<%BANNER%>

Mediating Role of Openness in the Links of Right-Wing Authoritarianism, Social Dominance Orientation and Traditional Gen...


PAGE 1

MEDIATING ROLE OF OPENNESS IN THE LINKS OF RIGHT-WING AUTHORITARIANISM, SOCIAL DO MINANCE ORIENTATION, AND TRADITIONAL GENDER ROLE ATTIT UDES TO ANTILESBIAN AND GAY ATTITUDES AND BEHAVIORS By MELINDA BETH GOODMAN A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2005

PAGE 2

Copyright 2005 by Melinda Beth Goodman

PAGE 3

This thesis is dedicated to the good people at Starbucks who let me work there for many hours and whose chai tea helped me through th e process. In addition, this document is dedicated to my cat Calvin who demonstrated his love for me and th e literature as noted by the countless hours spent sitt ing on my lap and the Journal of Counseling Psychology.

PAGE 4

iv ACKNOWLEDGMENTS First of all I would like to thank my advisor and committee chair, Dr. Bonnie Moradi, for all her guidance, support, and en couragement during every step of the way. I am thankful for the assistance given to me by my committee members, Dr. Martin Heesacker and Dr. Kendal Broad, and by my re search assistants Ashley Stripling and Korinne Chiu. In addition, I would like to th ank my parents for all of their support during this project and my entire life. I could never have reached this point without their love, encouragement, and devotion. Finally, I would like to thank my friends from Gainesville and all of my family and friends from home who have continually supported me despite my move to Florida.

PAGE 5

v TABLE OF CONTENTS page ACKNOWLEDGMENTS.................................................................................................iv LIST OF TABLES............................................................................................................vii LIST OF FIGURES.........................................................................................................viii ABSTRACT....................................................................................................................... ix CHAPTER 1 INTRODUCTION........................................................................................................1 2 REVIEW OF LITERATURE.....................................................................................11 Right-Wing Authoritarianism.....................................................................................12 Social Dominance Orientation....................................................................................17 Traditional Gender Role Attitudes..............................................................................20 Openness to Experience: An Underlying Construct...................................................24 Need to Study Anti-Lesbian and Gay Behaviors........................................................30 Purpose of Study.........................................................................................................37 3 METHODS.................................................................................................................39 Participants.................................................................................................................39 Procedures...................................................................................................................40 Instruments.................................................................................................................40 Criterion Variables..............................................................................................40 Predictor Variables..............................................................................................43 4 RESULTS...................................................................................................................49 Gender Differences.....................................................................................................49 Descriptive Statistics..................................................................................................50 Test for Order Effects.................................................................................................52 Intercorrelations Among Variable s of Interest: Hypothesis 1....................................53 Tests of Mediation: Hypotheses 2 and 3.....................................................................54 5 DISCUSSION.............................................................................................................63

PAGE 6

vi LIST OF REFERENCES...................................................................................................74 BIOGRAPHICAL SKETCH.............................................................................................81

PAGE 7

vii LIST OF TABLES Table page 1 Zero-Order Correlations Among Vari ables of Interest for Women.........................49 2 Zero-Order Correlations Among Va riables of Interest for Men..............................49 3 Summary Statistics Among Variab les of Interest for Women.................................50 4 Summary Statistics Among Vari ables of Interest for Men......................................51

PAGE 8

viii LIST OF FIGURES Figure page 1 Fully saturated model depicting relations hips among variables of interest for women......................................................................................................................56 2 Trimmed model depicting re lationships among variables of interest for women....58 3 Fully saturated model depicting relations hips among variables of interest for men...........................................................................................................................6 0 4 Trimmed model depicting relationships among variable s of interest for men.........62

PAGE 9

ix Abstract of Thesis Presen ted to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science MEDIATING ROLE OF OPENNESS IN THE LINKS OF RIGHT-WING AUTHORITARIANISM, SOCIAL DO MINANCE ORIENTATION, AND TRADITIONAL GENDER ROLE ATTIT UDES TO ANTILESBIAN AND GAY ATTITUDES AND BEHAVIORS By Melinda Beth Goodman December 2005 Chair: Bonnie Moradi Major Department: Psychology The present study examined c oncomitantly the links of ri ght-wing authoritarianism (RWA), social dominance orientation (SDO), and traditional gender role attitudes to antilesbian and gay (LG) attitudes. In add ition, the present study examined openness to experience as a mediator that might expl ain, partially or fully, the links of the aforementioned correlates to an ti-LG attitudes. Finally, this study examined the extent to which RWA, SDO, traditional gender role at titudes, and openness ar e related to anti-LG behaviors directly or indirec tly through anti-LG attitudes. Results showed that RWA and traditional gender role attitudes were the str ongest predictors of anti-LG attitudes and behaviors. The role of openness as a medi ator of the relations of RWA, SDO, and traditional gender role attitudes to anti-LG attitudes was not supported. Finally, anti-LG attitudes did indeed partially mediate the relations of RWA, tr aditional gender role

PAGE 10

x attitudes, and SDO to anti-LG behaviors. Futu re directions for rese arch and implications for practice are discussed.

PAGE 11

1 CHAPTER 1 INTRODUCTION A broad range of variables (e.g., age, inco me, education, gender, and interpersonal contact) has been shown to co rrelate with anti-lesbian and gay (LG) attitudes (e.g., Herek & Capitanio, 1995, Johnson, Brems, & Alford-Keating, 1997). Within this body of literature, some of the most consistent correlates of anti-LG attitudes have been personality-related variables. More specifi cally, right-wing author itarianism (RWA), social dominance orientation (SDO), and trad itional gender role at titudes have been linked strongly and consistently to anti-LG attitudes. Although studies that focus on examining the link of each of these variab les and anti-LG attitudes (e.g., Altemeyer & Hunsberger, 1992; Basow & Johnson, 2000; Whitley & Egisdttir, 2000) are informative, theoretical and empirical scholar ship is needed to integrate findings across such studies and explain the shared links of personality-r elated variables to anti-LG attitudes. More specifically, a question that need s to be examined is what, if any, underlying construct or set of constructs explains the common links of RWA, SDO, and traditional gender role attitudes to anti-LG attitude s. Conceptual definitions of each of the aforementioned constructs and prior literatu re on their links to anti-LG attitudes provide a basis for exploring this question. Therefor e, a brief overview of the liter ature associated with each construct and its link to anti-LG attitudes is provided next. Altemeyer (1981) developed the concept of right-wing authoritarianism (RWA), based on social learning theory, to addr ess a number of limitations (e.g., lack of conceptual clarity and falsifiability and psychometric problems in operationalization) in

PAGE 12

2 Adorno, Frenkel-Brunswik, Levinson, and Sanford’s (1950) earlier psychodynamic conception of authoritarianism. Altemeyer (1981) defined RWA as comprised of three attitudinal clusters. The first component, au thoritarian submission, is defined as a “high degree of submission to authoriti es who are perceived to be established and legitimate in the society in which one lives” (p. 148) The second component is authoritarian aggression and reflects “general aggression, directed at various persons, which is perceived to be sanctioned by established au thorities” (p. 148). A ggressiveness in this context is the predisposition to do physical, psychological, economic, or social harm to others. Persons with high levels of RWA te nd to be aggressive or support aggressiveness toward targets that they believe are social ly stigmatized. For example, persons high in RWA might support police violen ce against suspects in order to obtain confessions. The final component of RWA is conventionalism defined as a “high degree of adherence to the social conventions which are perceived to be endorsed by societ y and its established authorities” (p. 148). Altemeyer (1981) describe d “established and legitimate authorities” to include individuals in society who are usually considered to be legal or moral authorities such as parents for younger persons, religious officials, heads of government, police officers, judges, a nd military officials. Altemeyer’s (1981, 1988) theoretical conceptualization of right-wing authoritarianism as characterized by high degrees of submission to authorities, aggressiveness toward out-groups, and adhe rence to conventions perceived to be endorsed by society and its authorities, suggest s that people with hi gh levels of RWA are likely to exhibit anti-LG attitudes. More specifically, LG persons (a) are explicitly condemned by some religious and political leaders, making LG persons an acceptable

PAGE 13

3 target for prejudice and hostility, (b) represen t a socially stigmatized out-group relative to heterosexual women and men, and (c) are percei ved as advocates for political change and challenges to social conventions of traditiona l gender roles. Indeed, a number of studies have found significant positive links between RWA and anti-LG attitudes (e.g., Altemeyer, 1981, 1998, 2001; Basow & Johnson, 2000; Whitley & Lee, 2000) indicating that persons with high levels of RWA express more anti-L G attitudes than do persons with low levels of RWA. In addition to RWA, social dominance or ientation (SDO) also has been linked consistently with anti-LG attitudes. Pra tto, Sidanius, Stallwor th, and Malle (1994) defined SDO as the “extent to which one de sires that one’s in-group dominate and be superior to out-groups” (p 724). SDO taps a general attitudinal orientation toward intergroup relations that refl ects a preference for hierarchi cal rather than egalitarian relations. More specifically, the desire to main tain the superior position of their in-groups relative to out-groups motivates people hi gh in SDO to accept hierarchy-legitimizing myths that denigrate members of out-groups and enforce the status quo of their ingroups’ power position. Thus, persons high in SDO tend to have anti-LG attitudes because LG persons have lower perceived soci al status than do heterosexuals (Whitley & Lee, 2000). The empirical litera ture supports the expected link between SDO and anti-LG attitudes. Across a number of studies, signi ficant correlations have been found between SDO and anti-LG attitudes (e.g., Altemeyer, 1998; Whitley, 1999; Whitley & Lee 2000) such that higher SDO scores are related to higher levels of anti-LG attitudes. Although RWA and SDO both include an element of prejudice toward out-groups, Altemeyer (1988) argued that the two constructs are distinct in that SDO “does not have

PAGE 14

4 the same psychological roots that previous studies have unearthed in right-wing authoritarians” (p. 61). More specifically, a person high in RWA is very accepting of traditional values and institutional authorities and is likely to follow the instructions of those in positions of power and influence. By contrast, an individual high in SDO is not motivated by a sense of duty or morality (Heaven & Bucci, 2001). Thus, although RWA and SDO both can explain prejudicial beliefs and behaviors, such as anti-LG attitudes, they do so in a complementary but not re dundant fashion, such that each accounts for unique variance in a variety of prejudice-relate d beliefs, behaviors, a nd traits (Altemeyer, 1998). Indeed, empirical evidence indicates that RWA and SDO are only minimally correlated with each other (Altemeyer, 1998; Pr atto et al., 1994) and the relationship of SDO to prejudice is not accounted for by a joint relationship with RWA (Altemeyer, 1988; Whitley & Lee, 2000). The final personality-related construct that has been linked consistently to anti-LG attitudes is traditional gender ro le attitudes. Gender role at titudes are “beliefs about the appropriate role activities for women and men” (McHugh & Frieze, 1997, p. 4). Individuals who endorse traditional gender roles believe that women should follow traditional female-role norms and that men should follow traditional male-role norms. A commonly held stereotype about LG persons is that such persons violate traditional gender roles, with gay men acting more feminine and lesbian women acting more masculine than heterosexual men and wome n should, respectively (Kite & Deaux, 1987). In addition, LG persons often ar e perceived to deviate from the cultural exp ectations of the roles that men and women should perform in their lives, for example, by not getting married and having children (Kite, 1984; Whitley, 1987). Thus, heterosexual persons

PAGE 15

5 who hold traditional gender role attitudes are likely to express anti-LG attitudes because LG persons are stereotyped as having cr oss-gender traits, roles, and physical characteristics. Indeed numerous studies m easuring constructs re lated to traditional gender role attitudes (e.g., Attitudes Toward Women Scale, Sex Role Egalitarianism Scale) have found that persons who hold more traditional gender ro le attitudes express more anti-LG attitudes than do those with le ss traditional gender role attitudes (Basow & Johnson, 2000; Herek, 1988; Whitley & Egisdttir, 2000). Thus, conceptual definitions of RWA, SD O, and traditional gender role attitudes suggest that each of these vari ables is related to anti-LG at titudes and empirical evidence supports such links. In addition to their shar ed links to anti-LG attitudes, RWA, SDO, and traditional gender role attitudes each also has b een linked conceptually and empirically to openness to experience, a dimension of personality described in the Five Factor Model. Openness, in turn, has been show n to be related to anti-LG attitudes. Thus, openness to experience might be a mediating va riable that provides a more parsimonious explanation of and accounts for the links of RWA, SDO, and traditional gender role attitudes to anti-LG attitudes. Within the framework of the Five Fa ctor Model, openness to experience is conceptualized to reflect activ e imagination, aesthetic sensi tivity, receptiveness to inner feelings, preference for variety, intellectua l curiosity, and independence of judgment (McCrae & Costa, 2003). Additi onally, lack of openness to experience characteristically involves intolerance for ambiguity and lack of flexibility. A person who is high in openness is likely to value intellectual matters, is nonconf orming, introspective, has a vivid imagination, and admits that what is right and wrong for one person may not be

PAGE 16

6 applicable in other circumstances. On the ot her hand, a person scoring low in openness is likely to be rigid and uncomfortable with co mplexities, have a narrow range of emotions, favor conservative values, and judge others in conventional terms (McCrae & Costa, 2003). The characteristics associated with low levels of openness appear to overlap with the characteristics associated with high levels of RWA, SDO, and tr aditional gender role attitudes. For example, persons high in RWA ar e expected to reject any thoughts or experiences that are not sa nctioned by authority figures demonstrating a lack of flexibility. Characteristica lly, people high in RWA are rigid, conforming, and have conservative political values (Altemeyer, 1981, 1988). Indeed, RWA has been shown to relate significantly and negatively to openness to experience with correlations ranging from -.30 to -.57 (Altemeyer, 1996; Butle r, 2000; Heaven & Bucci, 2001; Peterson, Smirles, & Wentworth, 1997). Similarly, a lack of openness to experience is apparent in SDO. Persons high in SDO accept hierarchy-legitimizing myths in order to promote the superiority of one group over another (Pratto et al., 1994). Instead of viewing other people as individuals, persons high in SDO concentrate on the ta rgets’ out-group membership. Those high on SDO characteristically have a lack of symp athy, empathy, and tolerance. The lack of ability to view the complexities in the wo rld and the narrow range of emotions that characterize persons high in so cial dominance are consistent with theoretical definitions of lack of openness to experien ce. Indeed, SDO has been show n to correlate significantly and negatively (r = -.26 to -.28) with openness to experience (Heaven & Bucci, 2001; Pratto et al., 1994).

