribed need, drive, or affect. The pervasive effects of shame associated with identity may have given gay identity-based shame discriminant validity from self-esteem that global shame, associated with needs, drives and affects, does not have. This may have been the case with the subjects of tlnis study, because their identity shame was related to self-esteem, but distinct fiom it. The shame surrounding homosexuality is so “unpalatable” by societal norms that it may be split-off from other global feelings and judgnents about the self. Gay men often talk about hiding their identities, at times, from others and themselves. This “passing” as it is known by gays may first start with the self. Whereas the Rosenberg scale taps the general self-esteem, the hidden aspects of being gay may be removed, at times, from these considerations. In sum, the question of how gay identity-based shame and self-esteem are related is clearly articulated, regardless of the theoretical acumen, by the data Items about gay shame and self-esteem load most highly on their own scales, suggesting that they represent differing aspects of personality. 103 Another important issue was the relationship of gay identity-based shame and pride to the PANAS. Because these scales all purport to measure affects, it was possible that PANAS would subsume the Gay Identity-based Shame and Pride Scales. However, consistent with the relationship of gay shame and pride to other variables, PANAS is only moderately correlated with the Gay Identity-based Shame (positive affectivity r = - .32, p < .000; negative affectivity 1: = :53, p < .000) and Pride Scales (positive affectivity [ = .47, p < .000; negative affectivity [ = - .30, p < .000). And these correlations are in the predicted direction. As would seem appropriate, in the factor analysis, the PANAS item “Ashamed” appeared on the same factor as the Gay Identity-based Shame items. The addition of PANAS to the study was valuable, as the relation of the PANAS scales to one another served as a model for explaining the relationship between the Gay Identity-based Shame and Pride scales. In sum, it appears that the Gay Identity-based Shame and Pride Scales, while moderately correlated with the other psychological variables of the study, represent and measure unique constructs. This information provides support of their convergent validity. 104 r- 0'15'.t"1' t..tt‘..10’rr- err-10s. '3! 'st.' to 5 Results of the regression analysis were interesting for a number of reasons. In this study, the MAC significantly covaried with age. Past researchers (Appledorf & Hunley, 1975; Duckworth, 1983; MacAndrew, 1965) found no significant correlation between the MAC and age. Why this is so is not clear. Of prime interest in this regression analysis, however, was the question, “Does higher Gay Identity-based Shame positively covary with proclivity to abuse substances?” The answer from this study is yes: Gay Identity-based Shame positively covaries with the MAC. As was theoretically hypothesized, higher levels of internalized gay shame predicted higher substance abuse proclivity scores, afler the effects of the other study variables were taken into account. This relationship is opposite to the correlation analysis that found a slight negative relationship between shame and MAC scores. The “flipping” of the direction of a correlation in regession analyses may be due to the influence of a “suppressor” variable in the regression equation. Conger (1974) defined a suppressor variable as, “. . . a variable which increases the predictive validity of another variable by its inclusion in a regession equation (p36)”. Therefore, the change in direction and significance of the relationship between shame and the MAC could be due to a suppressor variable. By removing irrelevant variance 105 from the equation, theoretically, a clearer picture of the nature of the relationship between the other independent variables with the dependent variable is possible (Cohen & Cohen, 1975). The use of regession analysis may provide a more “real world view” of the nature of the relationship between shame and substance abuse proclivity because, under natural conditions, other variables may suppress, augnnent, and otherwise influence the strength and direction of the relationship between gay identity-based shame and substance abuse proclivity. Unexpectedly, a positive relation of Gay Identity-based Pride to MAC scores was also uncovered. One possible explanation for this unlikely correlation between pride and substance abuse proclivity comes from a detailed investigation of the MAC. F inney et. al. (1971) found that in addition to substance abuse proclivity, higher scorers on the MAC “...seemed to be bold, uninhibited, self- confident, sociable people who mix well with others. “ (p. 1058). This multidimensional composition of the MacAndrew Scale has also been endorsed by MacAndrew (1981) and Duckworth (1986), and is further supported by the low to moderate internal consistency of the MacAndrew Scale for this sample. The alpha value for the MAC was .52, the lowest alpha for any measure of the study. In terms of this sample, the reliability of the MAC is not at the same level of dependability as the rest of the variables measured. Therefore, some items of the 106 Gay Identity-based Pride Scale may also be tapping the same latent factor as the sociable and self-confident items of the MAC scale, resulting in a positive correlation. Likewise, the positive affectivity scale of the PANAS correlated positively with the MAC, as many of the items of the positive affectivity scale also reflect self-confidence and sociability. Another problem with the MAC scale was that numerous subjects did not complete the entire scale. This was particularly a problem with items 2, 4, 7, 38, 46, and 47 that deal with religious topics and have a decidedly Judeo-Christian emphasis. For example, item 46, “Christ performed miracles such as changing water into wine”, is typical of these items. Several subjects wrote, “I don’t know, I wasn’t there.” next to this item and many did not respond with a ‘T’ or ‘F’. SPSS was instructed to treat each occurrence of no response as missing values. In sum, the MacAndrew Substance Abuse Proclivity Scale may not have been the best choice of measures of substance abuse proclivity. It has only moderate internal consistency, and has a tendency to co-vary with positive affect, extroversion, or social facility. Additional measures of substance abuse proclivity, and documentation of actual substance usage, may have provided more precise measurements of substance abuse proclivity. In addition, multiple measures of substance abuse proclivity may have provided a referent by which the 107 accuracy of the present MacAndrew data could be compared. Future research would do well to explore the predictive validity of the shame and pride scales in relation to more adequate indicators of psychosocial functioning. Il"frEl' The results of this study, altlnough modest, may have some impact for clinicians working with gay men. The first issue is the readiness for clinical use of the gay identity-based shame and pride scales. This study provides preliminary reliability and validity data to support the use of the Gay Identity-based Shame and Pride Scales within a psychotherapy setting. Both overall scores and individual item responses could be utilized by clinicians to help clients understand and conquer their individual, specific areas of gay shame and also develop pride as gay men. Another issue with clinical implications is the relationship between gay identity-based shame and pride. The finding that these constructs are related, but distinct constructs should highlight for therapists the importance of assessing levels of each in their gay clients. Therapists dealing with very “out” clients may inadvertently overlook effects of shame if they detect areas of developed gay pride. Because these constructs may act somewhat independently from one another, the insidious effects of shame may still be present. 108 Consistent with shame theory, gay identity-based shame did predict substance abuse proclivity. Therefore, clinicians should carefully assess substance usage in their gay male clients. A thorough substance use history, and assessment of current types, amounts, and frequencies of all alcohol and drugs used by gay clients is warranted. Therapists who work with a number of gay clients should seek additional alcohol and drug assessment training to perform the tasks adequately. Although this study only addressed substance abuse proclivity, shame theory predicts other negative effects of shame. Because a significant relationship exists between shame and substance abuse, therapists might assess, and be cognizant of, other addictive forms of psychological pathology (e.g., eating disorders, over-spending, co-dependent relationships) that may coexist along with shame in their gay clients. Additionally, therapists might explore how gay shame derives from clients’ internalization of cultural messages about homosexuality. Therapists could utilize the Gay Identity-based Shame-Pride Scales clinically, through examination of individual shame items endorsed. Counselors could then encourage clients to directly examine the content of each item to address the internalized gay-shame. Clinicians could then help clients test out the validity of their internalized fears as they work at empowering themselves with gay pride. In 109 their process of facing their gay identity-based shame and developing a proud gay identity, many clients will have to face the sense of shame conveyed by item 16: “Because I’m gay, I feel as if I am somehow defective as a person, like there is something basically wrong with me.” Helping a gay man heal from the effects of this pain seems an essential goal of psychotherapy. Challenging gay clients further to develop gay-pride is a logical next step. Finally, the possible confounding in this study between the high degee of social facility tapped inadvertently by the MAC and gay pride-scale may serve as an additional reminder of the association of the gay community and gay culture with bars, alcohol related socialization, and gay-identification. The gay bar still plays a major social role in the gay male community. Gay men who have recently come out may look to men in these arenas to define social norms, codes of behavior, and ways of identifying themselves. Therapists need to educate tlnemselves about alternative social resources in their communities for their gay male clients. Helping clients to identify with healthy gay men in their communities can work to lessen shame, develop pride, and disconnect gay identity from addictive, closeted behaviors. 