PAGE 17

7 Finally, persons who adhere to traditional ge nder role attitudes expect others to fit into a relatively stable set of gender roles, traits, and physical attributes (e.g., Kite & Whitley, 1998). This acceptance of traditional roles demonstrates a lack of tolerance for ambiguity and variety within the gender spectr um. It is expected that persons who adhere to traditional gender role att itudes would also be low in openness to experience. Although the links between openness and adherence to traditional masculine and feminine gender role norms have not been examined directly, extant data indicate a negative link between constructs that are related to traditional ge nder role ideology and openness. For example, among men, openness to experience was found to be significantly and negatively related to masculine gender role conflict and masculin e gender role stress with correlations of .38 and -.21, respectively (Tokar, Fischer, Schaub, & Moradi, 2000). Gender role conflict is a “psychological state in which gender role s have negative consequences or impact on the person or others” (O’Neil, 1981, p. 203). Men’s gender role stress results from a failure to meet cultural standards of masculinity (Eisler & Blalock, 1991). Although men’s gender role conflict and stress are not equivalent to traditional gender role attitudes, there is significant theoretical overlap in that one must endorse traditional gender roles in order to expe rience gender role conflict or stress. Indeed, gender role conflict and stress are significantly correlated with traditional gender role attitudes (Thompson, 1991, as cited in Thompson, Pleck, & Ferrera, 1992). Thus, it is expected that traditional gender role at titudes will relate to low ope nness and the present study will examine directly this proposition. Taken together, the theoretical definitions and empirical evidence reviewed suggest that lack of openness to experience might be a global construct that ti es together the links

PAGE 18

8 of RWA, SDO, and traditional gender role attitudes to anti-LG attitudes. Furthermore, openness might be an underlying construct that e xplains, partially or fully, the links of the aforementioned correlates to anti-LG attit udes. In fact, in ad dition to openness to experience’s links with RWA, SDO, and traditional gender attitudes, openness is also significantly related to anti-LG attitudes. When examining the Five Factor Model of personality, Cullen, Wright, and Alessandri (2002) found that openness accounted for 25% of the variance in anti-LG attitudes, and none of the ot her Five Factor Personality traits were unique predictors. Due to th e correlations of openness to the predictor variables (i.e., RWA, SDO, trad itional gender role attitudes) and anti-LG attitudes, it is possible that openness will mediate the links of RWA, SDO, and traditional gender role attitudes to anti-LG attitudes. In othe r words, RWA, SDO, and traditional gender attitudes might be related to anti-LG att itudes through their shar ed link with openness. Thus, the first general aim of the present st udy is to examine openness to experience as a mediator of the links of RWA, SDO, and tr aditional gender role attitudes to anti-LG attitudes. Studying anti-LG attitudes alone, however, is not sufficient. An implicit reason for studying anti-LG attitudes is th at the expression of these at titudes as anti-LG behaviors might have serious negative interpersonal a nd intrapersonal conseque nces for LG persons and society at large. In fact, the preval ence of anti-LG behaviors (e.g., name-calling, threats, assaults) is staggering. D’Augelli (1989) found that 50% of the gay and lesbian college students surveyed reported ha ving overheard disparaging comments, 26% reported having experienced personal verb al insults, 26% reported having been threatened with physical violence and 23% reported having been victims of assault.

PAGE 19

9 Furthermore, such negative experiences have been linked to mental health-related concerns (e.g., suicidal ideation, psychologi cal distress) among LG persons (e.g., Diaz, Ayala, Bein, Henne, & Marin, 2001; DiPlac ido, 1998; Mays & Chochran, 2001). Thus, there is a need to study anti-LG behavi ors in addition to anti-LG attitudes. Despite the considerable pr evalence and potential negati ve consequences of antiLG behaviors and numerous calls for research examining such behaviors (Herek, 2000; Kite & Whitley, 1998; Whitley, 2001), the majority of extant research has not attended to predictors or correlates of anti-LG beha viors and has focused mostly on identifying correlates of anti-LG attitudes. The limited research available s hows that although hate crimes and anti-LG behaviors are inevitably influenced by complex situational factors (Franklin, 1998), anti-LG attitudes do indeed contribute to anti-LG behaviors. In three studies that assessed self-reports of both anti-LG attitudes and behaviors, anti-LG attitudes were a significant pr edictor of negative behaviors. Correlations between anti-LG attitudes and behaviors toward LG persons ranged between .25 and .40 (Franklin, 2000; Patel, Long, McCammon, & Wuensch, 1995; Whitley, 2001). Research also demonstrates that some of the major correlates of anti-LG attitudes are also related to anti-LG behaviors. For example, Haddock, Zanna, and Esses (1993) found that for college students, RWA is relate d to support of increased budget cuts to an LG student organization. Furthermore, Fr anklin (2000) and Whitley (2001) found that endorsement of traditional gender role at titudes was related positively to anti-LG behaviors. However, there is no research examining the correlations of SDO and openness to experience to anti-LG behaviors. Th us, the second general aim of the present study is to add to the literat ure by examining the extent to which psychological factors

PAGE 20

10 that have been shown to be related to an ti-LG attitudes are also related to anti-LG behaviors. Also, it is not know n if RWA, SDO, traditional gender roles, and openness to experience will relate directly to anti-LG behaviors, or if their relationship is mediated by anti-LG attitudes. Thus, the present study w ill examine whether anti-LG attitudes mediate the relationships of RWA, SDO, and traditional gender roles and openness to experience to anti-LG attitudes. In light of the conceptual and empirical l iterature reviewed he re, the aims of the present study are to address the following gaps: 1. Examine openness to experience as a medi ator of the links of RWA, SDO, and traditional gender role attitudes to anti-LG attitudes. 2. Examine anti-LG attitudes as a mediator of the links of RWA, SDO, traditional gender role attitudes, and ope nness to anti-LG behaviors.

PAGE 21

11 CHAPTER 2 REVIEW OF LITERATURE A broad range of variables (e.g., age, inco me, education, gender, and interpersonal contact) has been shown to co rrelate with anti-lesbian and gay (LG) attitudes (e.g., Herek & Capitanio, 1995; Johnson, Brems, & Al ford-Keating, 1997). Within this body of literature, some of the most consistent correlates of anti-LG attitudes have been personality-related variables. More specifi cally, right-wing author itarianism (RWA), social dominance orientation (SDO), and trad itional gender role at titudes have been linked most strongly and consistently to anti -LG attitudes. Although studies that focus on examining these relationships (e.g., Alte meyer & Hunsberger, 1992; Basow & Johnson, 2000; Whitley & Egisdttir, 2000) are informativ e, theoretical and empirical scholarship is needed to integrate these findings and expl ain the shared links of personality-related variables to anti-LG attitudes. More specifica lly, a question that needs to be examined is what, if any, underlying construc t or set of constructs explains the common links of RWA, SDO, and traditiona l gender role attitudes to anti-LG attitudes. In order to provide the groundwork for expl oring this question, the present chapter will review the extant literature on RWA, S DO, and traditional gender role attitudes and present a fourth variable, openness to expe rience, that might provide a potentially integrative and parsimonious explanation of th e links of the aforementioned constructs to anti-LG attitudes. The first three sections of the present chapter reflec t critical reviews of the literature linking RWA, SDO, and traditiona l gender roles to anti-LG attitudes. In the fourth section, literature will be reviewed th at points to openness to experience as an

PAGE 22

12 underlying construct that might explain the li nks of the aforementi oned variables to antiLG attitudes. Next, literature on the links of anti-LG attitudes and their predictors to antiLG behaviors will be described. Finally, a conceptual model and the hypotheses that are examined in the present study will be presented. Right-Wing Authoritarianism Adorno, Frenkel-Brunswik, Levinson, and Sanford (1950) first defined the authoritarian personality as a way to explain the rise of Fascism in the 1930s. The authoritarian personality style was conceptualized to include nine traits: Conventionalism, Authoritarian Submission, Authoritarian Aggression, Anti-intraception, Superstition and Stereotypy, Power and T oughness, Destructiveness and Cynicism, Projectivity, and exaggerated concern with sexual goings-on (Adorno et al., 1950). Authoritarianism highlighted the role of psycho-dynamic influences on the individual that, it was believed, predisposed one to become prejudiced. More spec ifically, Adorno et al. (1950) theorize that threatening, forbi dding, and status-consci ous parents who used harsh and arbitrary punishment raised au thoritarian children. According to this conceptualization, children repr ess hostility toward their pa rents and instead overglorify their parents and act submissively toward th em. Consequently, these children’s repressed aggression toward their parents b ecomes displaced onto out-groups. Authoritarianism became a popular variable to study, and a considerable amount of research was conducted using the F scale, the questionnaire developed by Adorno et al. (1950) to assess this construct. However, numerous problems were identified with this theory of authoritarianism and the F scale. First, definitions of the traits of authoritarianism were thought to be uncle ar, lack distincti on, and reflect some redundancy (Altemeyer, 1981). Furthermore, the theory was criticized for lacking

PAGE 23

13 testability and falsifiability. For example, if authoritarians report th at their parents were especially harsh, then this s upports Adorno et al.’s (1950) theo ry. If authoritarians report that their parents were not harsh, however, it can be claimed that they are overglorifing their parents, still supporting the theory. Lastly, the psychomet ric properties of scores on the F scale have been questioned (Altemeyer 1981). For example, there are no reverse scored items, making F s cale scores susceptible to res ponse set bias. Also many items were intended to tap several of the nine tra its at once, adding empi rical ambiguity to the distinctiveness of the nine traits. To address some of the limitations of Adorno et al.’s (1950) model, Robert Altemeyer (1981, 1988, 1996) developed the cons truct of right-wing authoritarianism (RWA) based on social learning theory. Alte meyer (1981) defined RWA as comprised of three attitudinal clusters. Th e first component, authoritaria n submission, is defined as a “high degree of submission to authorities who are perceived to be established and legitimate in the society in which one lives” (p. 148). The second component is authoritarian aggression and refl ects a “general aggression, di rected at various persons, which is perceived to be sanctioned by esta blished authorities” ( p. 148). Such aggression is manifested as the predisposition to inflic t physical, psychological, economic, or social harm onto targets perceived as socially s tigmatized. The final component of RWA is conventionalism defined as a “high degree of adherence to the social conventions which are perceived to be endorsed by society and its established authorities” (p. 148). Established and legitimate author ities could include parents, re ligious officials, heads of government, police officers, judges, and m ilitary officials, and other persons or institutions perceived as legal or moral authorities (Altemeyer, 1981).

PAGE 24

14 In contrast to Adorno et al.’s (1950) model of author itarianism, Altemeyer’s (1981, 1988) theory of RWA relies on a social learni ng explanation of development. According to this conceptualization, attitudes are shaped by other people, through direct teaching and modeling, and through direct and vicarious experiences with th e objects of these attitudes. Thus, a rightwi ng authoritarian personality can be shaped by the child’s parents, religious leaders, peers, teachers, and through ot her models in the community. Additionally, watching television, reading books and hearing the news can shape the person’s attitudes. Theoretical conceptualization of RWA sugge sts that persons with high levels of RWA are likely to exhibit anti-LG attitudes because LG persons (a) are explicitly condemned by some religious and political leaders, making LG persons an acceptable target for prejudice and hostility, (b) represen t a socially stigmatized out-group relative to heterosexual women and men, and (c) are percei ved as advocates for political change and challenges to social conventions of traditi onal gender roles. Indeed, Robert Altemeyer has conducted extensive research demons trating a link between RWA and anti-LG attitudes. Altemeyer has measured the relationship between RWA and anti-LG attitudes with thousands of participants using his Right -Wing Authoritarianism (RWA) scale and Attitudes toward Homosexuals (ATH) scale (1988). The ATH is a balanced 12-item, Likert-type scale with scores ranging from –4 to +4, and higher scores indicating more anti-LG attitudes. ATH scores have been dem onstrated to be relatively unidimensional, with average intercorrelations among items at about .45, pr oducing an alpha internal consistency reliability estimate of about .90 (Altemeyer, 1988). ATH scores assess

PAGE 25

15 cognitive, affective, and beha vioral components of anti-LG attitudes. Sample items include, “I find the thought of homosexual acts disgusting” an d “I won’t associate with a known homosexual if I can help it.” Every year since 1984, Altemeyer has us ed the RWA and ATH scales to assess attitudes of University of Manitoba Intr oductory Psychology students and their parents. Across these studies, Altemeyer has found a c onsistent significant relationship between RWA and anti-LG attitudes with correl ations in the .50-.60 range (Altemeyer 1988, 1996, 2001). Based on this research, Altemeyer cl aimed that RWA “may explain hostility toward gays and lesbians be tter than any other personal ity variables” (1996, p. 26). The research program of Robert Alteme yer has yielded insight into attitudes toward LG persons. However, there are pa rticular limitations to these studies. For instance, Altemeyer sampled primarily Canadian introductory psychology students and their parents. Findings based on this populat ion may not generalize to Canadian and North America populations as a whole. In fact, Altemeyer reported that Manitoba, Canada has recently elected an openly ga y mayor (2001), suggesting that attitudes towards lesbian and gay persons in Manitoba might be more positive that other areas. Another limitation of Alte meyer’s (1981, 1988, 1996, 2001) studies is that he uses the ATH scale to assess anti-LG attitudes. This scale refers to “homosexuals” rather than to “lesbians” or “gay men,” leaving open the possi bility that participants may consider their attitudes toward only gay men when re sponding to ATH items (e.g., Haddock, Zanna, and Esses, 1993). In addition, one questi on explicitly refers to “gay men” only. Therefore, it is difficult to generalize result s based on this scale to attitudes toward lesbians.

PAGE 26

16 Due to the lack of research on attitude s toward lesbians, Basow and Johnson (2000) specifically studied the atti tudes of 71 undergraduate women toward lesbians. Using Herek’s (1988) Attitudes Toward Lesbians scale, they found that RWA correlated strongly with negative attitudes at r = .79, p < .001. Furthermore, a multiple regression analysis revealed that RWA alone predicte d 62% of the variance in attitudes toward lesbians. This study demonstrated that RWA is an important predictor of attitudes toward lesbian women. Other studies also have found similarl y strong relationships between RWA and anti-LG attitudes. Whitely and Lee (2000) c onducted a meta-analysis that examined the relationships of authoritarianism, dogmatism, SDO, and political-economic conservatism to anti-LG attitudes. They analyzed the re sults of 36 studies that included a total of 16,218 research participants and provided 51 effect sizes. For aut horitarianism, Whitley and Lee found a mean d of –1.21 and an equivalent r of -.52, p < .001. The d statistic represents the standardized difference in mean anti-LG attitude scores between participants scoring high and low on th e predictor variable. These values for d and equivalent r are well above Cohen’s (1988) criter ia for large effect sizes (i.e., d = 0.8; r = 0.5) Furthermore, Whitley and Lee (2000) f ound that authoritarianism had a stronger association with anti-LG attitudes wh en measured with the RWA scale, d = -1.19, r = .51, p < .001, than with the F scale (based on Adorno et al.’s psychodynamic conceptualization of authoritarianism) d = -0.96, r = -.43, p < .001. The larger effect sizes found with the RWA scale could be due to the relatively stronger psychometric properties of this scale compared to that of th e F scale. In addition, a ll of the studies using

PAGE 27

17 Altemeyer’s (1998) RWA scale used Canadian research participants, whereas the studies using Adorno et al.’s (1950) F scale used United States participants. Therefore, differences in mean effect sizes found for th e two scales could also represent nationality differences. On the other hand, there was no significant difference in scores based on whether the ATH, the ATLG, or other measures of anti-LG attitude s was used. Overall, Whitley and Lee’s meta-analysis demonstrates that RWA is indeed related to anti-LG attitudes. Social Dominance Orientation In addition to right-wing authoritarianis m (RWA), social dominance orientation (SDO) has also been linked consistently with prejudice and specifically with anti-lesbian and gay (LG) attitudes. Pratto, Sidanius, St allworth, and Malle (1994) defined SDO as the “extent to which one desires that one’s in-group dominate and be superior to outgroups” (p 724). The desire to maintain the su perior position of thei r in-groups motivates people high in SDO to denigrate members of out-groups, to oppose equality-enhancing social programs such as affirmative actions, and to discriminate against members of outgroups in order to enforce the status quo of the in-groups’ power position. Thus, persons high in SDO tend to have anti-LG attitudes b ecause LG persons have lower social status than do heterosexuals (Whitley & Lee, 2000) and because LG persons are perceived as advocates for social equality (Whitley, 1999). Empirical research suppor ts the proposed relationship between SDO and anti-LG attitudes. In addition to examining RWA, th e meta-analysis described earlier by Whitley and Lee (2000) also examined the relati onship between SDO and anti-LG attitudes. Based on 9 effect sizes, Whitley and Lee found th at SDO was related to anti-LG attitudes with a mean d = -.76 and an equivalent r = -.35, p < .001, indicating effect sizes within

PAGE 28

18 the medium to large range (C ohen, 1988). Effect sizes did not vary significantly whether SDO was assessed with Pratto et al.’s (1994) SDO scale or a measure developed by Altemeyer (1988). Results of Whitley and Lee’ s meta-analysis demonstrate that SDO, no matter which measure is used to assess it, is an important correlate of anti-LG attitudes. Although RWA and SDO are both related to anti-LG attitudes, Altemeyer (1988) argued that the two constructs are distinct in that SDO “does not have the same psychological roots that previous studies have unearthed in right-wing authoritarians” (p. 61). RWA is conceptualized as an intragroup phenomenon, whereas SDO is conceptualized as an intergroup phenomenon (Altemeyer, 1998; Pratto et al., 1994). More specifically, RWA focuses on submission to in-group authority figures independent of whether they advocate intergroup domi nance, whereas SDO focuses on dominance over out-groups independent of the views of the in-group authorities. As such, a person high in RWA is very accepting of traditional values and institutional authorities and is likely to follow the instructions of those in positions of power and influence. By contrast, an individual high in SDO is not motivated by a sense of duty to authorities or morality (Heaven & Bucci, 2001). Thus although RWA and SDO freque ntly explain prejudicial beliefs and behaviors, such as anti-LG attit udes, they do so in a complementary but not redundant fashion. such that each accounts for unique variance in a variety of beliefs, behaviors, and traits (Altemeyer, 1998). Indeed, empirical evidence indicates that RWA and SDO are only minimally correlated (Altemeyer, 1998; Pratto et al., 1994 ) and the relationship of SDO to prejudice is not accounted for by a joint relationshi p with RWA. For example, even after controlling for the effects of RWA, SDO still explained variance in ethnocentrism,

PAGE 29

19 racism, and nationalism (Pratto et al. 1994). Furthermore, acro ss several studies reviewed and conducted by Altemeyer (1998), RWA a nd SDO were only minimally correlated ( r = .22), yet when entered into a regression e quation with numerous other personality variables, together accounted for over 50% of the variance in a generalized prejudice index (comprised of prejudice against Afri can Americans, women, and LG persons). Altemeyer (1998) stated that RWA and SDO each “explained different segments of the overall prejudice,” implying that both RW A and SDO accounted for unique variance in the equation, yet he did not describe the specif ic data that served as the basis for this statement. Whitley and Lee (2000)’s findings, however, provided more detail and were consistent with Altemeyer’s (1998) conclu sion that RWA and SDO each accounted for unique variance in anti-LG attitudes. More specifically, Whitley and Lee conducted a follow up study to the meta-analysis described above, in which they surveyed 316 Introductory Psychology students. The partic ipants completed Herek’s (1988) Attitudes Toward Lesbians and Gay Men (ATLG) s cale, Altemeyer’s (1998) RWA scale, Rokeach’s (1956) Dogmatism scale, Pratto et al.’s (1994) SDO scale, and Pratto et al.’s (1994) three-item scale for political-econom ic conservatism. These authors found that RWA and SDO were only slightly correlated with each other ( r = .28). Of all the variables studied, RWA correlate d most strongly with negative attitudes toward lesbians ( r = -.58) and negative attitudes toward gay men ( r = -.54). SDO was the second strongest correlate of negative attitudes toward lesbians ( r = -.39) and negative attitudes toward gay men ( r = -.44) (all p s < .001). Furthermore, with RWA and SDO controlled as covariates, political-economic conservatism di d not significantly add any unique variance

PAGE 30

20 to attitudes toward lesbians and gay men. Additionally, with RWA and SDO controlled as covariates, dogmatism added less than 2% unique variance to attitudes toward lesbians, F (1, 312) = 8.29, p < .01, and less than 1% of the variance in attitudes toward gay men, F (1, 312) = 4.46, p < .05. In another equation that included all four attitudinal variables as predictors, RWA and SDO each accounted for unique variance in attitudes toward lesbians ( = -.42 and = -.18) and gay men ( = -.43 and = -.18). Taken together, these results suggest that RWA a nd SDO are two primary correlates of anti-LG attitudes. Each variable was related to anti-L G attitudes with the other controlled and they jointly accounted for most of the relationships of dogmatism and conservatism to anti-LG attitudes. Traditional Gender Role Attitudes A third personality-related construct that has been demonstrated consistently to have a significant relationship with anti-le sbian and gay (LG) att itudes is traditional gender role attitudes. Gender role attitudes are “beliefs about the appropriate role activities for women and men” (McHugh & Frieze, 1997, p. 4). Those endorsing traditional gender roles have the attitude that women should abide by traditional femalerole norms and that men should abide by traditional male-role norms. Heterosexual persons who endorse traditional gender roles hold anti-LG attitudes because LG persons are stereotypically perceived as having cr oss-gender traits, roles, and physical characteristics. Empirical evidence suggests that a commonl y held stereotype is that LG persons violate traditional gender roles, with gay men acting more feminine and lesbian women acting more masculine than heterosexual me n and women should, respectively. Kite and Deaux (1987) examined gender stereotypes a bout LG persons with 206 participants from

PAGE 31

21 introductory psychology courses. Partic ipants rated lesbian women, gay men, heterosexual women, and heterosexual me n on how likely each group was to possess masculine and feminine traits (e.g., self-conf ident, kind), role be haviors (e.g., financial provider, cooks the meals), physical char acteristics (e.g., sturdy, graceful), and occupations (e.g., construction worker, telephon e operator). Results were analyzed using multivariate analysis of variance and a signi ficant interaction was found between sexual orientation of target and sex of target, multivariate F (8, 182) = 20.60, p < .001. This interaction demonstrated that participants accepted the implicit gender inversion theory for LG persons in that gay men were rate d lower on masculine qua lities, yet higher on feminine qualities than were heterosexual me n, and that lesbian women were rated lower on feminine qualities, yet higher on masculin e qualities than were heterosexual women. This was evident in all categories (i.e., tra its, role behaviors, phys ical characteristics, occupations). These findings demonstrate that college students ster eotypically describe LG persons in cross-gender ways. In general, people who act in ways c onsistent with the opposite gender are not viewed positively (e.g., Laner & Laner, 1979, 1980). Therefore, heterosexual persons might hold anti-LG attitudes because LG persons are, as previously shown, stereotypically viewed as deviating from the cultural expectations of the roles that men and women should perform in their lives. Thes e anti-LG attitudes should be particularly strong among people who hold traditional gend er role attitudes because homosexuality poses a threat to their worl d view that places “masculi ne” and “feminine” at opposite ends of a bipolar dimension (Kite & Whitle y, 1998) and stereotypical views of LG persons place them at inappropriate ends of the gender dichotomy.

PAGE 32

22 Indeed, research using various indicators of traditional gender role attitudes has found significant correlations with anti-LG attitudes. Whitley (2001) conducted a metaanalysis to examine the relationship between traditional gender role attitudes and anti-LG attitudes. This meta-analysis included 42 studies with a total of 9629 re search participants that yielded 50 effect sizes for the relationship between gender-role beliefs and attitudes toward LG persons. Whitley (2001) included se ven types of measures for gender role attitudes: Attitudes Toward Women Scale (ATW; Spence & Helmreich, 1978), Attitudes Toward Feminism (FEM) scale (Smith, Ferree, & Miller, 1975), a measure of attitudes toward sexual equality (MacDonald, 1974), measur es of old-fashioned sexism (i.e., belief in traditional gender stereot ypes or approval of unequal tr eatment of men and women), measures of endorsement of traditional ma le role norms, and measures that assessed attitudes toward both male and female roles. The mean overall effect size indicated that people who endorsed traditional gender role attitudes held more anti-LG attitudes, d = 1.11, r = -.48, Z = -102.03, p < .001, QW (49) = 1041.87, p < .001, with effect size in the large range (Cohen, 1988). Whitley (2001) also compared the magnitude of the anti-LG attitudes-gender role ideology link across different op erationalizations of traditi onal gender role attitudes. These comparisons revealed that endorsement of traditional gender roles and gender hierarchies, rather than awareness of and sensitivity to gender discrimination or genderrelated self-concept, are key co rrelates of anti-LG attitudes. More specifically, the mean effect size for studies that a ssessed endorsement of traditiona l gender hierarchies (i.e., old fashioned sexism) was similar to that of studi es that assessed endorsement of male role norms, but was significantly greater than that of studies that a ssessed perceptions of

PAGE 33

23 discrimination against women (i.e., modern sexism) (1) = 119.11, p = .001. With regard to gender-related self -concept, the mean effect size for the relationship of femininity/expressiveness to anti-LG attitude s was not statistically significant. Although the mean effect size of the relationship be tween masculinity/instrumentality and anti-LG attitudes was significant, it wa s quite small (Cohen, 1988) and th erefore of little practical significance, d = .07, r = .03, Z = 3.83, p < .001, QW (19) = 179.62, p < .001. Overall, these results indicate that modern sexism and gender-ro le self-concept are not closely linked to anti-LG attitudes. On th e other hand, traditional gender role attitudes defined to “represent people’s ideas of th e proper roles for men and women in society and of behavioral norms for men and wo men” (Whitley, 2001, p. 692) and typically assessed with measures such as the Masc uline Role Norms Scale (Thompson & Pleck, 1986) and the ATW (Spence & Helmreich, 1978) are closely linked with anti-LG attitudes. Indeed, Whitley’s (2001) results indi cated that traditiona l gender role attitudes accounted for a substantial (23%) amount of variance in anti-LG attitudes. The magnitude of this relationship was similar for attitudes toward lesbians and attitudes toward gay men. Finally, the relatively high consistency of mean effect sizes across different measures of traditional gender role attitudes and anti-LG at titudes attests to the robustness of the relationship (Whitley, 2001). Kite and Whitley (1998) noted that the re lationship between traditional gender role attitudes and anti-LG attitudes could reflect a relationship between attitudes toward LG persons and a generalized conservative beli ef system. As such, it could be that the relationship between traditional gender role attitudes and anti-LG attitudes could be explained by variables such as RWA and SDO. Whitley and Egisdttir (2000) examined

PAGE 34

24 this possibility with 266 introductory psyc hology students. Each participant completed the ATLG (Herek, 1988), AWS (Spence, He lmreich & Stapp, 1973), RWA (Altemeyer, 1988) and SDO (Pratto et al., 1994) scales. In this study, gender role attitudes accounted for an additional 4% of the varian ce in attitudes toward lesbians, F (1, 249) = 17.22, p < .001, and an additional 8% of the vari ance in attitudes toward gay men, F (1, 249) = 36.66, p < .001, beyond variance accounted for by RWA and SDO. Therefore, gender role attitudes appear to account for varian ce in attitudes toward LG persons beyond that accounted for by RWA and SDO. Overall, studies on the link of traditional gender role attitudes to anti-LG attitudes demonstrate that heterosexual persons stereotypically assume that lesbians and gay men violate traditional gender role norms. Thus, persons who endorse traditional gender role attitudes hold more anti-LG attitudes. This significant relationship has been demonstrated using various measures of traditional gender ro le attitudes. However, the key aspect of gender role-related variables that is related to anti-LG attitudes appears to be traditional gender role attitudes and not gender role self-concept or modern sexism. Finally, the relationship between traditional gender role attitudes and anti-LG attitudes is not accounted for fully by a joint relationship w ith either RWA or SDO. Taken together, these findings support the inclusion of tradit ional gender role att itudes in the present study. Openness to Experience: An Underlying Construct The literature reviewed thus far indicates that conceptual definitions of right-wingauthoritarianism (RWA), social dominance orie ntation (SDO), and trad itional gender role attitudes link each of these variables to anti-l esbian and gay (LG) attitudes and empirical evidence supports such links. In addition to th eir links to anti-LG attitudes, RWA, SDO,

PAGE 35

25 and traditional gender role at titudes, each have also been linked conceptually and empirically to openness to experience, a dimension of personality described in the Five Factor Model. Openness, in turn, has been show n to be related to anti-LG attitudes. Thus, openness to experience might be a mediating va riable that provides a more parsimonious explanation and accounts for the links of RWA, SDO, and traditional gender roles to antiLG attitudes. Within the framework of the Five Fa ctor Model, openness to experience is conceptualized to reflect activ e imagination, aesthetic sensi tivity, receptiveness to inner feelings, preference for variety, intellectua l curiosity, and independence of judgment (McCrae & Costa, 2003). A person who is high in openness values intellectual matters, is nonconforming, introspective, has a vivid imagin ation, and admits that what is right and wrong for one person may not be applicable in other circumstances. On the other hand, a person scoring low in openness is likely to be rigid, have intolerance for ambiguity, have a narrow range of emotions, favor conservati ve values, judge others in conventional terms, and is uncomfortable with complexities These characteristics associated with low levels of openness overlap with the characteri stics associated with high levels of RWA, SDO, and traditional gender role attitudes. First, a lack of openness is evident in the traditional conception of authoritarianism and in the more modern definition of RWA. Adorno et al. (1950), who first defined authoritarianism, argued that persons high in this variable are closed-minded, mentally rigid, and disinterested in imaginative pursu its. These traits are also evident in Altemeyer’s (1981) conceptualiz ation of RWA. More specifically, Altemeyer conceived RWA as being comprised of conventionalism or a “high degree of adhe rence to the social

PAGE 36

26 conventions which are perceived to be endorsed by society and its established authorities” (p. 148). Thus, persons high in RWA are likely to reject thoughts or experiences that are not sancti oned by proper authorities. Indeed, RWA has consistently been shown to relate significantly and negatively to openness to experience, with correlations ranging from -.30 to -.57 (Altemeyer, 1996; Butler, 2000; Heaven & Bucci, 2001; Pete rson, Smirles, & Wentworth, 1997). For example, Butler (2000) admini stered the full 240-item NEO Pe rsonality Inventory (PI; Costa & McCrae, 1992) and the RWA scale (Altemeyer, 1996) to 76 participants recruited from psychology courses. In this sample, RWA was significantly correlated with openness, r = -.48, p < .001, but not with any of the other personality dimensions (i.e., extroversion, neuroticism, agreeablen ess, and conscientious ness). Furthermore, RWA was significantly correlated with five of the six facets of openness. The exception was the fantasy facet, but even this relationshi p was in the predicted direction and fantasy did not correlate strongly with any of the ot her openness facets in this sample. Consistent with theoretical conceptions of RWA, RWA scores were mo st strongly correlated with openness to values, followed closely by fee lings and actions. Butler (2001) claimed “given that the mean correlation of RWA and the openness facets ( r = -.30), compares favorably to the magnitude of the mean inte rcorrelations among the facets themselves ( r = .22), we may be justified in viewing author itarianism as an additional facet of openness to experience” (p. 8). Furthermore, McCr ae (1996) suggested that openness versus closedness “is an indispensable element” (p. 327) in explaining an authoritarian personality.