110 11"EEEI As with all research projects, the process of bringing this study to fruition has raised many more questions than it has answered. There are multiple directions that future research could address. These fall into two main areas: further empirical studies for the refinement of the shame construct, and further research directed at the gay community with linkages to shame-pride. Several issues for each area are detailed below. Future research on the reliability and validity of the Gay Identity-based Shame and Pride Scales is warranted, particularly with expanded samples of the gay community. Further studies that assess the reliability of these scales could help to confirm their high degree of internal consistency. While the current study provides some information about the convergent validity of the gay shame and pride measures, additional research could address both convergent and divergent validity issues. Also, future research on the constructs of gay identity-based shame and pride might employ either a different measure of substance abuse proclivity or another index of psychopathology as a criterion in assessing predictive validity. Altlnough this study was able to trace a relationship of shame and pride to substance abuse proclivity, the content of the MacAndrew scale may have been lll confounded with the pride construct. An alternative to the MAC, such as direct documentation of alcohol use, might shed further light on these relationships. Alternatively, use of the Beck Depression Inventory, other indices of depression, or other forms of psychological difficulties relevant to shame theory would be helpful. Another issue to be addressed in subsequent research would be to extend the focus of the present study into a series of studies that address the issues of shame and pride in various segnents of the gay population. This series could expand the methods of obtaining subjects to gay community centers, youth progams, health departments, and traditional gay bars. This diversification of the sample would make it more representative and generalizable to the overall population of gay men. The focus of these studies could be substantially influenced by characteristics of heightened importance to each of these segnents of the gay male population. Assessing differences of rural versus urban gay men, younger versus older, poorer versus richer would be crucial. Further research on gay shame and pride has multiple applications for the gay community. Future study of shame and pride in gay men could help address health issues, including prevention of AIDS. For instance, would lower levels of shame and higher levels of pride facilitate adherence to safer sex practices? 112 Would levels of shame and pride support or undermine the efforts of HIV positive gay men to remain healthy or comply with treatment? A frightening new trend is developing for gay men to stop practicing safer sex, placing them once again at a very high risk for all sexually transmitted diseases, including AIDS (Rogers, 1994; Yam, 1995). What roles might gay shame and gay pride be playing? The use of the newly developed gay shame and pride measures may be useful to those researching ways of preventing the spread of, and treating, this disease. An additional area for further research might address gay identity developmental stages in relation to gay identity shame and pride. An iderntity developmental approach may provide clearer data about the relationship between shame and pride in gay men as they gow along developmental lines toward identity integration and actualization. This line of study might provide clinicians with further insights about the coming out process and filrther empowerment of gay male clients. Finally, the research paradign utilized in this study might be expanded to further examine shame and pride as constructs. For instance, might identity-based shame and pride be relevant constructs in lesbian women, bisexual-sexual men and women, or other under-represented/oppressed goups? Could shame about one’s etlnnicity and oppression predict gang violence, school drop-out rates, or 113 teenage pregnancy? The constructs of identity-based shame and pride may have numerous uses in investigations of other oppressed goups. Conclusion E . . l E' l' A major accomplishment of this study was to construct measures of gay identity-based shame and pride that demonstrated adequate reliability and validity. These scales were then utilized to examine the relationships of gay identity shame and pride to other variables in psychology that had either relevance for the gay community or were utilized to provide convergent reliability for these newly developed scales. Factor analytic techniques provided support for the validity of the constructs of gay shame and pride as independent from other latent variables. These new scales also appear to relate to other measures in directions predicted by theory. The study also collected data regarding demogaphic variables for this sample of gay men, which provided interesting information about these subjects. Men who were more highly educated and had higher income levels enjoyed significantly higher levels of gay pride, self-esteem, and positive affects and lower levels of gay shame and negative affectivity. Subjects who identified most strongly as gay, those who were comfortable with this label, and those who would 114 like to continue to use this label in the filture, also enjoyed significantly higher levels of gay pride, self-esteem, and positive affects and lower levels of gay shame and negative affectivity. Additionally, subjects who were sexually active had more favorable levels of self-esteem, gay pride, and positive affectivity, and lower levels of gay shame and negative affectivity, than those who were not currently sexually active. Preliminary findings suggest that the construct of gay identity-based shame has utility for predicting one measure of distress relevant to the gay community: substance abuse dependency. The construct of gay identity-based pride also significantly predicted abuse proclivity, but this finding is not in a direction predicted. This, however, may be an artifact of the substance abuse proclivity measure, rather than a defect in the pride measure. Further research of the utility of these scales is warranted. I I. . E I E. I I The reliability and validity of the Gay Identity-based Shame and Pride Scales, while preliminary, provide support for their use in future reset and clinical realms. Future research utilizing these scales could expand the methods of subject recruitment to diversify future sample characteristics of gay men to increase the generalizability of further findings. Other research recommendations would be to 115 assess the reliability and validity of these scales with other samples. The Gay Identity-based Shame and Pride Scales hold promise for studying other issues of import for the gay community. Several examples of such issues are exploring possible ties between gay shame and pride to addictive coping mechanisms. These scales may also be useful in furthering an understanding of why gay men are stopping safer sex practices. Clinically, psychotherapists might begin to use the shame and pride scales in their practice with gay men. Because shame and pride can be independent, therapists should be aware that clients presenting with initial high levels of gay pride may still have unresolved gay shame. Use of these scales could help identify specific areas of shame that remain problematic for their gay clients. Finally, althougth the results of this study are correlational and do not imply causality, therapists who work with gay male clients should be aware of tlne links between gay shame and a proclivity to abuse substances and should assess substance usage with these clients. TABLES Table l. 116 2.. 1m 1. Interest-Excitement: Eyebrows down, track, look, listen 2. Enjoyment-Joy: Smile, lips widened up and out 3. Surprise-Startle: Eyebrows up, eye blink MW 4. Distress-Anguish: Cry, arched eyebrows, mouth down, tears, rhythmic sobbing 5. Fear-Terror: Eyes frozen open, pale, cold, sweaty, facial trembling, with hair erect 6. Anger-Rage: Frown, clenched jaw, red face 7. Shame-Humiliation: Eyes down, head down 8. Dissmell: Upper lip raised 9. Disgust: Lower lip lowered and protruded Note. Taken from Kaufman (1989, p.12). Table 2. 117 WWW Year Investigators Number of Rate of Alcohol Dependency Subjects Alcohol Criteria Dependency 1973 Saghir, et. al 89 males 30 % 2 questions on substance usage 1974 Weinberg, et. 2497 males 29 % 1 question ‘ al. 1975 Fifield, et al. 483 gay bar 31 % Gay bar patrons’ and patrons, bartenders’ estimates bartenders & regarding patrons gay recovering alcoholics 1978 Lohrenz, et. al. 145 males 29 % MAST “ ‘ Weinberg, et. al. (1974) asked, “Do you ever drink more than you should?” " MAST = Michigan Alcohol Screening Test 118 Table 3. on '.'..":’\."I 1"0‘I'. si‘nl'éks .--r~ out '00.! “10011.1‘ ..1 .0 L'er 011 1,0. “I. to . MW Correlations with Measure and PANAS n PANAS NA PANAS PA time instructions HSCL Past few weeks 398 .74 -.19 Today 53 .65 -29 BDI Past few days 880 .56 -.35 Past few weeks 208 .58 -.36 A State Past few weeks 203 .51 -.35 N915. Taken from Watson, Clark & Tellegen (1988, p. 1068). 119 Table 4 WWW Status of Questionnaire n % Completed and returned 200 40 % Could not be delivered to subject (e.g., return to sender, no 4 1 % forwarding address, no such address) Subject deceased 7 1 % Subject indicated not interested 5 1 o/ in participation o Subject states he is not gay 6 1 % Total Questionnaires Accounted 222 44 % For. No response from subject or any 278 56 % other agent (e.g., relative, postal service) 120 o. _ an. _n. 0.. on. o.— ovm svm ova on. o.— vvm ~v. an. an. o.— 8'63. 33.331883 0. — av On av an on. aw an On 0v. av ov. .. an. _v. a. nv. a. No .v. _v. «N. on. an. on. an. nv. nv. vv. hv 3 an. ON. 0v. «m. 0v. nv. 5. no a Iv av nv. on. on. av nv. 9. 3 av. av no. 0v an. an. av. vn. on 9. -v on. ov on vn. _v. mm. on. vN on. on. on. «v On. an v" on. an 0N. an. on. «v. an «v pm av NM. nv 3 ON. av «n. «v. 3. on. uv. nv. nv 0v 0N. —v no. IN. nv. an. on. no. so. nv On a. on. 3. nu av. pm. No. 3. av. on. :1 5. on. 3 ma an. an. vn. an. nv. an 8. a. vN. o _ on. «v. «v. v0. av. 5. vv 3 «n. 0v Q— «N. «n. ma. an. «v. an vn. 2 an o — Nv 0v. 0v 0v. Mn vn. on FM o; no. so. on. nn. 3. pm. ~v 0 ~ 3. on. In on an. nv A: no. «v 6v en. an o — a. vn. an. vv 0 _ an on. _v o _ 2. on. o — vv. 9.— ona nun a; a; 2m 2m 2m :m an a vm an gag—835383.335 n 039—. .v. ova hvu ova vvu Nvu .eu can can fin am can «an 2m 2m 2» man go Zn on no vu no .a 121 0.. nu. un. 0.. nn. 0.. 0mm nvm nvm nu. 0.. nvm .vm nu. nn. an. 0.. onm .v. 0. . «n» 8513113. V r. 3.8 nnn nnm unn 0nm 0.. on. 0.. nn. bu. u. 2. 