PAGE 37

27 Similarly, a lack of openness to experience is apparent in SDO. Persons high in SDO accept hierarchy-legitimizing myths in order to promote the superiority of one group over another (Pratto et al., 1994). Instead of viewing other people as individuals, persons high in SDO concentrate on targets’ out-group membership a nd characteristically have a lack of sympathy, empathy, and to lerance for members of out-groups. These characteristics, along with the lack of ability to view the complexities in the world and the narrow range of emotions evident in persons high in SDO, are consistent with a lack of openness to experience. Consistent with this theoretical exp ectation, research has shown that SDO is correlated significantl y and negatively with openness. Heaven and Bucci (2001) used the Interna tional Personality Item Pool (IPIP; Goldberg, 1999) with a sample of psychology students and found a correlation of r = -.26, p < .001 between SDO and openness. Additionally Pratto et al. (1994) used John, Donahue, and Kentle’s (1992) Big-Five Personality Invent ory and, with a sample of college students, found a correlation of r = .28, p < .01 between SDO and openness. Finally, there is a theoretical link between an adherence to traditional gender role attitudes and a lack of openness to experience. Persons who adhere to traditional gender role attitudes expect others to fit into a re latively stable set of ge nder roles, traits, and physical attributes (e.g., Kite & Whitley, 1998) This acceptance of traditional roles demonstrates a lack of tolerance for ambigu ity and variety within the gender spectrum. Additionally, it represents traditional and conforming values. Therefore persons who adhere to traditional gender role attitudes ar e expected to have low levels of openness to experience. Although there is a theoretical connection betw een traditional gender role attitudes and openness, previ ous research has not exam ined this link directly.

PAGE 38

28 Nevertheless, extant findings indicate a negative link between ope nness and constructs that are linked with traditional gender role attitudes. For example, Tokar, Fischer, Schaub & Moradi (2000) administered to a sample of male university st udents the Gender Role Conflict scale (O’Neil, Helms, Gable, David & Wrightsman, 1986), the Masculine Gender Role Stress scale (Eisler & Skidmo re, 1987), and the 60 item NEO Personality Inventory (Costa & McCrae, 1985). These au thors found that masculine gender role conflict and gender role stress both were si gnificantly and negativel y related to openness to experience with correlations of -.38 and -.21, respectively. Gender role conflict is a “psychological state in whic h gender roles have negative consequences or impact on the person or others” (O’Neil, 1981, p. 203). Men’s gender role stress re sults from a failure to meet cultural standards of masculinity (Eisler & Blalock, 1991). Thus, although men’s gender role conflict and stress are somewhat different from traditional gender role attitudes, there is significant theoretical overlap in that one must endorse traditional gender roles in order to experi ence gender role conflic t or stress. Indeed, both gender role conflict and stress are correlated significan tly with traditional gender role attitudes (Thompson, 1991, as cited in Thompson, Pleck, & Ferrera, 1992). Thus, it is expected that traditional gender role at titudes will relate to low ope nness and the present study will examine directly this proposition. It follows from the theoretical conceptualization of RWA, SDO, and traditional gender role attitudes and from empirical ev idence, that lack of openness to experience might be a global construct that underlies thes e various correlates of anti-LG attitudes. Furthermore, openness might be an underlying c onstruct that explains partially or fully, the links of the aforementioned correlates to an ti-LG attitudes. In fact, in addition to its

PAGE 39

29 conceptual and empirical links to RWA, SDO, and traditional gender attitudes, openness is also significantly related to prejudicial attitudes in ge neral and more specifically to anti-LG attitudes. For example, Ekehammar and Akrami (2003) examined the Five Factor Model and a generalized prejudiced factor with a sample of 156 Swedish, nonpsychology university students. The generali zed prejudice factor included anti-LG attitudes in addition to mode rn and classical racism, se xism, and attitudes toward intellectually disabled indivi duals. The participants took a Swedish translation of the NEO-PI (Costa & McCrae, 1985) and Scandinavi an versions of instruments assessing the prejudice variables. In th is study, openness was significantly related to scores on the generalized prejudice factor, r = -0.45, p < .001. In another study, Cullen, Wright, and Alessandri (2002) focused more specifically on the link between anti-LG attitudes and the Five Factor Model dimensions of pe rsonality with a sample of 123 university students. Participants comple ted the Neuroticism, Extrovers ion, and Openness scales of the NEO-PI (Costa & McCrae, 1985) and the ATLG (Herek, 1984). In a regression equation that included openness, neuroticism, extroversion, ge nder, and contact with LG persons, 25% of the total variance in anti -LG attitudes was accounted for, and openness was the only personality variable that accounted for significant unique variance. Since openness is correlated with both the predictor variables and the outcome variable of anti-LG attitudes, it is hypothe sized that openness will mediate the links of RWA, SDO, and traditiona l gender role attitudes to anti-LG a ttitudes. In other words, it is predicted that RWA, SDO and traditional gende r attitudes have indir ect links to anti-LG attitudes through openness. According to Bar on and Kenny (1986) for a variable to be considered a mediator, there must be a signi ficant relationship (a) between the predictor

PAGE 40

30 and the mediator, (b) between the mediator and the criterion, and (c) and the predictor and the criterion. The literature reviewed thus far fits with this de finition and statistical requirements. More specifically, in the pr evious literature, op enness to experience (mediator) has been shown to be correlated significantly with RWA, SDO, and traditional gender role attitudes (predictor s) as well as with anti-LG a ttitudes (criterion). The present study will provide an empirical test of this mediational relationship. Need to Study Anti-Lesbian and Gay Behaviors Although much of extant literature has focu sed on the correlates of anti lesbian/gay attitudes, an implicit reason fo r studying anti-LG attitudes is that the expression of these attitudes as anti-LG behaviors might ha ve serious negative interpersonal and intrapersonal consequences for LG persons and so ciety at large. In fact, the prevalence of anti-LG behaviors (e.g., name-calling, threats, assaults) is staggering. D’Augelli (1989) found that 50% of the gay and lesbian co llege students surveyed reported having overheard disparaging comments, 26% repor ted having experienced personal verbal insults, 26% reported having been threatened with physical violence, and 23% reported having been victims of assault. Furthermore, such negative experiences have been linked to mental health related concerns (e.g., su icidal ideation, psychol ogical distress) among lesbian and gay persons (e.g., Diaz et al, 2001; DiPlacido, 1998; Mays & Chochran, 2001). Thus there is a need to study anti-LG be haviors in addition to anti-LG attitudes. Despite the considerable pr evalence and potential negati ve consequences of antiLG behaviors and numerous calls for rese arch examining such behaviors (Kite & Whitley, 1998; Herek, 2000; Whitley, 2001), the ma jority of extant research has not attended to predictors or correlates of anti-LG behaviors and has focused mostly on identifying correlates of anti-LG attitudes. The limited research available shows that

PAGE 41

31 although hate crimes and anti-LG behavior s are inevitably influenced by complex situational factors (Frankli n, 1998), anti-LG attitudes do indeed contribute to anti-LG behaviors. A review of extant literature on the link between anti-LG attitudes and anti-LG behaviors will now be discussed. Franklin (2000) conducted empirical research into prevalence ra tes and motivations for anti-LG harassment and vi olence with a noncriminal sample. Her sample included 489 students at six community colleges in fi ve counties in Northern California. The sample was racially and economically dive rse with 40% White, 30% Asian, 14% Latino, and 12% African American, and 59% reporting that their father’s highest educational degree was a high school degree. Approximately 93% of the participants identified as heterosexual. The participants completed an Antigay Behaviors I nventory developed for this study. This 89-item inventory assessed (a ) frequencies of vari ous types of anti-LG behaviors, (b) descriptions of specific inci dents, (c) the stated motivations for such behaviors, and (d) motivations reported by nonassailants for not engaging in anti-LG assaults. In addition, particip ants completed the ATLG (Her ek, 1988), a measure of social drinking, and the Attitude Function Invent ory (Herek, 1986), which measures four distinct social and psychol ogical functions served by attitudes toward LG persons. Results of this study indicated that large percentages of the participants admitted engaging in anti-LG behaviors, having friends who had engaged in such behaviors, and witnessing anti-LG incidents. In fact, 10% (n = 49) of the sample reported physically assaulting or threatening people whom they believed were lesbian or gay and 32% (n = 158) reported calling LG persons by insulting na mes (but not physically assaulting them).

PAGE 42

32 Additionally, of the 321 responde nts who denied any verbal or physical harassment, 23% (n =83) reported witnessing such incidents. As part of the Antigay Behaviors Invent ory, individuals who admitted to engaging in anti-LG behavior were asked to rate 19 motivations on a 4-point Likert scale. The motivations included such items as “I did it to have fun” and “I did it because homosexuals disgust me.” Franklin (2000) ex amined the structure of these motivations using exploratory factor analyt ic procedures. She identified four factors labeled peer dynamics, antigay ideology, thrill seeking, and self-defense, which accounted for 64% of the variance in the data. Although peer dyna mics (35%) accounted for the greatest percentage of variance, antigay ideology accounted for the second highest amount of variance at 15%. This factor reflected anti-LG attitudes such as disgust, hatred, religious and moral values, and the beli ef that LG persons spread AIDS. The emergence of this factor demonstrates that anti-LG attitude s were a motivation for anti-LG behaviors. Furthermore, in this sample, a signifi cant relationship between anti-LG attitudes and anti-LG behaviors was found. The mean a ttitude score for participants who reported anti-LG behaviors ( M = 25.3, SD = 6.92) was significantly greater than that for nonassailants ( M = 21.5, SD = 7.24), t (484) = .25, p < .001. In addition, Franklin separated the sample into those who had committed anti-LG behaviors and those who had not. A point-biserial correlation revealed a positive relation between committing anti-LG behaviors and anti-LG attitude s as measured by the ATLG, r = .25, p < .001. Other studies have provided further suppor t for the link between anti-LG attitudes and behaviors. For example, Patel, Long, McCammon, and Wuensch (1995) found a relationship between anti-LG attitudes and be haviors with 102 male college students.

PAGE 43

33 Approximately half of the participants were enlisted marines enrolled in college classes on two military bases, whereas the rest were ci vilians enrolled in classes at a university. Participants completed a modified version of the Index of Attitudes Toward Homophobia (IAH; Hudson & Ricketts, 1980). This scale wa s developed to measure fear, disgust, anger, discomfort, and aversion toward LG pe ople. Patel et al. ( 1995) created the SelfReport of Behavior Scale (SBS) for this study to measure past negati ve behavior toward gay men in a variety of situations. Simila r to Franklin’s (2000) findings, anti-LG behaviors were not uncommon. In this samp le of men, 5% reported getting into a “physical fight with someone who I thought was making moves on me,” 34% verbally threatened someone who had “checked me out,” and 42% moved away from someone who was believed to be gay. Anti-LG attitudes were significantly corr elated with anti-LG behaviors, r = .40, p < .01. A multiple regression analysis was conducted predicting antiLG behaviors with anti-LG attitudes and personality variables (social adjustment, emotional stability, defensiveness) measured with the Minnesota Muliphasic Personality Inventory (MMPI). In this e quation, anti-LG attitudes accounted for significant unique variance in anti-LG behaviors ( = .38, p < .001). A limitation of the Patel et al (1995) study is that all of the participants were male and that attitudes and behaviors were a ssessed toward only gay men. Whitley (2001) addressed this limitation by assessing attitude s and behaviors toward lesbians and gay men with a sample of 227 female and 207 male introductory psychology students. AntiLG attitudes were assessed using a composite score of an affective measure and the Beliefs About Lesbians a nd Beliefs About Gay Men scales (LaMar & Kite, 1998). Whitley (2001) assessed anti-LG behaviors using a revised version of Patel et al.’s (1995)

PAGE 44

34 Self Report of Behavior Scale (SRBS-R). Th e scale was revised to include questions about lesbians in addition to questions about gay men. In this study, anti-LG attitudes and anti-LG behaviors were significan tly and moderately correlated r = .39. In addition to research on the self-repor t of anti-LG behaviors, experimental research has demonstrated that anti-LG attitudes are related to actual anti-LG behaviors, measured in a variety of ways. Bernat, Calhoun, Adams, and Zeichner (2001) found that in a sample of college men, those who held more anti-LG attitudes as measured by the Homophobia Scale (HS) were more aggressive to ward an LG target. In this experiment, physical aggression was measured by the number of shocks given to a fictitious opponent during a competitive reaction time task. Additio nally, Kite (1992) found that college men who held more anti-LG attitudes as meas ured by the Homosexuality Attitude Scale (HAS) were less likely to choose to meet a partner perceived to be lesbian or gay. In another study, Kite and Deaux (1986) found that when male college student participants had an interpersonal interaction with an LG target, those with anti-LG attitudes formed more negative impressions and recalled le ss information about the target than did participants who did not hold negative attit udes. A limitation of the previously described studies is that they only a ssessed anti-LG behaviors for ma le participants. A study that addressed this limitation examined proposed budget cuts for student organizations. Haddock and Zanna (1998) found that male a nd female students with anti-LG attitudes proposed greater cuts to an LG student or ganization (mean percentage funding reduction = 42%) than did participants with relatively neutral (25%) or favorab le attitudes (27%). Additionally, Swim, Ferguson, and Hyers ( 1999) found that female college students engaged in greater social dist ancing (by altering their stated opinions to disagree with the