0.. cu. nu. n—. vu. on. nu. nn. on a. .v nv. nu. 0 .. nn. hv. a. on. .v. on. on .v. «v on. vu. h .. 0n. .n. nu. .. .. .u. n .. .n. «n. a. .n. 8. u. nn. v.. on. nv. .u. nv. nv .n. en. 0v. no. un. on. On. vu. on. nu. 3. sn nu. vu. nv. . .. n .. nu. .n. nu. an. nu. un. vv vv 0v. nv. 0n. nu. nu. vv. nu. nn. 0.. vn. vn nu. nu. 0n. 2.. nu. uu. no. vu. No. uv. 0n. .v .n. an. un. a . . vu. vn. uv. “N. .n. vu. nu. on. on. 0v. 0n. n .. an. nv. vu. 9. av. vu. nv. a. un. vn. 0n. 0H. 0.. nu. nv. uu. nn. vv. 8. 0n. uv. nn. 0n. vu. nv. 0v. nn. 0n. 0. . n n. .0. nu. a. .n. .v. vu. 8. 0u. 8. 0 .. 0. . nu. vn. nu. nu. an. .0. un. on. o .. .n. 0. . R. nn. vv nn. vu. 8. vu. fin. n .. 0 . vv. 9 0n. nu. nv. nn. .n. 3. 0 . vn. nv. vu. 5n. nn. 5. nn. 0.. vn. nu. on. nv. a. 0v. 0. . un. 0v. vv. nv. 3. 0. . n .. uu. vn. n .. 0. . nn. uu. uv. 0. . nu. 6n. 0. . h .. 0.. mu m nu a nu m .u n cu m u. n v.n n. m 0. m an on an 333%: 0 035... 0.. un 0mm 'n nvm nvu .vm .nnn unn nnn nnn unn an Na ”Na nun nun .un .an 2m v.» 2a . 0.u on no nm um 122 Table 7 f3 ...‘-.'1-. H’ H' .Ul~ .‘II \Il-l .0. l. C I. . Gay Shame Item Scale Item Mean Item Communalit Item Number Correlation Variance y 1 .64 1.85 .95 .57 3 .56 2.34 .94 .39 4 .44 2.42 .88 .35 7 .58 1.23 .56 .43 9 .58 2.14 .96 .43 1 1 .75 1.64 .82 .63 12 .74 1.3 1 .68 .61 15 .72 1.54 .81 .60 16 .72 1.22 .56 .59 18 .56 1.36 .84 .34 19 .48 1.81 .99 .32 22 .67 1.43 .85 .49 26 .47 1.86 .94 .23 29 .69 1.22 .65 .63 31 .71 1.37 .77 .49 34 .71 1.20 .63 .59 36 .62 1.25 .64 .47 38 .48 1.49 1.01 .25 40 .61 1.25 .67 .54 42 .72 1.53 .93 .64 44 .75 1.40 .85 .69 46 .67 1.76 1.15 .55 47 .74 1.47 .88 .65 49 .72 1.45 .80 .60 123 Table 8 r- O‘oirI\-t.,'r ’ot' ,.' 511- ,. ‘ ,su‘ .101 In U‘.-t «l-t ‘ C I. . Gay Shame Item Scale Item Mean Item Communalit Item Number Correlation Variance y 2 .47 3.43 1.44 .56 5 .45 4.16 .94 .43 6 .63 3.26 1.19 .61 8 .50 4.35 .87 .41 10 .42 3.64 1.02 .38 13 .47 3.85 1.43 .68 14 .70 4.15 1.06 .63 17 .58 4.47 .73 .55 20 .53 4.62 .63 .61 21 .60 4.15 .90 .61 23 .57 3.44 1.22 .53 25 .42 3.33 1.04 .34 27 .60 3.34 1.22 .56 28 .64 4.26 .86 .60 30 .61 3.73 1.17 .52 32 .53 4.67 .64 .52 33 .70 3.55 1.05 .64 35 .62 3.25 1.42 .76 37 .70 3.84 1.11 .60 39 .53 3.81 1.10 .54 41 .71 2.72 1.32 59 43 .45 4.14 .91 .49 45 .71 3.70 1.16 .60 48 .64 3.22 1.08 .53 50 .40 3.09 1.04 .36 124 service) Table 9 WWW Status of Questionnaire n % Completed and returned 783 78 % Could not be delivered to subject (e.g., return to sender, no 43 4 % forwardrng address, no such address) Subject deceased 14 1 % Subject states “not interested” 11 1 % Subject states he is not gay 18 2 % Total Questionnaires Accounted 369 37 % For. No response from subject or any 131 13 % other agent (e.g., relative, postal Table 10 125 . 0 Standard Vanable n /0 Mean Deviation Age: 973 47.05 1 1.22 Ethnic 1d.: 973 Black 10 1 % Himanic 16 2 % White 922 95 % Asian 8 < 1 % Native American 8 < 1 % Other 8 < 1 % Education Level: 973 Less than High school 9 1 % High School 42 4 % Some College 155 16 % Associates Degree 45 5 % Bachelor’s Degree 309 32 % Masters Degree & beyond 407 42 % Employed: 973 Yes 799 82 % No 21 2 % Laid OK 12 1 % Never Worked 2 < 1% Retired 137 18 % Income: 973 < $5,000 5 1 % $5,000-Sl4,999 44 5 % 315,000-324,999 91 9 % 325,000-339,999 234 24 % 840,000-359,999 256 26 % $60,000 or more 336 35 % Relationship Status: 973 Single, no sex 271 28 % Single, 1 partner 58 6 % Single, multiple partners 188 19 % Coupled, 1 partner, live-in 217 22 % Coupled, > lpartner, live in 109 11 % Coupled, 1 partner, live separate 55 6 % Coupled, >1 partner, live update 30 3 ”/0 Other 39 4 % Note: n = 973 to correct for social desirability response bis 126 0H ‘l Table 11 (a. '1' ’1 -'1.' . 'fl'r'll ”- ' '5 ' . 11.1 “I? '1. 1 ° 111-. H"! .Jlsll '3 "l 1U '1 D . . Variable n % Current Sexual Orientation: 973 Exclusively homosexual 854 88 % Predominately homosexual 88 9 % Bisexual 18 2 % Predominantly heterosexual 0 0 % Exclusively heterosexual 1 < 1 % Unsure 8 <1 % Future Sexual Orientation: 973 Exclusively homosexual 819 84 % Predominately homosexual 90 9 % Bisexual 31 3 % Predominantly heterosexual 7 1 % Exclusively heterosexual 2 < 1 % Unsure 16 2 % Comfort w/ Sexual Orientation: 973 Very Comfortable 699 72 % Mostly Comfortable 175 18 % Comfortable 66 7 % Mostly Uncomfortable 23 2 % Very Uncomfortable 6 1 % Note: 11 = 973 corrected for social desirability response bias in It 127 Table 12 ll lSIID"flEll'lll ElSl . Standard Variable It Mean Deviation Marlowe Crown Social 981 19.17 3.26 Desirability Scale: Form C Rosenberg Self-Esteem Scale 981 33.83 5.55 MacAndrew Substance 981 18.94 4.09 Abuse Proclivity Scale PANAS: Positive Affectivity 981 36.74 6.36 Negative Affectivity 981 18.29 6.62 Gay Identity-based Scales: Shame 981 36.66 12.94 Pride 981 94.52 17.63 Note: Scoring information for each scale above: ‘Mar'lowe-Crowne Form C Scale Mrmm' ° um: 13 Scale Maxrm' um: 26 ‘Rosenberg Self-Esteem Scale Scale Mrnrm' ' um: 10 Scale Maxrm' um: 40 ‘MAC - Substance Abuse Scale Mrnrm' ' um: O Scale Maxrm' um: 50 ‘Positive Afl‘ectivity Scale Mrmm' ' um: 10 Scale Maxrm' um: 50 ‘Negative Affectivity Scale Mrmm' ' um: 10 Scale Maxrm' um: 50 ‘Gay Shame Scale Mrmm' ' um: 25 Scale Maxrm' um: 125 ‘Gay Pride Scale Mrnrm' ' um: 25 Scale Maxrm' um: 125 ‘Higher scores on each of these scales represent more of that attribute. 128 Table 13 .ll' «'00 Ho". 6 0l‘ 0 H ‘1 O 0 o . Marlowe- Self- Positive Negative Mac- Shame Pride Crowne Esteem AflwfiViW Micah/“Y Andrew Shame .94 Pride -.53 .93 p=.000 Marlowe -.17 .14 .77 Crowne p-.000 p=.000 Self -.53 .50 .32 .90 Esteem p==.000 p==.000 p=.000 Positive -.32 .47 .24 .53 .88 Affectivity p=.000 p=.000 p=.000 p=.000 Negative .53 -.30 -.32 -.50 -.25 .90 Afl‘ectivity p=.000 p=.000 p=.000 p=.000 p=.000 Mac- -.03 .21 .10 .15 .27 -.04 .52 Andrew p=.314 p-.000 p=.002 p=.000 p-.000 p-.176 Nam; N = 983 Diagonal values indicate Cronbach alpha internal consistency reliability estimates for each measure. 129 Table 14 - ‘1 o -."H human r- . on BUOo‘aou; horn-u Xariahlcs Dependent Variables Shame Pride MAC Demographics: M 512 1 M 512 1 M SD 1 Ethnicity - Anglo 36.67 12.9 94.61 17.4 18.88 4.0 Non-Anglo 36.60 12.8 .04 93.10 21.3 .61 19.86 5.3 - 1.69 Education Level Less than College 37.40 14.5 92.52 17.4 20.43 3.9 Bachelor's and up 36.40 12.4 1.06 95.33 17.6 - 2.2 18.40 4.1 6.97 " Employed - Yes 36.51 12.7 94.76 17.5 18.78 3.9 No 37.36 13.9 -.78 93.80 18.1 .65 19.72 4.7 - 2.70 Income - Low 39.74 14.6 91.40 19.3 19.76 4.4 High 36.10 12.4 3.15 " 95.12 17.1 - 2.34 18.80 4.0 2.61 Current Sexual Orientation - Predom. Gay 36.52 12.5 95.19 16.8 18.91 4.0 Bisexual 42.14 23.8 - 2.27 ' 70.61 26.3 7.48 " 19.78 6.1 - 1.12 Future Sexual Orientation - Predom. Gay 36.06 11.79 95.79 16.3 18.89 3.9 Bisexual 46.23 22.91 - 5.86 " 74.12 23.2 9.44 " 19.82 5.4 - 1.67 Comfort Level Being Gay - Comfortable 34.50 9.65 97.10 15.56 19.00 4.0 Uncomfortable 57.02 20.16 - 18.79‘ 70.38 17.50 15.71 " 18.39 4.8 1.39 sn<.oor 130 Table 15 -551 0 -.'°.O .A'HO'O’a'l oI-. 5% ‘ H I 6L; .1H0.'IO'| Xan'ahlcs Dependent Variables Self-Esteem Positive Afl‘ectivity Negative Affectivity Demographics: M 512 1 M SD 1 M 512 1 Ethnicity - Anglo 33.87 5.4 36.71 6.3 18.29 6.6 Non-Anglo 32.92 7.3 1.2 37.38 7.5 -.75 18.21 6.9 .08 Education Level Less than College 33.07 6.1 35.81 6.7 18.23 7.2 Bachelor's and up 34.10 5.3 - 2.56 37.09 6.1 - 2.81 18.31 6.4 -.17 Employed - Yes 33.87 5.4 36.75 6.3 18.41 6.6 No 33.60 6.2 .60 36.64 6.7 .21 17.90 6.6 .87 Income - Low 31.84 6.6 34.51 7.4 18.91 7.2 High 34.18 5.3 - 4.71 ' 37.12 6.1 - 4.59 " 18.14 6.5 1.28 Current Sexual Orientation - Pmdom. Gay 33.90 5.4 36.82 6.2 18.26 6.5 Bisexual 30.61 8.7 3.11 ' 33.32 10.4 2.88 " 19.43 9.3 -.92 Future Sexual Orientation - Predom. Gay 34.03 5.3 36.89 6.1 18.10 6.4 Bisexual 30.16 7.75 5.18 " 33.91 8.9 3.45 ' 21.35 9.4 - 3.63 " Comfort Level Being Gay - Comfortable 34.38 5.17 37.20 6.1 17.70 6.2 Uncomfortable 28.53 6.2 10.27 " 32.27 7.0 - 8.83 ' 23.78 8.1 - 8.83 ' rn<.oor 131 Table 16 at. a 0 .A‘tto'r-H -I-0' '.1.-:'o|0110 HI .1011; 1H 0 WWW Dependent Sum of Mean . Source Vanable: Squares df Squares E 11 MAC Main Efl'ect 20.76 2 10.34 .618 .539 Error 16359.05 974 16.80 Total 16379.81 976 Shame Main Efl'ect 7596.14 2 3798.07 23.70 .000 Error 156074.97 974 160.24 Total 163671.11 976 Pride Main Efl'ect 18723.25 2 9361.63 32.00 .000 Error 284908.28 974 292.51 Total 30363 1.53 976 Self-Esteem Main Efi'ect 586.97 2 293.49 9.67 .000 Error 29553.75 974 30.34 Total 30140.72 976 Positive Afl'ectivity Main Efl'ect 1239.80 2 619.90 15.74 .000 Error 3854.04 974 39.38 Total 39593.83 976 Negative Afl'ectivity Main Efi‘ect 352.26 2 176.13 4.04 .021 Error 425 12.02 974 43.65 Total 42864.28 976 Monogamous Non-Exclusive Sex Not Sexual quendent: M SD M SD M SD MAC 19.04. 3.97 19.02 . 4.06 18.70 . 4.25 Shame 33.79 , 10.40 35.55 , 11.46 41.01., " 15.90 Pride 97.67 . 17.26 97.00 . 15.82 87.50 b" 19.04 Self-Esteem 34.45 _ 5.74 34.23 , 5.38 32.59 b" 5.55 Positive ” Afl‘ . ‘ty 37.39 , 7.07 37.45 . 5.62 34.92 b 6.68 Negative . Afl‘ . 'ty 17.40 , 6.18 18.24 , 6.56 19.10 b 6.99 ‘n<.05 "n<.000 Note: Means having difl‘erent subscripts are signifimntly different at p_< .05 132 Table 17 {s._-o o _o.or' o r- '1 \-r a 111 10.00'3111 «'10 Item Scale Item w 3 - . -' h Number is me Item Wordrng “‘ GIBSPS 29 Gay Shame I would like to shrink away when I feel like I am a .80 .71 mistake as a gay man. GIBSPS 4" Gay Shame Being gay, I feel like there is something missing inside .74 .68 of me. GIBSPS 34 Gay Shame At times I feel like 1 will break into a thousand pieces. .75 .63 because I am gay. GIBSPS 44 Gay Shame Being gay, I feel empty and unfulfilled. .72 .73 01389840 Gay Shame At times being gay makes me feel so exposed that I .72 .58 wish the earth would open up and swallow rne. GIBSPS 16 Gay Shame Because I am gay, I feel as if I am somehow defective as .74 .64 a person, like there is something basically wrong with me. GIBSPS 42 Gay Shame Being gay has left a painful hole within me that 1 have .67 .70 not been able to fill. GIBSPS 36 Gay Shame I feel as if I have lost control over my body, my life and .69 .57 my feelings because I am gay. GIBSPS 49 Gay Shame Compared to non-gay men, I feel lesser. .68 .63 GIBSPS 31 Gay Shame I replay painful events about my being gay over and .66 .56 over in my mind until 1 am overwhelmed. GIBSPS 12 Gay Shame Because of my sexual orientation, 1 see myself as being .69 .66 very small and insignificant. GIBSPS 22 Gay Shame 1 see myself striving to be the perfect man only to .65 .55 continually fall short because I am gay. GIBSPS 15 Gay Shame As a gay man. I feel intensely inadequate and full of self .68 .63 doubt. GIBSPS 07 Gay Shame I scold myself and put myself down for being gay. .64 .51 GIBSPS 46 Gay Shame As a gay man, my loneliness is more like emptiness. .59 .66 GIBSPS 11 Gay Shame Compared to non-gays, I feel like I somehow never .58 .68 measure up. GIBSPS 18 Gay Shame When I compare myself to non-gays I am just not as .58 .43 important. GISPS 01 Gay Shame Because I’m gay, I feel like I am never quite good .56 .62 enough. GIBSPS 38 Gay Shame Because of my being gay, sometimes I feel no bigger .48 .31 than a pea. GIBSPS 19 Gay Shame I have an overpowering dread that my being gay will be .44 .57 revealed in front of others. GIIPS 09 Gay Shame I feel insecure about others opinions of my gayness. .41 .53 omsrs 28 Gay Pride On the whole, I am pleased and satisfied with myself as -.42 .66 Eigenvalue = 16.57 a ay man. % ofvariance = 33.1 Table 18 133 Item - Scale Item . -rng It” Number is From Item Wording GIBSPS 27 Gay Pride My gayness is one of the finest aspects of myself. .78 .67 GIBSPS 33 Gay Pride Being gay fills me with joy. .75 .72 GISPS 23 Gay Pride Being gay makes me feel special. .75 .64 0113898 45 Gay Pride 1 am proud of myselfbecause I’m gay. .74 .68 GIBSPS 48 Gay Pride Being gay has proven to be a definite asset. .69 .61 GIBSPS 37 Gay Pride My gayness contributes to a sense of my being complete .67 .65 md whole. GIBSPS 30 Gay Pride 1 love to celebrate my gayness with others. .53 .68 GIBSPS 14 Gay Pride As a gay man, I feel I have much to be proud of. .50 .65 GIBSPS 17 Gay Pride As a gay man 1 have something valuable to contribute to .47 .63 society. GIBSPS 25 Gay Pride As a gay man I see myselfas powerful and masculine. .50 .70 Eigenvalue = 5.31 % of variance = 10.6 134 Table 19 O 1110,.10 .01 o r v1 .\ O,.\ .4 1.11 . t r 3.0‘ v: as . ’ .u-vt vx r t 11 Winner Gay Identity- - Gay Identity- . Based Sl Shame Factor based Pride Pnde-Factor Scale Scale Gay Identity- 1.00 .98 -.53 -.43 Based Shame = p =.000 p =.000 p 8.000 Scale 63' ’ [affix -.53 -.52 1.00 .90 Scale p =.000 p =.000 = p =.057 .98 1.00 -.52 -.43 Shame-Factor p = .000 = p = .000 p = .000 -.43 -.43 .90 1.00 Pride-Factor p = .000 p = .000 p = .000 p = .000 135 Table 20 E I l' E ! ll 5 l 5 Variable Factor 1 112 Gay Identity-based Shame -.79 .62 Gay Identity-based Pride .76 .57 Self-Esteem .85 .72 Negative Affectivity -.69 .47 Positive Afl‘ectivity .68 .45 Eigenvalue = 2.83 % of variance = 56.6 Tdrle 21 136 Item - Scale Item . .- 1 Number is From ltern Wording "3 1' GISPS 29 Gay Shame I would like to shrink away when I feel like I am a .79 .71 mistake as a gay man. GtBSPS 47 Gay Shame Being gay, I feel like there is something missing inside .75 .70 of me. GIBSPS 34 Gay Shame At times I feel like I will break into a thousand pieces, .75 .65 because I am gay. GIBSPS 44 Gay Shame Being gay, I feel empty and unfulfilled. .74 .74 GIBSPS 40 Gay Shame At times being gay makes me feel so exposed that I .72 .62 wishtheearthwouldopenupandswallowme. GIPS 16 Gay Shame Because I am gay, I feel as if 1 am somehow defective .72 .63 aperson, like there issomething basicallywrongwith me. GIBSPS 42 Gay Shame Being gay has left a painful hole within me that I have .69 .71 not been able to fill. GIBSPS 36 Gay Shame I feel as if I have lost control over my body, my life and .69 .57 my feelings because I am gay. GIBSPS 49 Gay Shame Compared to non-gay men, I feel lesser. .66 .64 6‘88 31 Gay Shame I rqilay painful events about my being gay over and .66 .55 overinmyminduntillamoverwhelmed. CUPS 12 Gay Shame Because of my sexual orientation, I see myself as being .64 .65 very small and insignificant. 013888 22 Gay Shame I see myself striving to be the perfect man only to .64 .57 continually fall short because I am gay. GIBSPS 15 Gay Shame As a gay man, I feel intensely inadequate and full of self .63 .65 doubt. amass 07 Gay Shame I scold myself and put myself down for being gay. .61 .54 GIBSPS 46 Gay Shame As a gay man, my loneliness is more like emptiness. .59 .68 GIBSPS 11 Gay Shame Compared to non-gays, I feel like I somehow never .57 .69 measure up. GISPS 18 Gay Shame When I compare myself to non-gays 1 am just not as .54 .42 important. GICPS 01 Gay Shame Because I’m gay, I feel like I am never quite good .53 .63 enough. GIBSPS 38 Gay Shame Because of my being gay, sometimes I feel no bigger .46 .39 than a pea. GfiPS 19 Gay Shame I have an overpowering dread that my being gay will be .44 .58 revealed in front of others. GIBSPS 28 Gay Pride On the whole. I am pleased and satisfied with myself as -.41 .66 a gay man. GIBSPS 09 Gay Shame I feel insecure about others opinions of my gayness. .41 .54 PANAS 13 Negative- Ashamed .43 .62 Affectivity Eigenvalue = 22.49 % ofvariance = 28.1 137 Table 22 Item Name . Loadin for and Sale Item rs From Item this F acgto r h: Number PANAS 04 Negative Affectivity Upset .76 .64 PANAS 15 Negative Affectivity Nervous .75 .69 PANAS 20 Negative Affectivity Afraid .74 .66 PANAS 07 Negative Afl‘ectivity Scared .73 .63 PANAS 02 Negative Affectivity Distressed .73 .62 PANAS 18 Negative Affectivity Jittery .71 .62 PANAS 11 Negative Affectivity Irritable .69 .57 PANAS 08 Negative Affectivity Hostile .56 .45 PANAS 06 Negative Afi'ectivity Guilty .54 .63 PANAS 13 Negative Afl‘ectivity Ashamed .50 .62 Eigenvalue = 6.32 % of variance = 7.9 Table 23 138 Load - - 2 a... 83:11“ ........... -- '- GIBSPS 27 Gay Pride My gayness is one of the fmest aspects of myself. .78 .67 GIBSPS 33 Gay Pride Being gay fills me with joy. .75 .72 class 3 Gay Pride Being gay makes me feel special. .75 .64 CUPS 45 Gay Pride 1 am proud of myself because I’m gay. .74 .68 031’s 48 Gay Pride Being gay has proven to be a definite asset. .69 .61 GIBSPS 37 Gay Pride My gayness contributes to a sense of my being complete .67 .65 and whole. GIBSPS 30 Gay Pride 1 love to celebrate my gayness with others. .53 .68 GIBSPS 14 Gay Pride As a gay man, I feel I have much to be proud of. .50 .65 0113888 17 Gay Pride As a gay man I have something valuable to contribute to .47 .63 society. Eigenvalue = 4.88 %ofvariance = 6.1 Table 24 139 Item Name . and Scale Item rs From Item Loading for h: Number this Factor PANAS 09 Positive Afl'ectivity Enthusiastic .77 .68 PANAS 16 Positive Afl'ectivity Determined .69 .57 PANAS 17 Positive Affectivity Attentive .69 .56 PANAS 19 Positive Afl'ectivity Active .69 .56 PANAS 01 Positive Afl'ectivity Interested .68 .52 PANAS 14 Positive Aflectivity Inspired .66 .53 PANAS 12 Positive Affectivity Alert .66 .56 PANAS 03 Positive Affectivity Excited .63 .53 PANAS 05 Positive Afi'cctivity Strong .55 .59 PANAS 10 Positive Aflectivity Proud .53 .58 Eigenvalue =- 3.08 % ofvariance = 3.9 140 Table 25 WES Item Name . Loadin“ h2 and Number Scale Item rs Item for thisg From Factor Self-esteem 03 Rosenng SES I feel I have a number of good .67 .62 qualities Self-esteem 09 Rosenberg SES All in all. I am inclined to feel -.66 .63 that I am a failure. Self-esteem 07 Rosenberg SES I feel that I’m a person of worth. .65 .62 at least on an equal plane with others. Self-esteem 02 Rosenberg SES At times I think I am no good at -.64 .61 all Self-esteem 10 Rosenba'g SES 1 take a positive attitude toward .63 .69 myself. Self-esteem 01 Rosenberg SES On the whole, I am satisfied with .62 .69 myself. Self-esteem 05 Rosenberg SES I feel I do not have much to be -.60 .49 proud of. Self-esteem 04 Rosenba'g SES 1 am able to do things as well as .58 .58 most other people. Self-esteem 06 Rosenberg SES I certainly feel useless at times. -.56 .53 Self-esteem 08 Rosenbu'g SES 1 wish 1 could have more respect -.55 .54 for myself. Eigenvalue = 2.26 960me = 2.8 141 mg "2 .nn 891a acorn 89!.— 8o.ln acorn 8a....- 8Qun ocean 58m cc.— 9... 2 . may 3.. 3.. 2. ca. 3. - 03r— .um 03.13 acorn 89!.— 8919 891.— 25.“... acorn ~88..— 8._ “Nov 3. Sr mm .. 2 .- 3.. no. one-am .na Segue 8a.“..— 8o.nn «8.....— 25.1.— Sntn 9.56: 84 3... ea. 2 . 2. .m. 8.- -32 .un 86. n .— 8c. u .— 8c. 1 a :8. u A. 8a. a .— §>a8b< co; 3.- an .1 «nr 8.. mm. 25.qu .1. 8o. 1 a So. u a :8. n .— Se. 1 a £385 8.. mn. ea. 2.. «nr 333.— .ne 8a."... Sore Sore gm 2: «m. 3. me..- ..=om .un 86. n a 8:. u a 0:320 84 Z . 2 .- 63232 013m .1. So. a a 03.5 8.. nnr ~88. 23m ."a— 033m 8.. 33. Essen Sauna g §>u8b< bm>uoob< 988%..— 2565 018m 018 .85 2.5 -82 2582 258.. e3 -8652 85 «can 38. .58. x... . 0 ...4'..uafi:.:é.r 3......” .... .....u:.. ea 035. 142 Table 27 Rehtive Parsimony Model x2 (If ledf GFI' NFI PGFI Absolute indices Five Factor: l.Self-Esteem 2.Neg. Afl’ect 3.Pos. Affect 4.Pride 5.Shame 2026.320 511 3.97 .91 .90 .83 ‘l‘wo Factor: 1 Self- Esteem Positive Afl'ectivity Pride 6389.805 526 12. 15 .65 .66 .62 2 Negative - Afl'ectivity Shame Note: GFI' = Goodness of Fit Index; NFI = Normed Fit Index; PGFI = Parsimonious Goodness of Fit Index. n = 973 to correct for social desiriaility response bias Table 28 \ 143 Maximum Factor Item Item Number Likelihood SE?“ B. E": Forced to Load 1* Value Estimate On Rosenberg SES .807 .027 Self-Esteem 29.78" 1 SES 2 .716 .029 Self-Esteem 25.07” SES 3 .585 .030 Self-Esteem 19.32” SES 4 .558 .031 Self-Esteem 18.25" SES 5 .603 .030 Self-Esteem 20.08” SES 6 .689 .029 SelfoEsteem 22.88“ SES 7 .692 .029 Self-Esteem 23.93” SES 8 .669 .029 Self-Esteem 22.92” SES 9 .737 .028 Self-Esteem 26.08" SES 10 .815 .027 Self-Esteem 30.21“ New“: .738 .028 Negative Afl‘ect 26.13“ Afl'ectivity 1 NA 2 .740 .028 Negative Afl‘ect 26.21” NA 3 .604 .030 Negative Affect 20.08" NA 4 .783 .028 Negative Afi'ect 28.40" NA 5 .537 .031 Negative Affect 17.42“I NA 6 .644 .030 Negative Afl’ect 21.78" NA 7 .636 .030 Negative Afl‘ect 21.43” NA 8 .745 .028 Negative Affect 26.43" NA 9 .698 .029 Negative Afl'ect 24.21" NA 10 .811 .027 Negative Afl'ect 29.94" ll’ositive Affectivity .623 .030 Negative Affect 20.66” PA 2 .776 .030 Negative Afl‘ect 18.75” PA 3 .656 .029 Negative Affect 22.06" PA 4 .771 .028 Negative Afl’ect 27.50” PA 5 .691 .029 Negative Affect 23.63“l PA 6 .614 .030 Negative Afi'ect 20.26“ PA 7 .656 .030 Negative Afl'ect 22.05” PA 8 .711 .029 Negative Afl‘ect 24.52” PA 9 .626 .030 Negative Afl‘ect 20.78" PA 10 .676 .030 Negative Afl'ect 22.91” Shame .966 .025 (W Identity- 39.29" Composite 1 based Shame Shame .946 .025 Gay Identity- 37.91“ Composite 2 based Shame Three .787 .030 Gay Idartity- 26.55" Composite 1 based Pride Pride .945 .029 Gay Identity- 33.16" Composite 2 based Pride Note: “ p < .001 144 Table 29 Factor 1 Factor_2 Factor 3 Item Loading Residual Loading Residual Loading Residual Rosenberg SES .807 .590 1 SES 2 .716 .698 SES 3 .585 .811 SES 4 .558 .830 SES 5 .603 .798 SES 6 .669 .744 SES 7 .692 .722 SES 8 .669 .743 SES 9 .737 .676 SES 10 .815 .580 Nee-live Affectivity .738 .674 1 NA 2 .740 .673 NA 3 .604 .797 NA 4 .783 .622 NA 5 .537 .844 NA 6 .643 .765 NA 7 .636 .772 NA 8 .745 .668 NA 9 .698 .716 NA 10 .811 .585 Positive Afi'ectivity .623 .782 1 PA 2 .576 .817 PA 3 .656 .755 PA 4 .771 .636 PA 5 .691 .723 PA 6 .614 .790 PA 7 .660 .755 PA 8 .711 .704 PA 9 .626 .780 PA 10 .676 .737 Note: Factor 1 = Self-Esteem: Factor 2 = Negative Afl'ectivity; Factor 3 = Positive Afi'ectivity n = 973 to correct for social desirability response bias 145 Table 30 Factor 4 Factor 5 Item Loading Residual Loading Residual Pride .786 .617 Composite 1 Pride .945 .327 Corrposite 2 Shame .966 .256 Conposite 1 Shame .949 .315 Composite 2 Note: Factor 4 = Gay Identity-based Pride; Factor 5 = Gay Identity-based Shame n= 973 to correct forsoeial desirability responsebias 146 Table 31 DEC l'll'fiEiE llll . Self- Positive Negative Sham” m“ Esteem Affectivity Afl’ectivity Shame 1.00 1" Pride -.54 1.00 p=.000 p=. Self-Esteem -.58 .60 1.00 Positive Affectivity -.35 .61 .61 1.00 Negative Afi‘ectivity .55 -.34 -.58 -.28 1.00 p=.000 p=.000 p=.000 p=.000 p=. n = 973 to correct for social desirability response bias 147 Table32 c. '1' -‘!