PAGE 45

35 target) from a lesbian confederate than from a heterosexual woman confederate. Furthermore the participants with greater anti-LG attitudes, as measured by the HAS, socially distanced themselves in more ways (i.e., by altering their op inion to agree with the majority as opposed to making indepe ndent responses which differed from all participants) than those participants with less anti-LG attitudes. Overall, these experiments demonstrate that for both men and women anti-LG attitudes are positively related to anti-LG behaviors. Just as there is evidence linking anti-LG attitudes to anti-LG behaviors, there is also some support for the notion that the predic tors of anti-LG attit udes (i.e., right-wing authoritarianism (RWA), social dominance orie ntation (SDO), and trad itional gender role attitudes are related to antiLG behaviors. Support for the link between RWA and antiLG behaviors can be inferred from Schope and Ellison’s (2000) study on religious fundamentalism. These authors found, in a sa mple of 129 college student participants, that those who self identified as having “very traditional or fundamentalist” religious beliefs were more likely than their less re ligiously fundamentalist counterparts to have engaged in anti-LG behaviors. More sp ecifically, one third of the religious fundamentalist participants re ported that they had threat ened a lesbian or gay person within the last year, compared to only 4% of participants wh o identified their religion as “somewhat traditional or cons ervative” or liberal. While th e religious fundamentalists comprised only 12% of the sample, they account ed for 50% of those who had threatened gay persons. Although RWA was not directly examined in this study, RWA has been shown (a) to be related strongly ( r = .68) to religious f undamentalism (Altemeyer & Hunsberger, 1992) and (b) to account for the li nk of religious funda mentalism to anti-LG

PAGE 46

36 (Hunsberger, 1995). Therefore, evidence linki ng religious fundamentalism to anti-LG behaviors suggests that RWA is likely to be related to anti-LG behaviors as well. Furthermore, direct support for the li nk between RWA and anti-LG behaviors is provided by an experiment by Haddock, Zanna and Esses (1993). In this study, 63 college student participants completed the RWA scale. Two weeks later these participants returned and were told that the undergraduate psychology so ciety was soliciting views of psychology students for the student governme nt about budget cuts for student groups. A median split was conducted for RWA scores and an ANOVA revealed that those with high RWA scores proposed significantly gr eater cuts (37%) to an LG student organization than those with low RWA scores (25%), F (1, 53) = 6.05, p < .05. This study demonstrates that RWA is indeed related to anti-LG behaviors. Unlike research on the link of RWA to anti-L G behaviors, no empirical research to date has examined the relationship between SDO and anti-LG behaviors. However, SDO has a strong relationship with anti-LG attitudes. Furthermore, SDO is based on a theory of group conflict, which postula tes that people will engage in unofficial terror, which is “violence or threat of viol ence perpetrated by private indi viduals from dominant groups against member of subordinate groups” (S idanius & Pratto, 1999, p. 42). Thus it is expected that persons high in SDO will report more anti-LG behaviors. This study will be the first to empirically examine this relationship. Lastly, extant data suggest that traditiona l gender role attitudes are related to antiLG behaviors. Franklin’s (2000) study provided some support for the relationship between traditional gender role attitudes and anti-LG behaviors. Her participants completed the Male Role Attitudes Scale (MRAS; Pleck, Sonenstein, & Ku, 1993), which

PAGE 47

37 assesses participants’ perceptions about the importance of men fulfilling masculinity standards. Franklin found that male assailan ts scored significantly higher on masculinity ideology than did their n onassailant counterparts, t (202) = 3.24, p = .001. Also, for men, there was a significant correlation between ma sculine ideology and anti-LG behaviors, r = .21, p <. 01. On the other hand, there was not a significant difference between female assailants and nonassailants on masculine ideology, and for women, the relationship between masculine ideology and anti -LG behaviors was non-significant. MRAS scores only had adequate reliability with a coefficient alpha of .68 for the entire sample, male and female. This coul d be one reason why a significant relationship between masculine ideology and anti-LG be haviors was not found among the female participants. In fact, when Whitley (2001) ex amined traditional ge nder role attitudes about women and men (using AWS and MRNS scores, which had better psychometric properties than did MRAS scores in Franklin’s sample), he found that scores on the AWS (r = .49, p < .001) and MRNS ( r = .45, p < .001) were both correlated with anti-LG behaviors for women and men participants. In summary, traditional masculine ideology, a component of gender roles atti tudes was shown to be relate d to anti-LG behaviors for men. Furthermore, gender role attitudes assessed more broadly were also related significantly to anti-LG behaviors for both women and men. Purpose of Study Previous research has demonstrated that right-wing authoritarian ism (RWA), social dominance orientation (SDO), and traditi onal gender role attitudes are important predictors of anti-LG attitudes. Each of these variables is also related conceptually and empirically to openness to experience. Openness in turn is related to anti-LG attitudes. Therefore, openness to experience may serve as a mediator of the relationships of RWA,

PAGE 48

38 SDO, and traditional gender role attitudes to anti-LG attitudes. Furt hermore, research has also shown that anti-LG attitudes, RWA, and traditional gender roles are predictors of anti-LG behaviors. It is expected that SDO and openness to experience will also be related to anti-LG behaviors. However, it is not known if RWA, SDO, traditional gender roles, and openness to experience will relate directly to anti-LG behaviors, or if their relationship is mediated by anti-LG attitude s. Thus, the present study will examine if anti-LG attitudes mediate the relationships of RWA, SDO, traditional gender role attitudes, and openness to anti-LG behaviors. The model tested in the current study is presented in Figure 1 and exam ines the following hypotheses: 1. RWA, SDO, and traditional gender ro le attitudes will be significantly and positively related to anti-LG attitudes and behaviors and openness to experience will be significantly and negatively related to anti-LG attitudes and behaviors. Support for these links is a prerequisite for examining corresponding links in Hypothesis 2 and 3. 2. Openness to experience will either partially or fully mediate the relationships of RWA, SDO, and traditional gender role attitudes to anti-LG attitudes. 3. Finally, anti-LG attitudes will either partially or fully mediate the relationships of RWA, SDO, tradit ional gender role attitudes, and openness to experience to anti-LG behaviors.

PAGE 49

39 CHAPTER 3 METHODS Participants A total of 238 undergraduate students we re recruited at a large southeastern university from the Psychology 100 participant pool and from two other undergraduate psychology courses where extra credit was gr anted. Given the present study’s focus on heterosexual persons’ attitudes toward lesbia n and gay persons, tw o participants who identified as bisexual were not included in analyses, because it is expected that such persons’ attitudes and behaviors toward LG individuals could stem from different constructs than those focused on in the pres ent study. In addition, 4 participants were not included in analyses because they did not co mplete significant portions of the survey packet, and 1 participant was excluded becau se she or he did not indicate gender. Therefore analyses were conducted using da ta from 231 participants; 97 men and 134 women. Participants ranged in age from 17 to 27 years ( M = 19.48; Mdn = 19; SD = 1.51) Eighty-nine percent of the sample identi fied as exclusively heterosexual and 11% as mostly heterosexual. With regard to race /ethnicity, 58% of the sample identified as White, 16% Latina/Hispanic, 11% African Amer ican/Black, 6% Asian American/Pacific Islander, 0.4% Native American, and 10% multiracial or other. Overall, 50% of the participants were in their fi rst year of college, 23% in th eir second year, 17% in their third year, and 10% in their fourth year. Fiftytwo percent of the participants reported that they were single, 37% were in a dating re lationship, and 11% were married or in a

PAGE 50

40 committed relationship. In terms of social class, 44% of the sample identified as middle class, 34% upper middle class, 15% working class, 4% lowe r class, and 3% upper class. With respect to religious identification, 32% identified as Catho lic, 23% Christian, 11% Baptist, 8% Methodist, 5% Jewish, 4% Presbyterian, 1% Muslim, .5% Buddhist, 3% Agnostic, and 3% Atheist or no religious orientation. Furthe rmore, 31% of the sample reported that they never attend religious se rvices, 29% attend less than once a month, 9% once a month, 9% twice a month, 14% once a we ek, and 9% attend services more than once a week. Procedures Participants completed a survey packet that included the instruments described below as well as a demographic questionnaire. Before beginning the survey, participants received written informed cons ent, and after completing the survey, they received written debriefing. The order of the instruments in the survey packet was counterbalanced to control for order effects. Instruments Criterion Variables Attitudes toward LG persons were assessed with the Attitudes Toward Lesbians and Gay Men (ATLG) scale. Herek’s (1988) scale is one of the more commonly used measures of anti-LG attitudes and assesses th e affective component of anti-LG attitudes referred to as “condemnation-tolerance.” Th e ATLG is a 5-point Likert scale ranging from 1 = disagree strongly to 5 = agree st rongly. The scale has 20 items; 10 items assess attitudes toward lesbians (ATL), and 10 items assess attitudes toward gay men (ATG). Sample items include “Homosexual behavior between two men is just plain wrong” and “The growing number of lesbians indicate a decline in American morals.” With this

PAGE 51

41 scale, separate scores for attitudes toward lesbian persons and attitudes toward gay persons can be computed by averaging or adding ATL or ATG items, respectively. However, given that the focus of the present study is on anti-LG attitudes in general, and that ATL and ATG scores are highly correlated ( r = .88), overall ATLG scores were used in the present study. Appropriate items were reverse scored and item ratings were averaged to yield an overall ATLG score with higher scores indicating more negative attitudes. Construct validity for ATLG scores has been demo nstrated through consistently high correlations with variable s associated conceptually with anti-LG attitudes such as dogmatism, conservative political ideology, and l ack of personal contact with LG persons (Herek, 1988; Whitley & Lee, 2000). ATLG scor es have yielded satisfactory levels of internal consistency. More specifically, for sc ores on the combined scale, attitudes toward lesbians subscale, and attitudes toward gay men subscale, Cronbach’s alphas were .95, .90, and .90 respectively with Herek’s (1988) sample of 405 college students. In the sample for the present study, the Cronbach’s alpha for ATL scores was .90 for men, .92 for women, and .91 for the combined sample. AT G scores yielded a Cronbach’s alpha of .93 for men, women, and combined samples. Finally, overall ATLG scores (used in the analyses) yielded a Cronbach’s alpha of .95 for men, .96 for women, and a .96 for the combined sample. Behaviors toward LG persons were assessed with the Self-Report of Behavior Scale-Revised (SRBS-R) The SRBS-R is a 30 item, 5-poi nt Likert-type scale with item ratings ranging from 1 = never to 5 = always Item ratings were averaged to yield an overall scale score and higher scores indicat e more negative behaviors. The scale was originally created by Patel, Long, McCamm on, and Wuensch (1995) and then revised by

PAGE 52

42 Roderick, McCammon, Long, and Allred (1998). It was designed to measure the extent to which respondents report having engaged in anti-LG behaviors in the past. A factor analysis demonstrated that the SRBS-R consisted of two factors that reflected passive avoidance or moving-away-from behaviors and active aggression or moving-against behaviors Although the original version of the SRBS-R refe rs to “a gay person” as the target of the behaviors, the current study as ked participants to report their behaviors separately for lesbians and gay men as used by Whitley (2001). Sample items include: “I have spread negative talk about someone because she was a lesbian,” and “I have participated in playing jokes on some one because I suspected he was gay.” In addition, validity is demonstrated by the relations hip between SRBS-R scores and negative affective responses to gay persons as m easured by the Index of Attitudes Toward Homosexuals scale (IAH; Roderi ck et al., 1998). Furthermore, SRBS-R scores are related to Scale 9 of the Minnesota Muliphasic Pers onality Inventory (MMPI), which measures overactivity, emotional liability, hostility, aggr essive outbursts, and thrill seeking (Patel et al., 1995). In the current sample, for both men and women, SRBS-R scores were not significantly related to scores on the Balanced In ventory of Desirable Responding (BIDR; Paulhus, 1994). SRBS-R scores have good internal consistency, with Cronbach’s alphas of .88 for the lesbian scale and .93 for the gay male scale (Whitley, 2001). Whitley reported that anti-lesbian behaviors and an ti-gay behaviors were strongly correlated ( r = .41) and loaded strongly on a single factor in a factor analysis of attitudinal and behavior measures. Thus, Whitley (2001) argued that th e two subscales should be combined into a single scale score labeled anti-LG behavior as was done in th e present study. The

PAGE 53

43 obtained alpha internal consis tency reliability estimate for SRBS-R scores with the current sample was .93 for men, .89 for wo men, and .91 for the combined sample. Predictor Variables Right-wing authoritarianism was measured with the Right-Wing Authoritarianism (RWA) scale. Developed by Altemeyer (1981, 1988), this scale assesses authoritarian submission, author itarian aggression, and conven tionalism. It is a 30-item, 9-point Likert-type scale, with item ratings ranging from -4 = very strongly disagree to 4 = very strongly agree, as recommended by Altemeyer (1998). Item ratings were transposed to a 1-9 scale and then averaged to yield an overall score, with higher scores indicating higher levels of RWA. Sample items include: “Obedience and respect for authority are the most important virtues ch ildren should learn,” and “What our country really needs is a strong, determined leader w ho will crush evil and take us back to our true path.” One item assesses attitudes toward LG persons. Consistent with prior research (Whitley & Egisdottir, 2000; Whitley & Lee, 2000), this item was not scored in the present study to avoid an inflated correla tion with the ATLG due to item overlap. With regard to validity, Altemeyer (1981) demonstrated th at RWA scores were correlated positively with scores on other measures of authoritarian personality including the F scale (Adorno et al., 1950), the Dogmatism scal e (Rokeach, 1960), the Conservatism scale (Wilson & Paterson, 1968), the Balanced F s cale (Kohn, 1972), and the Authoritarianism Rebellion scale. Additionally, RWA scores were significantly correlated with scores on theoretically related variables such as orient ation to established authority and the law, acceptance of the law as the basis of moralit y, and punitiveness toward sanctioned targets (Altemeyer, 1981, 1988, 1996). In terms of reliability, RWA scores have yielded

PAGE 54

44 Cronbach’s alphas ranging from .81 to .95 (Hil l & Wood, 1999). In the current sample, the obtained alphas were .95 for the me n, women, and combined samples. Social dominance orientation was assessed with the Social Dominance Orientation (SDO) scale. Pratto et al. (1994) devel oped this scale to measure respondents’ preference for in equality among social groups. The SDO is a 14 item, 7point Likert-type scale and participants rate their responses on a range from 1 = very negative to 7 = very positive. A scale score was computed by averaging item ratings and higher scores indicate greater levels of SDO. Sample ite ms include: “It’s OK if some groups have more of a chance in life than ot hers,” and “We would ha ve fewer problems if we treated people more equally.” With regard to validity, SDO scores have been shown to correlate positively with cultural elitism, ethnic prejudice, sexism, political-economic conservatism, and a desire to end affirmativ e action, thus demonstrat ing a high degree of construct validity (Pratto et al., 1994). A dditionally, SDO scores have yielded high reliability estimates with an average Cr onbach’s alpha of .90 across multiple samples (Pratto et al., 1994). In the current samp le, alpha was .87 for men, .87 for women, and .88 for the combined sample. Traditional gender role attitudes was measured with two scales, the AntiFemininity subscale of the Male Role Norms Scale (MRNS) and the Anti-Masculinity subscale of the Female Role Norms Scale (FRNS). These scales assess the degree to which participants endorse traditional gender roles, that women should abide by traditional female-role norms and that men should abide by traditional male-role norms. Thompson and Pleck (1986) developed the Anti -Femininity subscale to assess the degree to which participants endors ed the traditional male role norm that men should avoid

PAGE 55

45 anything feminine. The scale consists of 7 ite ms and uses a 7-point Likert-type scale, with item ratings ranging from 1 = very strong ly disagree to 7 = very strongly agree. A scale score was computed by averaging item ratings and higher scores indicate more traditional masculine ideology. Sample items include: “It bothers me when a man does anything that I consider ‘femini ne’,” and “I might find it a litt le silly or embarrassing if a male friend of mine cried over a sad love scene in a movie.” Although this scale was originally normed on men, it has since been used with women (e.g., Kurpius & Lucart, 2000; Whitley, 2001). With samples of men, the va lidity of MRNS scores is supported by their relationship to holding an adversarial view of sexual relations hips, lower levels of self-disclosure, and a greater number of re ported sexual encounters (Sinn, 1997). With a sample that included women and men, MRNS scores were related positively to AWS scores and endorsement of traditiona l gender roles for oneself (Whitley, 2001). Furthermore, MRNS scores were minimally correlated with the masculinity and femininity scales of the Bem Sex Role I nventory (Bem, 1974), demonstrating that the MRNS assesses a construct that is unique fr om trait based masculinity and femininity (Sinn, 1997). In addition, MRNS scores ha ve good internal consistency with a Cronbach’s alpha of .88 for a male sample (S inn, 1997). Of the two studies that used the MRNS with women, one did not report reliab ility information (Kur pius & Lucart, 2000). The other reported an alpha of .87 for a comb ined sample of men and women and for the total MRNS score including the Toughness and Status Seeking subscales. With the current sample, alpha for MRNS scores was .80 for men, .77 for women, and .82 for the combined sample.