° U‘r" 'U_-.H 11 .‘10101' .111: '10,.“ ll. 0 1 ‘1'". Maximum Factor Item Item Number Likelihood 53:5“?! 5'2“ Forced to Load 1' Value Estimate On Rosenberg SES .771 .028 Factor 1 27.95“ 1 SES 2 .649 .029 Factor 1 22.16” SES 3 .561 .030 Factor 1 18.54” SES 4 .543 .030 Factor 1 17.79" SES 5 .582 .030 Factor 1 19.33" 5155 6 .623 .030 Factor 1 21.05" SES 7 .655 .029 Factor 1 22.42“ 8128 8 .614 .030 Factor 1 20.61“ SES 9 .683 .029 Factor 1 23.68“ SES 10 .782 .027 Factor 1 28.57“ " Nee-u“ ve '40" activity .730 .028 Factor 2 25.75" NA 2 .723 .028 Factor 2 25.44" NA 3 .607 .030 Factor 2 20.23" NA 4 .772 .028 Factor2 27.86" NA 5 .532 .031 Factor 2 17.26“ NA 6 .634 .029 Factor 2 21.46“ NA 7 .653 .029 Factor2 22.33" NA 8 .735 .028 Factor 2 25.97" NA 9 .689 .028 Factor 2 23.84" NA 10 .802 .027 Factor2 29.50“ forum Afiecuvity .493 .031 Factor 1 15.89" PA 2 .407 .032 Factor 1 12.81“ PA 3 .619 .029 Factor 1 20.87" PA 4 .579 .030 Factor 1 19.18" PA 5 .664 .029 Factor 1 22.79" PA 6 .493 .031 Factor 1 15.89" PA 7 .512 .030 Factor 1 16.59" PA 8 .589 .030 Factor 1 19.61“ PA 9 .487 .031 Factor 1 15.66" PA 10 .529 .031 Factor 1 17.24" Shame .609 .030 Factorz 20.34" Composite 1 Shame .595 .030 Factor 2 19.73“ Composite 2 hide Comm: .520 .031 Factor 1 16.88” Pride Comm 2 .652 .030 Factor 1 21.11" Note: ” p < .001. Factor 1 = Self-esteem, Positive Affectivity, Wide; Factor 2 = Negative Afi‘ectivity , Shame. 148 Table 33 .' no .010 unr- .01. 0401' 0. '5' em 0 ‘. 35"?!“ ”Writ-31" W Factor 1 Factor 2 Item Loading itesiduai Loading Read—um ' Rosenberg SES 1 .770 .638 SES 2 .649 .761 SES 3 .562 .827 SES 4 .543 .840 SES 5 .582 .813 SES 6 .624 .782 SES 7 .655 .755 SES 8 .613 .790 SES 9 .683 .730 SES 10 .782 .623 Nesflive Afi'ectivityl .730 .683 NA 2 .724 .690 NA 3 .607 .795 NA 4 .772 .636 NA 5 .532 .847 NA 6 .636 .772 NA 7 .656 .755 NA 8 .735 .679 NA 9 .690 .724 NA 10 .502 .600 Positive Affectivity 1 .493 .870 PA 2 .401 .914 PA 3 .620 .785 PA 4 .578 .816 PA 5 .664 .748 PA 6 .493 .870 PA 7 .512 .859 PA 8 .589 .808 PA 9 .487 .874 PA 10 .529 .549 Note: Factor l= Self-esteem, Positive Afl'ectivity, Pride; Factor 2 = Negative Affectivity , Shame. n = 973 to correct for social desirability response bias. 149 Table 34 q. a. .fi. “0‘0"- .0}. I._IQ' O OHIO‘O' M" O I."'.l.0.‘...0' Emma: Factor 1 Factor 2 Item Loading Residual Loading Residual Pride Composite 1 Pride Composite 2 Shame Composite 1 Shame Composite 2 .520 .855 .625 .781 .610 .793 .595 .803 Note: Factor 1== Self-esteem. Positive Afl'cctivity, Pride; Factor 2 = Negative Afl‘ectivity , Shame n = 973 to correct for social desirability response bias. 150 Table 35 CEECI’M°EIE HII Factor 1: Self-Esteem Factor 2: Negative Positive Afl‘ectivity Afl‘ectivity Pride Shame Factor 1: 1.00 Self-Esteem Positive p=. Afl‘ectivity Pride Factor 2: -.57 1.00 Negative Affectivity p=.000 p=. Shame n = 973 to correct for social desirability response bias 151 Table 36 Model Test Model Comparison Model x2 (if Comparison x2 at Fm Fm“ 2026 320 511 M '- M2 . . - M2 4363.49 15 Two Factor. 6389.805 526 M1 Note: " p < .001 n = 973 to correct for social desirability response bias 152 Table 37 n ’ I‘x‘fil‘!!!‘ {cursor at... ' 'tu. stir-t. a um 11 in 's 9? w W Predictor Beta l R l R2 Change F Change Demographics - entered as a .28 .08 8.91 to. block: Current Comfort w/ Sexual 02 Orientation -' Ethnicity .04 Education Level -.24 ”‘ Age .17 ‘” Relationship Status -.02 Current Sexual Orientation 01 Identity ' Income Level -.05 Employment Status -.05 Future Sexual Orientation 09 Identity ° SE & PANAS entered as a .40 08 32.41‘” block: Self-Esteem .03 Positive Afl‘ectivity .22 “‘ Negative Affectivity .05 GIB Shame & Pride - .44 03 20.31‘” entered as a block: 613 Pride .23 "‘ GIB Shame .15 “‘ GIB Pride x 618 Shame .04 Note: 11 = 973 to correct for social desirability response bias " p< .05 " p< .01 ”" p< .001 Table 38 O'quOt O 153 sent; 0.5.0 l-H .gt mm...” o ‘ 05' 1°16 r1 21 '15] E IE'I E E EE! ICE! Gay Identity- EFA & CFA Gay Identity- EFA & CF A Based Shame Shame-Factor based Pride Pride-Factor Scale Scale Gay Identity- 1.00 .98 -.53 -.43 Based Shame =. p =.000 p =.000 p 8.000 Scale gaggfig' -.53 -.52 1.00 .90 Scale p =.000 p =.000 p = p 8.057 Marlowe- -.17 -.16 .13 .13 Crowne p =.000 p =.000 p =.000 p =.000 Self-Esteem -.53 -.53 .50 .40 pemm pamm pawn paw” Mac-Andrews -.03 -.03 .21 .17 p =.314 p =.426 p =.000 p =.000 Positive -.32 -.32 .46 .41 Afi'ectivity p =.000 p =.000 p =.000 p =.15 Negative .53 .52 -.29 -.23 Affectivity p =.000 p =.000 p =.000 p =.000 n = 973 to correct for social desirability response bias 154 Table 39 u'e—sr-tt-re: {hut-5.01“» . A lU't ant-1.82310 n '41. \\ or.» Predictor Beta R R2 Change F change Demographics - entered as a .28 .08 8.91 ‘” block: Current Comfort w/ Sexual _ 03 Orientation ' Ethnicity .04 Education Level -.24 ‘” Age .16 ‘” Relationship Status -.02 Current Sexual Orientation 01 Identity ' Income Level -.05 Emfloyment Status -.04 Future Sexual Orientation 09 Identity ' SE & PANAS - entered as a .40 .08 32.41 ‘” block: Factor 1: Self-Esteem .05 Factor 2: Negative Affectivity .04 Factor 3: Positive Affectivity .23 ”‘ GIB Pride & Shame - entered .43 .03 14.34 ”‘ as a block: Factor 4: GIB Pride .17 ‘” Factor 5: GIB Shame .14 ‘” GIB Pride x GIB Shame .001 Note: _n = 973 to correct for social desirability response bias " p< .05 " p< .01 ”‘ p< .001 155 Table 40 .A it’- 'a‘a°|".\ ‘. l‘fi‘h ' - \' r'i‘ Uh" 'H‘O-t'.’ .‘ 1' \' ~‘1' Uh". Relative Parsimony Absolute mdlces indices lndrces Model 36 or {lot GFI NFI PGFI Original Five- Factor. 1.Se1f- Esteem 2.Neg. Afl'ect. 3.Pos. Afl’ect. 4.Pride 5.Shame 2026.320 511 3.97 .9104 .905 .83 lst. Split-Half Five Factor. 1.Self- Esteem 2.Neg. Afl'ect. 3.Pos. Afi‘ect 4.Pride 5.Shame 2nd Split- Half Five Factor. l.Self- Esteem 2.Neg. Afl'ect. 3.Pos. Affect. 4.Pride 5.Shame 1608.800 511 3.15 .87 .85 .79 1680.629 511 3.29 .86 .84 .78 Note: GFI' = Goodness of Fit Index; NF1= Normed Fit Index; PGFI = Parsimonious Goodness of Fit Index. n = 973 to correct for social desirability response bias 156 Table 41 OvHU'U.I"I'-.'F 011.»:- {nut-racial. u. “tr-t1; 01011:“. \t' W Predictor(s): Beta R R2 F Beta R R2 F Chang Chmg Chang Chang 7 e e , 8 e Demographics - .34 .12 7.02‘ 25 .07 323‘ entered as a block: Current Comfort w/ 01 05 Sexual Orientation " " Ethnicity .04 .02 Education Level -.26‘ -.23‘ Age .15. .18‘ Relationship Status -.01 -.06 . cm?“ 3"“?! .11' .05 Orientation Identlty Income Level -.07 -.04 Employment Status .05 -.03 Future Sexual Orientation Identity '06 '08 SE & PANAS .43 .06 12.96 .41 .09 19.67 entered as a block: ’ ' Self-Esteem .01 .06 Positive Affectivity .22‘ 22’ Negative Afl'ectivity .05 .01 GIB Shame & Pride .46 03 7.74. ' ' .45 .04 11.25 - altered as a block: ‘ GIB Pride .22‘ .23‘ GIB Shame .11 .20‘ GIB Pride x GIB .07 .09 Shame Note: n = 973 to correct for social desirability response bias ’p<.001 FIGURES 157 Figure Captions Hum. Regression model for the study. Bum]... Scree plot of eigenvalues from principle components EFA of the Gay Identity-based Shame and Pride Scales. Final. Scree plot of eigenvalues from principle components EFA of all items from all scales. Eignmi Five factor confirmatory factor analysis model. Emmi. Histogram of the residuals for regression analysis. W. Scatterplot depicting homoscedasticity of residuals of the regression equation. Figural. Normal probability plot of standardized residual. 158 n Joe—fl £2 58...? 243qu be: .565. 0333 808mm ”=20. N due—m + 82%..» .259 — sue—fl 8:? 83am mmcr>=>zuot< gamed v w 23:66.? gamed 2|— ( k ( or .. — 2:2. EOQuflNtu-Ow N t F 9:2. 238500 onto. >emv N .. p mEo: 0:89:00 2:96 30 ’ a u .528 .. gow ) a u cute :60 ) _ M 9:95 :60 162 N EXPN (’=3Cases, .:=NormalCurve) 2 .75 0111‘I 3 1.49 3.00 “ 6 3.80 2.67 2" 7 8.68 2.33 "'. 17 17.75 2.00 “”"2 33 32.53 1.67 ”“nn”: 58 53.33 1.33 eeeeeeeeeeeeeaeeeze 77 73 43 l. 00 canteen-teeeaeaeeeeeeeaeeaz 33 103 34 67 eecueateeeeeeeeeeaeaeseeeeae 109 121 90 33 Intetanus-aaeeeeeeeeeeeeeeeeattenuate _ 129 128.79 .00 teen-u-atattenuateeeeeeeeeaeeeeeeeeeeeaeez 132 121.90 -33 sect-nun"eacetateseeeeeeeeeeeeeeeeeeeeezeta 109 103.34 -.67 crustacean-aeeeeeaaeeeeeeeeeeeeee:ea 90 73.48 -100 accentu-eeeeaaeeeeeeaeeeezeeee 53 53.33 -133 eeeeeeeeeeeeeeeee: 38 32.53 -157 tenets-rumpus 14 17.75 -2.00 "W", 6 8.68 -2.33 “. 2 3.80-2.67: 0 1.49 -3.00 0 .75 Out 163 Across - ‘PRED Down - *SRESID Out ++~——+-. , +__+— , .—.+——++ 3 + Symbols: Max N . 3.0 : 6.0 ‘ 12.0 .1. l . l ' l I + . tttttttfit-t. -- "-.‘fitfitttfiiot- -l-.-.‘ .o.- - -2 .. ..'.Z'."."."I ' + I .3 + i . + Out H--+..—+——+—...+—-+—.-+r -3 -2 -1 0 1 2 3 Out 164 t. 1‘" 0‘. I ‘0‘ -HMmd l+||||+l|l|+||l|+ u can use. + .3 one ace .8 t can + can. can. can I." llll+llll+llll+lllle 5 5 7 5. 2. Expected 1.0 .25 APPENDICES SUBJECT DEMOGRAPHICS 1) AGE: _ 165 Appendix A 2) GENDER: _ Male __ Female Please Chad: One Below 3) ETHNIC OR RACIAL IDENTIFICATION: __ Black /African American _ Hispanic, Chicano, Latino or Mexican American White or Caucasian Asian/Oriental/Pacific Islander Native American Other (specify) 4) HIcuEsT LEVEL OF EDUCATION: _ Less than High School. diploma High School diploma Some college, no degree Associates degree Bachelor's degree Masters, professional degree or beyond 5) DID YOU WORK AT ALL INTnE PAST FEW warms? _ Yes- Employed _ No - Unemployed __ No - Laid off _ No - Never have worked NO - Retired If employed, what is your occupation: (please write in, for exanple: registered nurse, grinder operator, supervisor, etc.) 6) Wmcn OF TIIE FOLLOWING CATEGORIES BEST DESCanEs YOUR TOTAL INCOME IN 1993 ? Less Than $5,000 _ $5,000 to $14,999 $15,000 to $24,999 $25,000 to $39,999 $40,000 to $59,999 $60,000 Or More 166 7) What is your current relationship status: Single, no sexual partners Single. one committed partner Single, multiple partners Coupled, living together (Committed to an exclusive sexual relationship) Coupled, living together (Relationship permits other partners under certain circumstances) Coupled, living apart (Committed to an exclusive sexual relationship) Coupled. living apart (Relationship permits other partnas lmder catain circumstances) Other (please specify) 8) In tenns of my present sexual 9) In the future, I would like to orientation, I identify myself as . . . identify myself as . . . Exclusively homosexual Exclusively homosexual Predominantly homosexual Predominantly homosexual Bisexual Bisexual Predominantly heterosexual Predominantly heterosexual Exclusively heterosexual Exclusively heterosexual Unsure Unsure 10) In terms of comfort with my current sexual orientation, I would say that I am . . . __ Very comfortable __ Mostly comfortable _ Comfortable _ Not very comfortable Very uncomfortable 167 Appendix B THE MARLOWE-CROWNE SOCIAL DESIRABIIJTY ScALE - SHORT FORM 0 Listed below are a number of statements concerning personal attitudes and traits. Read each one and decide whether the statement is true or false as it pertains to you personally. 1. It is sometimes hard for me to go on with my work ifl T F am not encouraged. 2. I sometimes feel resentful when I don‘t get my way. T F 3. On a few occasions. I have given up doing something T F because I thought too little of my ability. 4. There have been times when I felt like rebelling T F against people in authority even though I knew they were right 5. No matter who I'm talking to, I'm always a good T F listener. 6. There have been a few occasions when I took T F advantage of someone. 7. I'm always willing to admit it when I make a mistake. T F 8. I sometimes try to get even rather than forgive and T F forget. 9. I am always courteous, even to people who are T F disagreeable. 10. l have never been irked when people expressed ideas T F very different from my own. 11 . There have been times when I was quite jealous of the T F good fortune of others. 12. I am sometimes irritated by people who ask favors of T F me. 13. l have never deliberately said something that hurt T F someone‘s feelings. 168 Appendix C GAY IDENTITY-BASED SHAME-PRIDE SCALE (GIBS-PS) (title for subjects: Gay Lifestyle Issues) W Below is a list of statements describing feelings or experiences that you may have from time to time or that are familim' to you bemuse you have had these feelings and experiences for a long time. Some people will seldom or never have had many of these feelings. Everyone has had some of these feelings at some time, but if you find that some of these painful statements describe the way you feel a good deal of the time, it ean be difficult just reading them. Try to be as honest as you can in responding Read each statement carefully and circle the number to the right of the item that indicates the frequency with which you find yourself feeling or experiencing what is described in the statement. Use the seale below. . Almost Never Seldom Someumes Frequently Always 1.Beeause1'm gay, I feel like I am never 1 2 3 4 5 quite good enough. 2. I am "out" to most everyone I work l-—-—-—-----2 3 4 5 with. 3. Being gay, I feel somehow left out. 1 2 3 4 5 4. Ithinkthatpeople look down on measa I---—--—-—-—2 3 4 5 gay man. 5.Allinall,includingmysexual l 2 3 4 5 orientation, I am inclined to feel that I am a success. 6.1readi1y disclosetoothersthatlamgay. 1 2 3 4 5 7. I scold myself and put myself down for 1 2 3 4 5 being gay. 8. Asagaynmlholdmyhmdupand 1 2 3 4 5 look other people directly in the eye. 9. I feel insecure about others opinions of l—-——-—--—-2 3 4 5 my gayness. 10. Ifeel atmseandrelaxedasagayman l—-—-----2 3 4 5 when I'm around non-gay men 11. Compared to non-gays, I feel like I somehow never measure up. 12. Because of my sexual orientation, 1 see myself as being very small and insignifieant. I3. I am ”out" to most everyone in my family. 14. Asagaymanlfeellhavemuchtobe proud of. 15. As a gay man, I feel intensely inadequate and full of self doubt. 16. Because I am gay, I feel as ifl am somehow defective as a person, like there is something basically wrong with me 17. As a gay man I have something valuable to contribute to society. 18. When I compare myself to non-gays I am just not as important. 19. I have an ovapowering dread that my being gay will be revealed in front of others. 20. As agay man, I feel I haveanumberof good qualities. 21. As a gay man I feel good about my body, my life and my feelings. 22. I see myself striving to be the perfect man only to continually fall short beeause I am gay. 23. Being gay makes me feel special. 24.1thinkothersareabletoseethatlam 833'- 25. Asagay manlseemyselfaspowerful and masculine. 169 . Almost N ever Seldom Sometimes Frequently Always l 2 3 4 5 l 2 3 4 5 l 2 3 4 5 I 2 3 4 5 1 2 3 4 5 l 2 3 4 5 1 2 3 4 5 l 2 3 4 5 l 2 3 4 5 l 2 3 4 5 l 2 3 4 5 I 2 3 4 5 l 2 3 4 5 l 2 3 4 5 l 2 3 4 5 26. Iworry thatbeeauselamgay mybody does not appear masculine enough. 27. My gayness is one of the finest aspects of myself. 28. On the whole, I am pleased and satisfied with myself as a gay man. 29. Iwould like to shrinkaway whallfeel likelamamistakeasagayman 30. I love to celebrate my gayness with others. 31. I replay painful events about my being gay over and over in my mind until I am overwhelmed. 32. Asagaymamlfeellamapasonof worth at least on an equal plane with others. 33. Being gay fills me with joy. 34. At times I feel like I will brmk into a thousand pieces, because 1 am gay. 35. I feel both free and comfortable telling other people that I am gay. 36. I feel as if I have lost control over my body, my life and my feelings because I am 833’- 37 . My gayness contributes to a sense of my being complete and whole. 38. Beeause of my being gay, sometimes I feel no bigger than a pee 39. I feel a sense of connection with others in the gay community. 170 Never Seldom Sometimes 142011116le 23:; 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 l 2 3 4 5 l 2 3 4 5 l 2 3 4 5 1 2 3 4 5 1 2 3 4 5 l 2 3 4 5 l 2 3 4 5 l 2 3 4 5 1 2 3 4 5 l 2 3 4 5 l 2 3 4 5 40. At times being gay makes me feel so exposed that I wish the earth would open up and swallow me. 41. I enjoy my being gay. 42. Being gay has left a painful hole within me that I have not been able to fill. 43. I feel very relaxed and comfortable when I am around other gay men. 44. Being gay, I feel empty and unfulfilled. 45. I am proud of myself because I am Kai’- 46. Asagaymnmylonelinessismore like emptiness. 47. Being gay, I feel like there is something missing inside of me. 48. Being gay has proven to be a definite asset 49. Compared to non-gay men, I feel lesser. 50.1arnpmudofmyselfasagay rmn when I am around other men. 171 Never 1 Almost Seldom Sometimes Frequently Always 2 3 4 5 2 3 4 5 2 3 4 5 2 3 4 5 2 3 4 5 2 3 4 5 2 3 4 5 2 3 4 5 2 3 4 5 2 3 4 5 2 3 4 5 172 Appendix D ROSENBERG SELF-ESTEEM SCALE QIBEQJJQNE: This questionnaire contains 10 items describing your self-esteem. For each question, please indicate your rating on the scales to the right of each question. Remember, there are no right or wrong answers. just ratings of how you feel. STRONGLY STRONGLY AGREE AGREE DISAGREE or SA GREE 1. On the whole, I am satisfied with myself. 1 2 3 4 2. At times I think I am no good at all. 1 2 3 4 3. I feel that I have a number of good qualities. 1 2 3 4 4. I am able to do things as well as most other 1 2 3 people. 5. I feel I do not have much to be proud of. 1 2 3 4 6. I certainly feel useless at times. 1 2 3 4 7. I feel that I'm a person of worth. at least on an 1 2 3 4 equal plane with others. 8. I wish I could have more respect for myself. 1 2 3 4 9. All in all, I am inclined to feel that I am a 1 2 3 4 failure. 10. I take a positive attitude toward myself. 1 2 3 4 173 Appendix E MACANDREW ALCOHOUSM SCALE (MAC) (title for subjects: Personal Interests Inventory) Directions: Please indicate if you agree or disagree that the following items apply to your life. 444444 444 44444 444444 444 444444 44444 'n'n'n'nm'n 'I'I'TI'TI 1111111111 111111111111 'Tl'l'l‘fl TITITI'TI'H'I'I 1111111111 9995’." Ad 13. 14. 15. 16. 17. 18. 19. 20. 21. 23. 24. 25. 27. 28. 29. 30. 31. 32. 33. 999949 I like to read newspaper articles on crime. Evil spirits possess me at times. I have a cough most of the time. My soul sometimes leaves my body. As a youngster I was suspended from school one or more times for cutting up. I am a good mixer. Everything is turning out just like the prophets of the Bible said a would. I have not lived the right kind of life. I think I would like the kind of work a forest ranger does. I am certainly lacking in selfconfidence. I do many things that I regret afterwards (I regret things more or more often than others seem to). I enjoy a race or a game better when I bet on it In school I was sometimes sent to the principal for cutting up. My table manners are not quite as good at home as when I am out in company. I know who is responsible for most of my troubles. The sight of blood neither frightens me nor makes me sick. I have never vomited or coughed up blood. I like to cook. I used to keep a diary. l have had periods in which I canted on activities without knowing later what I had been doing. I liked school. I am worried about sex matters. I frequently notice that my hand shakes when I try to do something. I have used alcohol excessively. My parents have often objected to the kind of people I went around with. I have been quite independent and free from family rule. I have few or no pains. I have had blank spells in which my activities were interrupted and I did not know what was going on around me. I sweat very easily even on cool days. I have often felt that strangers were looking at me critically. Ifl were a reporter I would very much like to report sporting news. I have never been in trouble with the law. I seem to make friends about as quickly as others do. Many of my dreams about sex matters. 44444 4444 444444 44 1111111111 11111111 111111111111 11'“ 174 MAcANoREw ALCOHOUSM SCALE (MAC) - cormuuen 35. 38. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. I cannot keep my mind on one thing. I have more trouble concentrating than others seem to have. I do not like to see women smoke. I deserve severe punishment for my sins. I played hooky from school quite often as a youngster. l have at times had to be rough with people who were rude or annoying. I was fond of excitement when l was young (or in childhood). I enjoy gambling for small stakes. l have used alcohol moderately (or not at all). Ifl were in trouble with several friends who were equally to blame, I would rather take the whole blame than to give them away. While in trains. busses. etc., I often talk to strangers. Christ performed miracles such as changing water into wine. I pray several times every week. I readily become one hundred per cent sold on a good idea. I have frequently worked under people who seem to have things so that they get credit for good work but are able to pass off mistakes onto those underthem. I would like to wear expensive clothes. The one to whom I was most attached and whom I most admired as a child was a woman. (Mother. sister. aunt or other woman.) 