PAGE 56

46 The Anti-Masculinity scale assesses the de gree to which respondents believe that women should abide by traditional female role norms by avoiding anything masculine (Lefkowitz, Shearer, Gillen, & Espinosa-Herna ndez, 2003). The scale consists of 7 items and uses a 7-point Likert-type scale with item ratings ranging from 1 = strongly disagree to 7 = strongly agree. An overall score wa s computed by averaging the item ratings and higher scores indicate more traditional gender role attitudes. Sample items include: “A woman whose hobbies are fishing, fixing cars, and watching sports probably wouldn’t appeal to me,” and “Unless she was really de sperate, I would probably advise a woman to keep looking rather than accept a job as a cons truction worker.” In this sample, scores on the Anti-Masculinity scale were significan tly correlated with scores on the AntiFemininity scale ( r = .55 for women; r = .71 for men), and in other samples to beliefs about males’ and females’ roles in the family such as marital and childrearing roles, and endorsement of a sexual double standard (L efkowitz et al., 2003). Anti-Masculinity scores yielded Cronbach’s alpha of .79 with a sample of males and females (Lefkowitz et al., 2003). In the current sample, FRNS scor es yielded an alpha of .66 for men, .78 for women, and .75 for the combined sample. Since there is theoretical evidence that an overall endorsement of traditional gender role attitudes for the roles of women and men is related to anti-LG attitudes and behaviors, in this study an ove rall composite score of traditi onal gender role attitudes was created by computing the average of Anti-Femi ninity and Anti-Masculinity item ratings. Evidence of reliability for the composite tr aditional gender role attitudes scores was obtained by examining item-total correlati ons and the alpha internal consistency reliability estimate for the sample. With the current sample, the composite scale scores

PAGE 57

47 yielded a Cronbach’s alpha of .82 for me n, .86 for women, and .87 for the combined sample. The range of corrected item -total correlation was .05-.61 for men ( M = .45), .33.70 for women ( M = .48), and .29-.64 for the combined sample ( M = .53). The item with the lowest corrected item-total correlation of .05 for men was “I might find it a little silly or embarrassing if a female friend of mine shouted at the tele vision during a football game.” With this item removed, the range of corrected item-total co rrelations would have been .21-.61, but overall alpha would not have improved substantially (alpha = .84 if the item is deleted). Furthermore, this item ha d a corrected item-total correlation of .41 for women and .29 for the combined sample. Theref ore, this item was retained in computing overall traditional gender role attitudes scores. Based on theoretical conceptualizations and past research, positive correlations between traditional gender role attitudes a nd RWA and SDO as well as non-significant or small correlations with social desirability would support the construc t validity of scores on the composite measure. As such, validity of composite traditional gender role attitudes scores was supported by significant po sitive correlations with RWA ( r = .39 for men and .44 for women) and SDO ( r = .29 for men and .40 for women) scores (convergent validity) and non-significant correlations w ith the Balanced Inventory of Desirable Responding (BIDR; Paulhus, 1994) for men and the combined sample and a negative correlation for women ( r = -.19). Openness to experience was measured with the International Personality Item Pool (IPIP) openness scale. The IPIP was designed by Goldberg (1999) and is available through the website: http://ipi p.ori.org. Like the NEO-PI-R (Costa & McCrae, 1992), the IPIP scales assess the personality dimensions of the Five Factor Model (International

PAGE 58

48 Personality Item Pool, 2001). The 20-item ope nness to experience subscale was used in the present study. Respondents rated how accura tely each item described them on a fivepoint scale from 1=very inaccurate to 5=ve ry accurate. Sample items include: “Have difficulty understanding abstra ct ideas” and “Enjoy hearing new ideas.” IPIP Openness scores correlate highly ( r = .83) with the corresponding Openness scores on the NEO-PIR (International Personality Item Pool, 2001). In addition, subscale scores of the IPIP’s openness scale correlate with corresponding subscale scores of the Sixteen Personality Factors Questionnaire (16PF; Conn & Rieke, 1994); more specifically, Intellect scores correlate to Reasoning at r = .51, Imagination scores corr elate with Abstractedness at r = .65, and Emotionality scores co rrelate with Tension at r = .59 (International Personality Item Pool, 2001). Openness scores yielded a Cron bach’s alpha of .89 in previous research (International Personality Item Pool, 2001) a nd with the current sample, alphas were .87 for men, .88 for women, and .88 for the combined sample.

PAGE 59

49 CHAPTER 4 RESULTS Preliminary analyses were conducted to explore potential gender differences, examine descriptive statistics, and test poten tial order effects. Tests of hypotheses were conducted following these preliminary analyses. Gender Differences Examination of correlation matrices for vari ables of interest by gender revealed a few differences in the pattern of intercorre lations for women and men (see Tables 1 and 2). Table 1. Zero-Order Correlations Among Variables of Interest for Women Variables 1 2 3 4 5 1. Negative Attitudes Toward LG Persons 2. Negative Behaviors toward LG persons .53** 3. Right-Wing Authoritarianism .79** .34** 4. Social Dominance Orientation .19* .22* .36** 5. Traditional Gender Role Attitudes .46** .47** .44** .40** 6. Openness to Experience -.23** -.23** -.35** -.34** -.52** Note p < .05. ** p <.001. Table 2. Zero-Order Correlations Among Variables of Interest for Men Variables 1 2 3 4 5 1. Negative Attitudes Toward LG Persons 2. Negative Behaviors Toward LG persons .47** 3. Right-Wing Authoritarianism .79** .34** 4. Social Dominance Orientation .17 .24* .08 5. Traditional Gender Role Attitudes .50** .50** .37** .29** 6. Openness to Experience -.37**-.37**-.45** -.16 -.32** Note p < .05. ** p <.001.

PAGE 60

50 Previous research also has found gender diffe rences in levels of negative attitudes towards lesbian and gay (LG) persons (Kerns & Fine, 1994; Kite & Whitley, 1996; LaMar & Kite, 1998) and in relations among pred ictors of negative at titudes towards LG persons (Basow & Johnson, 2000). Due to the pa ttern of gender difference in the present data and prior literature, all analyses were conducted separately for women and men in order to ensure that combining all particip ants into one sample did not mask unique patterns of findings for women and men. Descriptive Statistics Although the present sample’s scores on some of the variables of interest indicated some restriction in range (see Tables 3 and 4), m eans and standard deviations for the present sample were generally comp arable to those obtained using the same instruments in previous sample s of undergraduate students. Table 3. Summary Statistics Among Va riables of Interest for Women Variables Possible Range Sample Range M SD 1. Negative Attitudes Toward LG Persons 1.00-5.00 1.00-4.80 2.25 .95 .96 2. Negative Behaviors toward LG persons 1.00-5.00 1.00-2.90 1.23 .28 .89 3. Right-Wing Authoritarianism 1.00-9.00 1.03-6.55 3.82 1.53 .96 4. Social Dominance Orientation 1.00-7.00 1.00-4.64 2.34 .86 .87 5. Traditional Gender Role Attitudes 1.00-7.00 1.00-5.00 2.53 .92 .86 6. Openness to Experience 1.00-5.00 2.45-4.95 4.01 .55 .88 Note. Higher scores indicate higher le vels of the construct assessed. More specifically, levels of anti-LG attit udes were generally close to the mid range of possible scores for both women and men a nd the current sample’s means and standard deviations for anti-LG attitudes (women: M = 2.25, SD = .95; men: M = 2.59, SD = .91) were comparable to those reported by Kerns and Fine (1994) in a sample of undergraduate women ( M = 2.88, SD = .80) and men ( M = 3.26, SD = .99). Relatively

PAGE 61

51 few anti-LG behaviors were reported in the current sample by women ( M = 1.23, SD = .28) and men ( M = 1.42, SD = .42). These low scores are comparable to the scores reported by Whitley (2002) with an underg raduate university sample of women ( M = 1.25, SD = .28) and men ( M = 1.55, SD = .56). Scores for right-wing authoritarianism (RWA) were below the possible mid-point demons trating lower levels of the variable. In addition, this sample’s scores for RWA (women: M = 3.82, SD = 1.53; men: 3.92, SD = 1.40), were lower than those reported by Whitley and Egisdttir (2000) for undergraduate women ( M = 5.28, SD = 1.00) and men ( M = 5.03, SD 1.18), but comparable to those reported by Basow and Johnson (200 0) for undergraduate women ( M = 3.95, no SD reported) and those reported by Butler (2000) for undergraduate men and women ( M = 4.10, SD = 1.00). Like RWA scores, social domin ance orientation (SDO) scores also were below the possible mid-point demonstr ating lower levels of this variable. Nevertheless, the current sample’s means and standard deviations for SDO scores (women: M = 2.34, SD = .86; men: M = 2.99, SD = .99) were comparable to those reported by Whitley (1999) with a sa mple of undergraduate women ( M = 2.79, SD = .96) and men ( M = 3.46, SD = 1.17). Table 4. Summary Statistics Among Variables of Interest for Men Variables Possible Range Sample Range M SD 1. Negative Attitudes Toward LG Persons 1.00-5.00 1.00-4.78 2.59 .91 .95 2. Negative Behaviors toward LG persons 1.00-5.00 1.00-3.10 1.42 .42 .93 3. Right-Wing Authoritarianism 1.00-9.00 1.00-7.45 3.92 1.40 .95 4. Social Dominance Orientation 1.00-7.00 1.00-5.57 2.99 .99 .87 5. Traditional Gender Role Attitudes 1.00-7.00 1.00-5.57 3.29 .86 .82 6. Openness to Experience 1.00-5.00 2.45-4.95 3.75 .56 .87 Note. Higher scores indicate higher le vels of the construct assessed.

PAGE 62

52 Since the traditional gender role attitudes measure used in the present study was a combination of previously used scales, the pr esent sample’s scores on overall traditional gender role attitudes (women: M = 2.53, SD = .92; men: M = 3.28, SD = .86) could not be compared directly to that of prior samples. However, this sample’s scores on the Antifeminity subscale (women: M = 2.80, SD = 1.04; men: M = 3.76; SD = 1.14), which comprised half of the items in the composite measure, were comparable to Kurpius and Lucart’s (2000) scores for a sample of college students (women: M = 2.67, SD = 1.02; men: M = 3.36, SD = .80). Past research has not reporte d means and standard deviations for the Antimasculinity subscale of the Fema le Role Norms Scale, which comprised the second half of the traditional ge nder role attitudes scale. Last ly, with respect to openness to experience, scores were generally close to the mid range of possible scores for both women and men. The current sample’s means and standard deviations for openness to experience (women: M = 4.01, SD = .55; men: M = 3.75, SD = .56) were comparable to those reported by Chan, Rounds and Drasgow (2000) for openness with a combined sample of college women and men ( M = 3.79 SD = .55). Overall, the pattern of similarity of descriptive statistics for the present sa mple, and that of previous samples of undergraduate students, suggests that any restriction in range of variables of interest found in the present sample likely reflects the range of attitudes among undergraduate students rather than pointing to sample specific idiosyncrasies. Test for Order Effects To test for order effects across the tw o orders of the survey, a MANOVA was conducted with survey order as the independent variable and th e variables of interest (i.e., anti-LG attitudes, anti-LG behaviors, RWA, SDO, traditional gender role attitudes, and openness to experience) as dependent variable s. The overall model was not significant ( F

PAGE 63

53 [3, 229] = .77, p > .05, p 2 = .20), demonstrating that survey order did not significantly impact participants’ responses. Intercorrelations Among Variables of Interest: Hypothesis 1 Zero-order correlations were computed to test hypothesized relations among variables of interest. More specifically, Hypothesis 1 st ated that (a) RWA, SDO, and traditional gender role attitudes will be significantly and positively related to anti-LG attitudes and behaviors and (b) openness to experience will be significantly and negatively related to anti-LG attitudes a nd behaviors. Analyses were conducted separately for women and men participants. Intercorrelations for women. For women, Hypothesis 1 wa s fully supported (see Table 1). As expected, antiLG attitudes were correlat ed positively with RWA ( r = .79, p < .001), SDO ( r = .19, p < .05), and traditional gender role attitudes ( r = .46, p < .001), and negatively with openness to experience ( r = -.23, p < .001). Additionally, anti-LG behaviors were correlated positively with RWA ( r = .34, p < .001), SDO ( r = .22, p < .05), and traditional gender role attitudes ( r = .47, p < .001), and negatively with openness to experience ( r = -.23, p < .01). Furthermore, openness to experience was negatively correlated with RWA ( r = -.35, p < .001), SDO ( r = -.34, p < .001), and traditional gender role attitudes ( r = -.52, p < .001). Intercorrelations for men. For the men, Hypothesis 1 was partially supported (see Table 2). Anti-LG attitudes were correlated positively with RWA ( r = .79, p < .001) and traditional gender role attitudes ( r = .50, p < .001), and negatively with openness to experience ( r = -.37, p < .001). Additionally, anti-LG behaviors were correlated positively with RWA ( r = .34, p < .01), SDO ( r = .24, p < .05), and traditional gender