175 Appendix F THE POSITIVE AND N EGATIVE AFFECTIVI'I'Y SCALE (PANAS) (title for subjects: Feelings Questionnaire) W This scale consists of a number of words that describe different feelings and emotions. Read each item and then mark the appropriate answer in the space next to that word. Indicate to what extent you generally feel this way. that is. how you feel on the average. Use the following scale to record your answers. GENERALLY, I FEEL .... ..... PLEASE CRCLE ONE 1: very eligrtty or e lltle moderately qtlte a bit extremely not at al interested 1 2 3 4 5 distressed 1 2 3 4 5 excited 1 2 3 4 5 upset _ 1 2 3 4 5 strong 1 2 3 4 5 guilty 1 2 3 4 s seared 1 2 3 4 5 hostile 1 2 3 4 5 enthusiastic 1 2 3 4 5 proud 1 2 3 4 5 irritable 1 2 3 4 5 alert 1 2 3 4 5 ashamed 1 2 3 4 5 inspired 1 2 3 4 5 nervous I 2 3 4 5 detemrinad 1 2 3 4 5 attentive 1 2 3 4 5 jittery 1 2 3 4 5 active 1 2 3 4 5 afraid I 2 3 4 5 (“33$ Tniversity l'nivcrsity Counseling Center . Fort Collins. (Colorado 50523 Appendix G Mini 4') I 4.053 July XX, 1994 F.\.\': (Jim ~I'll-2382 Subject Name 123 Oak St. Anytown, USA 99999 Gay men are often portrayed in the media in simplistic and stereotypic ways. Speaking from the perspective of a gay male professional. I have been disappointed with the lack of information in psychology. In particular, the lack of knowledge about the diversity of gay men and their perceptions of themselves and their experience is most troubling. Therefore. I am asking for your help so that we might collectively add to the factual body of knowledge about our unique group. You are among a small number of gay men who are being asked to give your thoughts. feelings. and opinions. Your name was randomly drawn from the National Community Masterfile, a database of direct mail-responsive gay men. In order that the results will truly represent the thinking of the diverse population of gay men in the United States, it is important that each questionnaire be completed and returned. It is important, however, that you understand that your participation is completely voluntary. You may refuse to answer any or all questions in this survey. If you do complete and return this survey, you indicate your voluntary agreement to participate by completing and returning this questionnaire. You can expect that it will take about 20-30 minutes of your time. You may be assured of complete anonymity. Your name will never and should never be placed on the questionnaire. In addition. all results of the survey will be reported as group averages and individual responses will never be listed. Our main goal is to see what gay men have to say. not identify who they are. Therefore. we ask that you carefully follow the instructions in your survey packet so as to keep your identity anonymous. The results of this research will be used for a doctoral dissertation through Michigan State University, as well as to add to the growing knowledge base reflecting issues of diversity and wellness within the gay community. You may receive a summary of the results by checking the box marked "copy of results requested” on the enclosed yellow postcard, and printing your name and address on it. In order to maintain your confidentiality, please 51m put this information on the questionnaire itself. A complete set of detailed instructions can be found on the first page of your survey packet. I would be most happy to answer any questions that you might have. It is possible, however unlikely. that some of the questions in this survey may remind you of, or cause you to think about. some things that you feel the need to process or talk about with someone. If this is the case. please feel free to call me collect. My telephone number is (303) 491- 7850. If you have any questions about the survey of any kind. please feel free to contact me about them Thank you for your assistance. Siegemly. . I Wattu/ L/vl- 124,? Daniel W. Socall, MA. Senior Staff Therapist Accredited by the International Association of Counseling Services Internship Training Program approved by the American Psychological Association 177 Appendix H Pro-Survey Postcard Dear Potential Survey Participant. You are among a select group of gay men who have been randomly chosen. from national database of direct mail responsive gay men, to participate in a study. This study conducted by a gay male therapist. is designed to look at gay men's attitudes in a ga affirmative manner. In about a week. you should receive a survey packet titled,‘A Male- Mail Survey: Gay Men Report on their Perceptions of Themselves.‘ In order for the resu to accurately reflect the health and diversity of gay men in this country. it is highly importan we have as many gay men participating in this study as possible. Please take time t complete and return this survey when it arrives. Your participation in this study will b completely anonymous and will help provide information that can greatly benefit ou community. A full set of instructions and explanations will be included with your surve packet. If you know at this time that you will not complete the survey. or if you are not gay male. please contact me at (303) 491-7850. I thank you in advance for your help and participation in this study I Sincerely. Daniel W. Socall, MA. Senior Staff Counselor 178 Appendix I Post-Survey Follow-up Posteard Dear Survey Recipient. In the past few weeks. you should have received a survey titled.'A Male-Mail Survey: Gay Men Report on their Perceptions of Themselves" You are among a select group 0 gay men who have been randomly chosen to participate in this study. In order for th results to accurately reflect the thinking of gay men in this country. it is highly importan we have as many gay men participating in this study as possible. Your participation in thi study will be completely anonymous and will help provide information that can grea benefit our community. If you have already completed and returned the study. on beha of myself and our community I sincerelythank you for your time and help! If you have not completed the survey, please do so and return it today. If for some reason you did not receive the survey. or if you have any questions, please contact me at (303) 4914 7850. If you have chosen not to complete the survey. please complete the green colored sheet in your survey packet. Thank you very much for your cooperation in this study. Sincerely. Daniel W. Socall, M.A. Senior Staff Counselor 179 Appendix I Results Requested and/or Not Interested Posteard In order to know who has completed the survey. we ask that you please completJ this postcard and mail itseparately from your survey packet and blue consent form. This will keep your responses completely confidential. This postcard requires no postage if mailed in the United States. Name: Address: City, State. Zipcode: Check here if you would like a copy of the results sent to you Check here if you are not willing to complete the questionnaire Thank you for your cooperationll BIBLIOGRAPHY I80 BIBLIOGRAPHY Apfeldorf, M., & Hunley, PI. (1975). Applications of MMPI alcoholism scales to older alcoholics and problem drinkers. WM, 16, 645-653. Beaton, S., & Guild, N. (1976). Treatment for gay problem drinkers. SociaL 2mm 51, 302-308. Binder, IL. (1970). WWW charactenstxle. Unpublished doctoral dissertation, University of Michigan. Blascovitch, J., & Tomaka, J. (1991). Measurement of self-esteem. In J.P. Robinson, PR Shaver, & L.S. Wrightsman (Eds), Measumfl 031$.‘l-5 «I. 0.. 0‘ 100°4-.-HI'€-.< 0 U‘vri‘ 6.x“ .021 WW (pp. 121-123). San Diego, CA: Academic Press, Inc. Harcourt Brace Jovanovich. Bollen, K.A. (1989). Smrmmmafignmnmmm. New York: John Wiley & Sons. Bollen, K. A. (1990). Overall fit in covariance structural models: Two types of sample size effects. EsyghglggicaLBnflgfin, 191, 256-259. Borkenau, P., & Ostendorf, F: (1990). Comparing exploratory and confirmatory factor analysis: A study on the five-factor model of personality. Remnaflhmnd IndeualDiffeLenees 11. 515-524. Branden, N. (1983). Homfingjheself. Los Angeles: 1P. Tarcher. Bradshaw, J. (1988). WWW. Deerfield Beach, Florida: Health Communications, Inc. 181 Buss, A-H- (1980)-521W San Franciscoz Freeman. Campbell. RN. (1984). WWW Lanham, Maryland: University Press of America. Chang, SS. (1988). W W. Unpublished doctoral dissertation, Michigan State University, East Lansing, Michigan. Cheek, J.M. & Buss, AH. (1981). Shyness and sociability. W Watchman. i1. 330-339. Cohen, J. & Cohen, P. (1975). W W. New York: Wiley. Coleman, E. (1987). Assessment of sexual orientation. In E. Coleman (Ed), I l’l’fi 111°‘Ell . amachesfizumntinnflmflieing (pp. 9 - 24). New York: Harrington Park Press. Conger, A.J. (1974). A revised definition for suppressor variables: A guide to their identification and interpretation. EducafignaLandlsychglggigaL Measurement, 3.4.. 3546. Cook, D. R. (1987). Measuring shame: The internalized shame scale. Alcoholism Ireannemfluarterlx. A. 197-215. Cook, DR. (1988). Shame, attachment, and addictions: Implications for family therapists. WM 11. 405-419. 182 Cook, DR. (1988). WWW. Menomonie, WI: Channel Press. Coopersmith, S. (1967). W. San Francisco: Freeman. Crouvpen, GA. (1976). Eielddenmdenmndmndenmndmmmd 6‘ 1 : : 1] l” l o lo I E l e I .l I eggjmcfinning. Unpublished doctoral dissertation, California School of Professional Psychology. Crowne, D.P. & Marlowe, D. (1960). A new scale of social desirability independent of psychopathology. MW 24, 349-3 54. Church, A.T., & Burke, P.J. (1994). Exploratory and confirmatory test of the big five and Tellegen’s three and four-dimensional models. lonmaigfi Wand-519131 25191101081. 6.6. 93-114. Dawis, R V. (1987). Scale construction. Ioumalomounselinumholoa. 15. 481-489. DeVellis, R. (1991). Scale development: Theory and applications. Anniiedjncial WM (Vol. 26). Newburry Park, CA.: Sage. Dillman, D.A. (1978). W. New York: John Wiley & Sons. Dobson, C., Goudy, W.J., Keith, P.M., & Powers, E. (1979). Further analysis of Rosenberg's self-esteem scale. MW, 45, 639-641. Duckworth, J. & Anderson, W. (1986). W W. Muncie, IN: Accelerated Development, Inc. 183 Duckworth, J. (1983). IheMACjealezAnpenemhreemdies Paper presented to the 18th Annual Symposium on the MMPI, Tampa, Florida. Enya, & Ryan, N. (1991). Shenherdmmns [Compact Disc]. New York, New York: Reprise Records. F ifield, L., De Crescenzo, T.A., & Latham, J .D. (1975). mm mmunity. Summary of "On my way to nowhere: Alienated, isolated drunk; An analysis of gay alcohol abuse and a evaluation of alcoholism rehabilitation services for the Los-Angeles County." Los Angles: Gay Community Services Center. F inney, J ., Smith, D., Skeeter. D. & Avuenshire, C. (1971). MMPI alcoholism scale: Factor structure and content analysis. WW onAlmhol, 32. 