PAGE 64

54 role attitudes ( r = .50, p < .001), and negatively with openness to experience ( r = -.37, p < .001). Furthermore, openness to experien ce was correlated nega tively with RWA ( r = .45, p < .001) and traditional gender role attitudes ( r = -.32, p < .01). However, inconsistent with Hypothesis 1, anti-LG attitu des were not correlated significantly with SDO, and SDO was not correlated signifi cantly with openness to experience. Tests of Mediation: Hypotheses 2 and 3 Hypothesis 2 stated that openness to expe rience would either partially or fully mediate the relations of RWA, SDO, and traditional gender role attitudes to anti-LG attitudes. In addition, Hypothesi s 3 stated that anti-LG attitu des will either partially or fully mediate the relations of RWA, SDO, tr aditional gender role at titudes, and openness to anti-LG behaviors. To test the mediati ons proposed in these hypotheses, Baron and Kenny’s (1986) procedures were followed. Accord ing to these authors, for a variable to be considered as a mediator, significant rela tions must exist between (a) the predictor and the mediator, (b) the mediator and the cr iterion, and (c) the pr edictor and criterion. According to Baron and Kenny (1986), if precond itions are satisfied, a variable acts as a mediator to the extent that it accounts for the relationship between the predictor and criterion. In order to test the significance of mediations Amos 5.0 (Arbuckle, 2003) was used to conduct path analyses of a fully satu rated model in which al l direct and indirect paths were estimated (see the models pres ented in Figures 1 and 3). To test the significance of mediation, appropriate standard ized path coefficients were multiplied to compute indirect effects, a procedure recommended by Cohen and Cohen (1983), and Sobel’s formula (1982) was used to determine whether or not the indirect effects were significantly different from zero. Again, sepa rate analyses were conducted for women and men. Maximum likelihood estimation was utili zed with the covarian ce matrix of the

PAGE 65

55 variables of interest as input. Given that the models tested were fully saturated, values for the Goodness of Fit Index (GFI), Incremen tal Fit Index (IFI), Comparative Fit Index (CFI), and the Normed Fit Index (NFI) al l were 1.0. The model for women accounted for 67% of the variance in anti-LG attitudes, 38% of the varian ce in anti-LG behaviors, and 30% of the variance in openness to experience. The model for men accounted for 67% of the variance in anti-LG attitude s, 35% of the variance in an ti-LG behaviors, and 24% of the variance in openness to experience. Tests of mediation with women. Hypothesis 2 proposed that openness would mediate the links of RWA and traditional gende r role attitudes to an ti-LG attitudes. For Hypothesis 2, preconditions for mediation were satisfied for the role of openness as a mediator of the link of traditional gender ro les to anti-LG attitudes, but not for openness as a mediator of RWA and SDO to anti-LG atti tudes (see Figure 1 for path coefficients). That is, zero order correlations indicated that traditional gender role attitudes (predictor variable) were correlated significantly with anti-LG attitudes (criterion). Furthermore, the path model indicated that traditional gender role attitudes (predi ctor variable) were related significantly and uniquely to openness (p otential mediator), which in turn was related significantly and uniquely w ith anti-LG attitudes (criterion). Sobel’s test indicated, however, that the indirect link of trad itional gender role attitude to anti-LG attitude s through openness was not signifi cant. Given that in the full model preconditions for mediation were not me t for RWA or SDO (i.e., neither variable was related uniquely to openness) and the in direct effect was not significant for traditional gender role attitudes, openness did not act as a mediator and thus Hypothesis 2 was not supported.

PAGE 66

56 Right-Wing Authoritarianism Traditional Gender Attitudes Social Dominance Orientation Openness to Experience Anti-LG Behaviors Anti-LG Attitudes .36** .40** .44** -.12 -.33* .78** -.41** .24** .27* -.13 .12* .64** -.03 -.14* .09 Figure 1. Fully saturated model depicti ng relationships am ong variables of interest for women. Values reflect st andardized coefficients. *p < .05. **p < .001. Hypothesis 3 proposed that an ti-LG attitudes would fully or partially mediate the links of RWA, SDO, traditional gender role attitudes, and openness to anti-LG behaviors. The preconditions for mediation were satisfied for the role of anti-LG attitudes as a mediator of the links of RWA, SDO, and tr aditional gender role attitudes to anti-LG behaviors, but not for anti-LG attitudes as a mediator of openness to anti-LG behaviors (see Figure 1 for path coefficients). That is zero order correlations indicated that RWA, SDO, and traditional gender role attitudes, but not openness (predict or variables) were correlated significantly with anti-LG behavior s (criterion). Furthermore, the path model indicated that RWA, SDO and tr aditional gender role attitudes (predictor variables) were related significantly and uniquely to anti-LG att itudes (potential mediator), which in turn was related to anti-LG behaviors (criterion).

PAGE 67

57 Multiplication of appropriate standardized path coefficients and Sobel’s formula indicated that consistent with Hypothesis 3, through anti -LG attitudes, a significant indirect link of .50 (.78 x .64; z = 5.00, p < .001) was obtained from RWA to anti-LG behaviors. Since there was also a signifi cant direct relationship between RWA and antiLG behaviors, anti-LG attitudes acted as a partial mediator. Additionally, through antiLG attitudes, a significant indirect link of -.09 (-.14 x .64; z = -2.25, p < .05) was obtained from SDO to anti-LG behaviors. Since there was not a significant direct relationship between SDO and anti-LG behavi ors, anti-LG attitudes acted as a full mediator. Also, through anti-LG attitudes, a significant indirect link of .15 (.24 x .64; z = 3.04, p < .01) was obtained from traditional gender role attitudes to anti-LG behaviors. Since there was a significant direct relationshi p between traditional gender role attitudes and anti-LG behaviors, antiLG attitudes acted as a pa rtial mediator. Therefore, Hypothesis 3 was partially supported in that an ti-LG attitudes partially mediated the links of RWA and traditional gender role attitudes to anti-LG behaviors and fully mediated the links of SDO to anti-LG behaviors, but an ti-LG attitudes did not mediate the link of openness to anti-LG behaviors. In the model for women, there were three si gnificant unique links that were in the unexpected directions. There was a positive si gnificant unique link from openness to antiLG attitudes ( = .12, p < .05), a significant negative un ique link from SDO to anti-LG attitudes ( = -.14, p < .05), and a significant negative un ique link from RWA to anti-LG behaviors ( = -.33, p < .01).

PAGE 68

58 Right-Wing Authoritarianism Traditional Gender Attitudes Social Dominance Orientation Openness to Experience Anti-LG Behaviors Anti-LG Attitudes .36 .40 .44 -.27 .77 -.52 .24 .32 .12 .60 -.14 Figure 2. Trimmed model depicting relati onships among variables of interest for women. Values reflect standardized coefficients. All paths depicted are significant at p < .05. Finally, the fit of a fully saturated model was compared to that of an alternative trimmed model that eliminated the non-si gnificant direct path s (a) from RWA to openness, (b) from SDO to openness, (c) from SDO to anti-LG behaviors, and (d) from openness to anti-LG behaviors (see Figure 2 for the trimmed model). The goodness of fit indices for this model were above the acceptabl e cut offs and nearly identical to those obtained from the original model (GFI = .98; IFI = .99; CFI = .99; NFI = .98). The trimmed model explained 68% of the variance in anti-LG atti tudes, 38% of the variance in anti-LG behaviors, and 27% of the varian ce in openness. As indica ted in Figures 1 and 2, the amounts of variance accounted for in e ndogenous variables as well as significance

PAGE 69

59 decisions and magnitudes of path s coefficients were similar across the fully saturated and trimmed models. Specifically, in the trimme d model the paths from traditional gender role attitudes to openness increased from -.41 to -.52, from RWA to anti-LG attitudes decreased from .78 to .77, from traditional ge nder role attitudes to anti-LG behaviors increased from .27 to .32, and from RWA to an ti-LG behaviors decreased from -.33 to .27. Thus, compared with the fully saturated model, the trimmed model appears to be more parsimonious but equally appropriate in explaining the relationships among the variables of interest. Tests of mediation with men. To test the mediations proposed in Hypotheses 2 and 3 for men in the sample, Baron and Kenny’s (1986)’s procedures, multiplication of appropriate standard path coefficients and S obel’s formula were used as outlined above. Hypothesis 2 stated that openne ss to experience would either partially or fully mediate the relationships of RWA, SDO, and traditional gender role attitudes to anti-LG attitudes. The precondition that openness (p otential mediator) must be uniquely related to anti-LG attitudes (criterion) was not met (see Figure 3), thus H ypothesis 2 was not supported. With regard to Hypothesis 3, preconditions for mediation were satisfied for the role of anti-LG attitudes as a mediator of the links of RWA, and traditional gender role attitudes to anti-LG behaviors, but not for anti-LG attitude s as a mediator of SDO and openness to anti-LG behaviors (see Figure 1 for path coefficients). That is, zero order correlations indicated that RWA and traditional gender role attitudes (predictor variables) were correlated significantly with anti-LG be haviors (criterion). Furthermore, the path model indicated that RWA and traditional gender role attitudes, but not SDO or openness (predictor variables) were related signi ficantly and uniquely to anti-LG attitudes

PAGE 70

60 (potential mediator), which in turn was related significantl y and uniquely with anti-LG behaviors (criterion). Right-Wing Authoritarianism Traditional Gender Attitudes Social Dominance Orientation Openness to Experience Anti-LG Behaviors Anti-LG Attitudes .08 .29* .37** -.39** -.13 .71** -.15 .24** .29* -.08 .03 .34* -.20* .04 .08 Figure 3. Fully saturated model depicting relationships among variables of interest for men. Values reflect standardized coefficients. p < .05. ** p < .001. Consistent with Hypothesis 3, through anti-LG attitudes, a significant indirect link of .24 (.71 x .34; z = 2.31, p < .05) was obtained between RWA and anti-LG behaviors. Because no significant direct link existed be tween reported RWA and anti-LG behaviors, anti-LG attitudes acted as a full mediator of this link. However, through anti-LG attitudes there was not a significant indirect link be tween traditional gende r role attitudes or openness and anti-LG behaviors. Therefore, H ypothesis 3 was partially supported in that

PAGE 71

61 anti-LG attitudes fully mediated the link of RWA to anti-LG behaviors, but not the links of SDO, traditional gender role attitudes, or openness to anti-LG behaviors. Finally, the fit of the full y saturated model was compared to that of an alternative trimmed model that eliminated the non-si gnificant direct paths (a) from SDO and traditional gender role attitudes to openness, (b) from SDO and openness to anti-LG attitudes, and (c) from RWA and SDO to an ti-LG behaviors (see Figure 4 for the trimmed model). The goodness of fit indices for this model were above the acceptable cut offs and nearly identical to those obt ained from the original model (GFI = .98; IFI = 1.00; CFI = 1.00; NFI = .97). The trimmed model explained 67% of the variance in anti-LG attitudes, 33% of the variance in anti-LG behaviors, and 20% of the variance in openness. As indicated in Figures 3 and 4, the amount s of variance accounted for in endogenous variables as well as significance decisions a nd magnitudes of paths coefficients were similar across the fully saturated and trimmed models. Specifically, in the trimmed model the paths from RWA to openness increased from -.39 to -.45, from traditional gender role attitudes to anti-LG attitudes increased from .24 to .25, from RWA to an ti-LG attitudes decreased from .71 to .70, from traditional ge nder role attitudes to anti-LG behaviors increased from .29 to .32, from openness to anti -LG behaviors decreased from -.20 to .17, and from anti-LG attitudes to anti-LG behaviors decreased from .34 to .25. Thus, compared with the fully sa turated model, the trimmed m odel appears to be more parsimonious, but equally appropriate in expl aining the relationships among the variables of interest.

PAGE 72

62 Right-Wing Authoritarianism Traditional Gender Attitudes Social Dominance Orientation Openness to Experience Anti-LG Behaviors Anti-LG Attitudes .27 .35 -.45 .70 .25 .32 .25 -.18 Figure 4. Trimmed model depicting relati onships among variables of interest for men. Values reflect standardized coefficients. All paths depicted are significant at p < .05.

PAGE 73

63 CHAPTER 5 DISCUSSION Numerous studies have examined sepa rately the relations of right-wing authoritarianism (RWA), social dominance orientation (SDO), traditional gender role attitudes, and openness to experi ence with anti-lesbian and ga y (LG) attitudes. However, no empirical study has examined this set of va riables together to determine the relative strength of their associations with anti-LG attitudes. The present study addressed this gap and also extended the literature by examining th e relations of this set of variables with self-reported anti-LG behaviors in addition to anti-LG attitude s. Furthermore, this study provided the first test of ope nness to experience as a potentia l mediator of the links of RWA, SDO, and traditiona l gender role attitudes to anti-LG attitudes, and the first test of anti-LG attitudes as a mediator of the li nks of RWA, SDO, traditional gender role attitudes, and openness to anti-LG behaviors. By examining concomitantly the relations among RWA, SDO, traditional gender role att itudes, openness, anti-LG attitudes, and anti-LG behaviors, the present findings build on prior research to identify the strongest correlates (among individual difference variable s) of anti-LG attitudes and behaviors. Identifying the most critical correlates will be important in informing future research and interventions that aim to reduce anti-LG attitudes and behaviors. In order to reduce anti-LG attitudes and be haviors, it is important to identify the variables that are most strongl y related to these co nstructs. Of the individual difference variables examined in the pr esent study, RWA emerged as th e strongest unique correlate of anti-LG attitudes for women and men and al so had a substantial relation with anti-LG

PAGE 74

64 behaviors. More specifically, RWA had larg e zero-order correlations with anti-LG attitudes ( r = .79) for both women and men. Furt hermore, in path models, once interrelations among all exogenous variables were accounted for, RWA again had the strongest relation with anti-LG attitudes for both women ( = .78) and men ( = .71). Additionally, RWA had a signifi cant indirect relation to anti-LG behaviors through antiLG attitudes for both women and men. For men, anti-LG attitudes fully mediated the link of RWA to anti-LG behaviors. That is thr ough anti-LG attitudes, a significant indirect link of .24 was obtained between RWA and anti-LG behaviors. For women, anti-LG attitudes partially mediated the link of RWA to anti-LG behaviors. That is, through antiLG attitudes, a significant indirect link of .50 was obtained between RWA and anti-LG behaviors. Thus, overall, the role of RWA was strong and as expected. An unexpected relation also was found, however, between RWA a nd anti-LG behaviors for women. Although there was a significant positive zero-order correlation between RWA and anti-LG behaviors ( r = .34), a significant ne gative unique relation ( = -.33) emerged between RWA and anti-LG behaviors when all variables of interest were entered into the path model. Since anti-LG attitudes partially mediated the link of RWA to anti-LG behaviors, a portion of the positive link of RWA to an ti-LG behaviors occurred through anti-LG attitudes, but the additional unique direct link of RWA to an ti-LG behaviors was negative for women. One possible explanation for the negativ e unique link of RWA with anti-LG behaviors (after other vari ables were accounted for) emerges from the theoretical conceptualization of RWA. Theoretically, RWA is comprised of three attitudinal clusters

PAGE 75

65 including authoritarian submission, author itarian aggression, and conventionalism (Altemeyer, 1981). Altemeyer (1981) theorized that aggression is manifested as the predisposition to inflict phys ical, psychological, economic, or social harm. However, most studies of RWA have examined its li nk with prejudicial at titudes but not with prejudicial aggressive beha viors (e.g., Altemeyer, 1998, 2001; Heaven & St. Quintin, 2002; Whitley, 1999). Therefore, there is a lack of clarity about the link between RWA and aggressive behaviors, although such a link would be consistent with the conceptualization of the sec ond attitudinal cluster of authoritarian aggression. Additional empirical research is needed to determine th e degree to which persons with high levels of RWA are also aggressive. Attention should be paid to the type of aggression exhibited by persons high in RWA, and the range of situational variables that can impact aggressive behaviors. The present study suggests that af ter accounting for the indirect link of RWA to anti-LG behaviors through anti-LG attitude s, for women, RWA negatively relates to anti-LG behaviors. Perhaps this is because RW A also reflects an adherence to authorities and conventionalism that does not coincide wi th aggressive or even criminal behaviors such as writing graffiti, yelling insu lting comments, and hitting an LG person. Furthermore, the significant negative relationship with an ti-LG behaviors was only found with women and not men. Perhaps there are ge nder differences in the expression of high RWA, with women acting less physically a ggressive than men because physical aggression may contradict the conventi onal view of women, and adherence to conventional roles is an important part of RWA. Thus, future empirical research is needed to examine potential gender differen ces in the expression of the aggressive component of RWA.