1055-1060. Fleming, J.S., & Courtney, BE. (1984). The dimensionality of self-esteem. II. Hierarchical facet model for revised measurement scales. .[numamfl Whales! 46, 404-421. F ossum, M & Mason, M. (1986). W. New Yorlg New York: Norton & Co. Fox, R.J., Crask, MR, & Kim, J. (1988). Mail survey response rate: A meta- analysis of selected techniques for inducing response. Euhligflnininn. mm. 22. 467-491- Gay Council on Drinking Behavior. (1982). W andsmlmheliss. Washington, DC: Whitman-Walker Clinic. 184 Gelso, C. (1980). Research in counseling psychology: Methodological and professional issues. Cnunsnljng Emhnlngist, 8, 7-35. Glass, G.V., & Hopkins, K.D. (1984). WWW psynhnlngy. Englewood Cliffs, NJ: Prentice-Hall, Inc. Greene, R L. (1980). W Orlando, FL: Grune & Stratton, Inc. Hammersmith, S. (1987). A sociological approach to counseling homosexual clients and their families._1numal_nf_Hnmnnmnlity, .15, 173-190. Harder, D.W., & Lewis, SJ. (1987). The assessment of shame and guilt In J.N. Butcher & CD. Spielberger (Eds), WW. Vol. 6 (pp. 59-74). Hillsdale, NJ: Lawrence Erlbaum Associates. Hetrick, E.S., & Martin, AD. (1987). Developmental issues and their resolution for gay and lesbian adolescents. Jnumnlnfflnmnsnmnlim, 15, 25-43. Hoblitzelle, W. (1987). Differentiating and measuring shame and guilt: The relation between shame and depression. In H.B. Lewis (Ed), Ihnmlmfl shmnnmjymntnmhnnatinn. Hillsdale NJ: Lawrence Erlbaum Associates. Huber, N.A., & Danahy, S. (1975). Use of the MMPI in predicting completion and evaluation changes in a long-term alcoholism treatment program. Inumalnflfihrdimnflmhol. 3.6. 1230-1237. Israelstam, 8., & Lambert, S. (1983). Homosexuality as a cause of alcoholism: A historical review. The IntemationaLloumaLoftheAddictions 13. 1085- 1107. 185 Jackson, MR. (1984). W. Albany: State University of New York Press. Jacobson, E. ( 1964). W111. New York: International Universities Press. Kanuk, L., & Berenson, C. (1975). Mail surveys and response rates: A review. MW. 12, 440-453. Kaufman, G. (1974). The meaning of shame: Toward a self-affirming identity. IQumaLQLCnunselineRmholm, 21. 563-574- Kaufman, G. (1989). basedmdromes. New York: Springer. Kaufman, G. (1990. April). Mubhamsjmhedifi‘eremialmneminsnf. WW Paper Framed at the National Conference on Re-Visioning Knowledge and the Curriculum: Feminist Perspectives, Michigan State University, East Lansing, MI. Kaufman, G. (1992). W (3rd ed.). Cambridge, Massachusetts: Schenkman Books. Kaufman, G., & Raphael, L. (1991). Ihnnynammnfinnwnnfighfingshmand rebuildingnnifl-fimnm (2nd ed.). Cambridge, Massachusetts: Schenkrnan Books. chlinger. F.N., & Pedhazur, EJ- (1973). MummsressionjnhehaxioraL Wit. New York: Holt, Rinehart and Winston, Inc. 186 Kim, J .O., & Mueller, CW. (1978). Factor analysis. Sullivan, J .L., & Niemi, KG. (Ed-1). Quandtatimnnlicafiensjnlhemial sciences (Vol. 14) (pp. ). Beverly Hills, CA.: Sage. Korpi, DR. (1977). WWW nnlitinaljncninu. Unpublished doctoral dissertation, California School of Professional Psychology, San Francisco. Kranitz, L. (1972). Alcoholics, heroin addicts and nonaddicts: Comparisons on the MacAndrew alcoholism scale of the MMPI. Quanndyjnumalnfi W. 3.3.. 807-809. Kus, RJ. (1988). Alcoholism and non-acceptance of gay self: The critical link. loumaLofHomosexualin, 15. 25-41. Lachar, D., Berman, W., Grisell, J.L., & Schooff, K. (1976). The MacAndrew alcoholism scale as a general measure of substance misuse. Inuannfl Shrdiemulmhel, 31, 1609-1615. Lewes, K. (1988). Wmflnhnmmflim New York: Penguin Books. Lewis, H.B. (Ed), (1987). Ihnminnflshamnjnmmnmjnnnatinn. Hillsdale, New Jersey: Erlbaum. Lewis, M. (1992). W New York: Free Press - MacMillan. Lewis-Beck, MS. (1980). Applied regression: An introduction. Sullivan, J.L., & Niemi, KG. (EdS). WWW (V01- 22). Beverly Hills, CA.: Sage. 187 Lohrenz, L.J., Connelly, J.C., Coyne, L., & Spare, K.E. (1978). Alcohol problems in several midwestem communities. JnnannffimdiennnAinnhni, 32, 1959-1963. Long, 1.8. (1983). Confirmatory factor analysis. Sullivan, J.L., & Niemi, KG. (5618). WWW (Vol- 33)- Beverly Hills, CA: Sage. Lynd, H. (1958). W. New York: Harcourt, Brace. MacAndrew, C. (1983). Alcoholic personality or personalities: Scale and profile data from the MMPI. In W.M. Cox (Ed), Wang, W (pp. 73-85). San Francisco: Jossey-Bass Inc. MacAndrew, C. ( 1981). What the MAC scale tells us about men alcoholics. Ionmalnflfitudies onAlmhoL :12. 604-625. MacAndrew, C. (1979). MAC scale score of three samples of men under conditions of conventional versus independent scale administration. InumaLeLSmdim Alcohol. :10. 138-141. MacAndrew, C. (1965). The differentiation of male alcoholic outpatients from nonalcoholic outpatients by means of the MIW Smdiesnnflsohol. 2e, 238-246. McDonald, R.P., & Marsh, H.W. (1990). Choosing a multivariate model: Noncentrality and goodness of fit. Psychninginaifinflefin, 191, 247-255. 188 McWhirter, D.P. & Mattison, AM. (1984). WM signing. Englewood Cliffs, New Jersey: Prentice-Hall, Inc. Mirman. MC- (1984). smmmmmmmnmmm W. Unpublished doctoral dissertation. Michigan State University. Nardi, PM. (1982). Alcoholism and homosexuality: A theoretical perspective. MW. 1. 9-25. Nathanson. D- (1992).Shameand.nnde..Af£ecLsex_and_the.h1rth.o£the.aclf New York: Norton. Nathanson, D. (1987). The shame/pride axis. In H.B. Lewis (Ed), Ihnrninnfl WW (pp. 183-205). Hillsdale, New Jersey: Erlbaum. National Institute on Alcohol Abuse and Alcoholism [NIAAA]. (1978). IhinL WWWMW. Rockville. MD: US. Department of Health, Education, and Welfare. Negri. TF- (1978). W Unpublished doctoral dissertation, Michigan State University. Nunnally, J.C. (1978). EsychnmntrinIhnnndZnnm New York: McGraw- Hill. 189 Paulhus, D. (1991). Measurement and control of response bias. In J.P. Robinson, P.R. Shaver, & L.S. Wrightsman (Eds), MW e._ e gee! ._.qqe:.,. e u‘.__‘-,.e e . e ree' W (pp. 27-31). San Diego, CA: Academic Press, Inc. Harcourt Brace Jovanovich. Pelham, B.W., & Swann, W. (1989). From self-conceptions to self-worth: On the sources and structure of global self-esteem. .[numainffinrsnnalinmnd Sndallmholou. 51, 672-680. Perlman. M. (1953). WWW anxinty. Unpublished doctoral dissertation, University of Chicago. Piers, G., & Singer, M. (1953). Shamnnnnguin. New York: Norton. Pride Institute reports mortality due to substance abuse. (1989, July). 1111 Adm. p. 13. Reynolds, BS. (1991). The role of shame in anorexia nervosa and bulimia nervosa (Doctoral dissertation, Boston University, 1991). W Ahmmlmemarional. 5215.41). 2783. Reynolds, W.M. (1982). Development of reliable and valid short forms of the Marlowe-Crowne Social Desirability Scale. JnnmainleininaL Radiology. 38. 199-125. Rhodes, R.J. (1969). The Mac Andrew alcoholism scale: A replication. inumaL effilrmealfimhelogx, 25, 189-191. Rich, C .C., & Davis, HQ (1969). Concurrent validity of MMPI alcoholism scales. loumalef ClinicaLEsxchelm 25. 425-426. 190 Robinette, R.L. (1991). The relationship between the Marlowe-Crowne Form C and the validity scales of the MMPI. inummmfinmmhnlngy, 41, 396-399. Rogers, P. (1994, September 19). Surviving the second wave. Gay community deals with resurgence in HIV infection rate. Wk, 124, 50-52. Rohan, W.P., Tatro, R.L., & Rotman, SR. (1969). MMPI changes in alcoholics during hospitalization. MW 20, 389- 400. Rosenberg, M. (1965). WWW. Princeton, NJ: Princeton University Press. Rosenberg, M. (1979). annniyingjhnsnlf. New York: Basic Books. Ryback. C. J- (1991). Axalrdatreumdxeflhesmemahzedshamessale. Ihnnmtinainriginnandjmniinatinns. Unpublished masters thesis, Southern Illinois University. Saghir, M.T., & Robins, E. (1973). Mainnnnfimainhnmnmm mmhnnsimnynsjigatinn. Baltimore: Williams & Wilkins. Schwartz, B. (1984). W (2nd Edition). New York: W. W. Norton and Co. Shilts. R- (1993). W. New York: St. Martin’s Press. Smith. RL- (1972). Iherelatmenronenesmshammulumjndisammf. defensiynstyln Unpublished doctoral dissertation, Northwestern University. 191 Spielberger. CD. (1983).MmuaLfoLthestate-_1rartmxremmnm Palo Alto, CA: Consulting Psychologists Press. Staff. (1992, November 6). Amendment 2 Aftermath. W. p. 12. Strahan, R & Gerbasi, K. (1972). Short, homogenous version of the Marlowe- Crowne Social Desirability Scale. Jnumnlnfifliininaifisynhnlngy, 23, 191- 193. Tinsley, H., Bowman, S.L., & York, DC. (1989). Career decision scale, My vocational situation, Vocational rating scale, and Decision rating scale: Do they measure the same constructs? loumaLoLCounselinsmholm 35. 115-120. Tomkins, SS. (1963). WWW, Vol. 2. New York: Springer. Tomkins, SS. (1987). Shame. In D.L. Nathanson (Ed), WW: (pp. 102-25). New York: Guilford Press. Uecker, A.E. (1970). Differentiating male alcoholics from other psychiatric inpatients: Validity of the MacAndrew scale. WWW mm. 3.1, 379-383. US. Bureau of the Census. (1991). MW. Characteristics of the Population Series. Washington, DC: US. Government Printing Office. Vega, A. (1971). Cross-validation of four MMPI scales for alcoholism. Quarterly lournaleffihrdiemnAlcohol. 3.2. 791-797. 192 Waite-O’Brien. N. (1991). We alsnhnlmndnnnflnnhnhnmmflnflnns. Unpublished doctoral dissertation, United States International University, San Diego. Wang, D. (1992). The relationship between locus of control, depression, shame and self-esteem. thnnsnMnnmfinalmlnumal, 6, 207-210. Watson, D., Clark, L.A., & Tellegen, A. (1988). DeveIOpment and validation of brief measures of positive and negative affect: The PANAS Scales. WW, :4, 1063-1070. Watson, D., & Tellegen, A. (1985). Toward a consensual structure of mood. EsxcholoeisaLflulletin, a, 219-235. Weinberg, M.S., & Williams, C]. (1974). Malnhnmnsnxunisglhninnmbinmmd ananminns. New York: Oxford University Press. Wong, M. (1992). Shame and its contribution to PTSD. JnumaLan‘mumatin fitness, 5, 557-562. Wurrnser, L. (1981). Ihnmanknflshamn. Baltimore: The Johns Hopkins University Press. Yam, P. (1995, February). Dangerous sex: New signs of risk taking prompt rethinking about AIDS prevention. ngnntifinAmnrim 212, 10-12. Zevon, M.A., & Tellegen, A. (1982). The structure of mood change: An ideographic/nomothetic analysis. InnmaLnflEersnnamLandjnniaL Melon. 43., 111-122. 193 look A., & Sipps, G.L. (1985). Cross-validation of a short form of the Marlowe- Crowne Social Desirability Scale. LnumalnLClininaljmhnlngy, 4.1, 236-238. IGRN STRTE UNIV. llIIIllllllllllllllllmIll"WIHIIIIHIINIIIHIIIVHI