PAGE 76

66 The second variable that was demonstrated to have an important relation with both anti-LG attitudes and behaviors was traditional gender role attitudes. Next to RWA, this variable had the strongest links with anti-L G attitudes in both zero order correlations for women ( r = .46) and men ( r = .50) and in the path model ( = .24) for both women and men. Furthermore, traditional ge nder role attitudes had the strongest links with anti-LG behaviors for both women and men. More specif ically, traditional ge nder role attitudes had large zero order corr elations with anti-LG behaviors for both women ( r = .47) and men ( r = .50). Furthermore, for women, traditiona l gender role attitude s had a significant indirect link (through the partial mediating role of anti-LG atti tudes) of .15 and additional unique direct link of .27 with anti-LG beha viors. For men, there was no significant mediation through anti-LG attitudes, but trad itional gender role attitudes had a unique direct link of .29 with anti-LG behaviors. Thus overall, traditional gender role attitudes emerged as having significant unique relations with anti-LG attit udes and behaviors for both women and men. These findings point to the important unique ro le of traditional gender role attitudes in shaping anti-LG attitudes and behaviors. In addition to the role of RWA and trad itional gender role at titudes in anti-LG attitudes and behaviors, the present findings provide furt her support for the connection between anti-LG attitudes and behaviors. More specifically, the path analysis showed that the relation between anti-LG a ttitudes and self-reported be haviors was strong for women ( = .64) and moderate for men ( = .34). Perhaps situational va riables play a larger role in shaping men’s anti-LG behaviors, whereas attitudes seem fairly strongly related to self-reported behaviors for women. Also, Fr anklin (2000) found that the gender gap in anti-LG behaviors grew wider as the specifi c behaviors became more violent, with

PAGE 77

67 women reporting much lower levels of more violent behaviors than men. Perhaps as antiLG behaviors become more violent, the relation between anti-LG attitudes and behaviors decreases, particularly for men because other individual difference and situational variables play a more important role in shap ing violent behavior. Future research could explore if the relation between anti-LG attit udes and behaviors is different for persons engaging in non-violent versus violent anti-L G behaviors. Future research could also examine the magnitude of gender difference in the relation of anti-LG attitudes across various levels of violen ce of those behaviors. The role of anti-LG attitudes in shaping the links of individua l difference variables to anti-LG behaviors was supported. More sp ecifically, for wome n, anti-LG attitudes partially mediated the links of RWA (indirect link of .49) and traditional gender role attitudes (indirect link of .14) to anti-LG be haviors and fully mediated the link of SDO (indirect link of -.09) to anti-LG behaviors. Similarly, for men, anti-LG attitudes fully mediated the link of RWA (indirect link of .25) to anti-LG behaviors. This shows that the strongest correlates of anti-LG behaviors had indirect links to anti-LG behaviors through anti-LG attitudes. This pattern of findings demonstrates the importance of anti-LG attitudes in understanding anti -LG behaviors. Longitudinal a nd experimental studies can build on the present findings and evaluate the direction a nd causal role of anti-LG attitudes in shaping anti-LG behaviors. Futu re research is needed to examine whether lowering anti-LG attitudes does indeed lead to a reduction of anti-LG behaviors. One important pattern in the present findi ngs is the apparent minimal role of openness to experience in anti-LG attitudes and be haviors. In relation to anti-LG attitudes and behaviors, openness had small zero-or der correlations for women with anti-LG

PAGE 78

68 attitudes ( r = -.23) and anti-LG behaviors ( r = -.23) and moderate zero-order correlations for men with anti-LG attitudes ( r = -.37) and anti-LG behaviors( r = -.37). However, once all predictor variables were entered into th e path model, the rela tion of openness to antiLG attitudes was not significant for men, and was minimal and in the opposite direction than expected for women ( = .12). Perhaps, once the shared variance of openness with the other variables is taken in to account, the positiv e relation with anti -LG attitudes could reflect a greater openness and willingness to report negative attitudes. Similarly, in the path model, openness was not significantly un iquely related with anti-LG behaviors for women, but was uniquely relate d to anti-LG behaviors ( = -.20) for men (in the expected direction). Finally, the role of openness as a mediator of the relations of RWA, SDO, and traditional gender role attitudes to anti-LG at titudes was not supported. In fact, not only did openness not act as a mediator, after a ccounting for the roles of RWA, SDO, and traditional gender role attitudes, the links of openness to anti-LG attitudes and behaviors was non-significant or minimal for both wome n and men. Thus, in the context of the other individual difference variables exam ined in the present study, openness to experience did not appear to play an importa nt unique role in anti-LG attitudes or behaviors. Similar to openness, after accounting for the role of RWA and traditional gender role attitudes, the links of SDO to anti-LG attitudes and behaviors was minimal. For men there was no significant zero-order correlati on between SDO and anti-LG attitudes and there was a small positive relation be tween SDO and anti-LG behaviors ( r = .24). This small relation disappeared, however, once all of the variables were entered into the path model. For women, there were small signi ficant zero-order correlations between SDO

PAGE 79

69 and anti-LG attitudes ( r = .19) and behaviors ( r = .22). Once in the path model, SDO and anti-LG attitudes ( = -.14) had a small relation in th e unexpected direction. Additionally, for women, anti-LG attitudes mediated the relation between SDO a nd anti-LG behaviors and a small indirect effect of -.09 was found (i n an unexpected direction). It was expected that persons with high levels of SDO w ould report greater an ti-LG attitudes and behaviors because they would view LG pers ons as an out-group. These relations were indeed demonstrated through the positive zero order correlations. In the context of variance accounted for in anti-LG attitudes a nd behaviors by RWA, traditional gender role attitudes, and openness, however, S DO was related negatively to the remaining variance in anti-LG attitudes and behaviors. Perhaps, the observed positive correlations of SDO with anti-LG attitudes and behavior, in part, reflected the correlation with SDO with RWA, traditional gender role attitudes, and openness. Overall, the link of SDO with anti-LG attitudes and behaviors was minimal, especially when other individual difference variables were considered. T hus, the present findings sugge st that future empirical research and interventions should be aimed at RWA and traditional gender role attitudes, which demonstrated the strongest and most consistent relations with anti-LG attitudes and behaviors. An additional important pattern in the pres ent findings is that there were some differences in the pattern and magnitude of in tercorrelations and medi ational relations for women and men. Previous rese arch has consistently found gender differences in the levels of anti-LG attitudes and behaviors; with men holding more ne gative attitudes and reporting more negative behaviors than do women (e.g., Herek, 2002; Kite, 1984). In addition, Basow and Johnson (2000) found a different pattern in relations among

PAGE 80

70 predictors of anti-LG attitudes for their sample of college women than was found by Theodore and Basow (2000) for their sample of college men. Furthermore, Whitley and Egisdttir (2000) found that gende r of participants had a uniqu e link to anti-LG attitudes in their path model that also examined RWA, SDO, and attitudes toward women. However, studies that have examined anti-L G attitudes and behaviors typically have not reported their findings separately by gender (e.g., Roderick et al., 1998; Whitley & Lee, 2000). Thus, it is not clear whethe r the gender differences in th e pattern of interrelations found in the present study also existed in previous samples. Future researchers should explicitly explore the patter n of interrelations by gende r and report their findings separately for women and men in order to evaluate the replicability of the present findings and ultimately tailor interventions most appropriately to each gender if needed. The present findings must be interpreted in light of a number of limitations. For example, the current sample consisted of student s, most of whom were in their first year of college, White, and identified as middle or upper middle class. Thus, results from the present study may not generalize to the general population, which includes greater diversity in terms of age, leve l of education, and social class. Furthermore, given that the present sample did not target convicted cr iminals, the results may not generalize to persons who have been convicted of hate crimes against LG persons. Therefore, future research is needed to examine the replicabil ity of the present results in a variety of samples in order to capture the broader dive rsity of the general popul ation. Such research can also explore the extent to which the pa ttern of gender differences in interrelations found in the present study will ge neralize to other samples.

PAGE 81

71 Another limitation of the pr esent study reflects needs for advancement in the state of the art of measuring anti-LG behaviors. Although the present st udy utilized the most appropriate measure of anti-LG behaviors avai lable in the literature, such measurement can be improved by addressing a number of limitations. First, although SRBS-R scores were not correlated significantly with social de sirability, it is possi ble that participants did not report the full extent of negative behavi ors that they have enga ged in, or that they are not consciously aware of a ll of their negative behaviors. For example, a person could act rudely to an LG person by creating more personal distance or avoiding eye contact, yet not be conscious of such actions. In addi tion, the fact that the SRBS-R assessed only overt negative behaviors and crimes (e.g., ve rbal harassment, writing graffiti, hitting) may have contributed to the restriction in ra nge and floor effect of SRBS-R scores (see Tables 3 and 4), with women scoring a mean of 1.28 and men scoring a mean of 1.42, when the scale midpoint is 2.50. One way to a ddress these limitations is to integrate assessment of a continuum of behaviors from overtly negative to overtly positive (which might include items assessing helpful behavi ors such as attending a pride parade, befriending an LG person, and supporting legisl ation for increased LG rights). Using a more inclusive measure of hostile and affirming behaviors would allow for a fuller exploration of the links of RWA, SDO, traditional gender role attitudes, openness, and anti-LG attitudes to behavi ors toward LG persons. Another important area for advancing understanding of behaviors toward LG persons is to build on the present findings regarding the roles of individual difference variables by integrating attention to situational variables, such as contact with LG persons and youthful thrill seeking, which also have been linked to anti-LG behaviors (Franklin,

PAGE 82

72 1998; Schope, 2000). In the presen t study, individual difference variables accounted for a much greater percentage of the variance of anti-LG attitudes than behaviors. Thus, attention to other potential co rrelates of anti-LG behaviors is warranted and data from such studies can inform the development and evaluation of interventions aimed at reducing anti-LG behaviors. By advancing scientific unde rstanding of the predictors of negative attitudes and behaviors towards LG persons, the present study can serve as groundwork for developing theoretically and empirically based interventi ons that aim to reduce anti-LG attitudes and behaviors. This study points to the impor tance of RWA and traditional gender role attitudes in predicting anti-LG attitudes and behaviors of college women and men. The present data raise questions, however, about the roles of SDO and openness in anti-LG attitudes and behaviors. Base d on these results, in terventions aiming to prevent anti-LG attitudes and behaviors could focus limited time and resources specifically on modifying RWA and traditional gender role attitudes. Since persons with hi gh levels of RWA are very submissive to authority figures, a possible intervention c ould involve legitimate authority figures (e.g., university presidents, religious leaders, a nd elected officials) voicing affirming-LG attitudes and behaviors. Additionally, to combat traditional gender role attitudes, interventions could focu s on explaining the range of possible gender expressions and restrictions that traditional gend er role attitudes place on individuals. If these interventions are successful at lowering RWA and traditional gender role attitudes, then it can also be assessed if anti-L G attitudes and behaviors are also reduced. Evaluating the impact of such interventions would also test th e causal direction of relations among RWA, traditiona l gender role attitudes, antiLG attitudes, and anti-LG

PAGE 83

73 behaviors. Thus, the present findings can provide the basis for such research and ultimately inform prevention and educational interventions aimed at reducing anti-LG attitudes and behaviors.

PAGE 84

74 LIST OF REFERENCES Adorno, T. W., Frenkel-Brunswik, E., Levi nson, D. J., & Sanford, R. N. (1950). The authoritarian personality. New York: Harper & Row. Altemeyer, B. (1981). Right-wing aut horitarianism. Winnipeg, Canada: University of Manitoba Press. Altemeyer, B. (1988). Enemies of freedom: Understanding right-wing aut horitarianism. San Francisco: Jossey-Bass. Altemeyer, B. (1996). The authoritarian specter Cambridge, MA: Harvard University Press. Altemeyer, B. (1998). The othe r “authoritarian personality.” Advances in Experimental Social Psychology, 30, 47-91. Altemeyer, B. (2001). Changes in attitudes towards homosexuals. Journal of Homosexuality, 42(2), 63-75. Altemeyer, B., & Hunsberger, B. (1992). Au thoritarianism, religious fundamentalism, quest, and prejudice. The International Journal for the Psychology of Religions, 2(2), 113-133. Arbuckle, J. L. (2003). Amos (Version 5.0) [Computer software]. Chicago, IL: SmallWaters Corporation. Baron, M. R. & Kenny, D. A. (1986). The mode rator-mediator variab le distinction in social psychological research: Conceptual, strategic, and statis tical considerations. Journal of Personality and Social Psychology, 51, 1173-1182. Basow, S. A., & Johnson, K. (2000). Predic tors of homophobia in female college students. Sex Role, 42(5/6), 391-403. Bem, S. L. (1974). The measur ement of psychological androgyny. Journal of Consulting and Clinical Psychology, 42, 155-162. Bernat, J. A., Calhoun, K. S., Adams, H. E., & Zeichner, A. (2001). Homophobia and physical aggression toward homosex ual and heterosexual individuals. Journal of Abnormal Psychology, 110 (1), 179-187.

PAGE 85

75 Butler, J. C. (2000). Personality and emotiona l correlates of right-w ing authoritarianism. Social Behavior and Personality, 28(1), 1-14. Chan, K., Rounds, J., & Drasgow, F. (2000). The relation between vocational interests and the motivation to lead. Journal of Vocational Behavior, 57, 226-245. Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). Hillsdale, NJ: Lawrence Earlbaum Associates. Cohen, J., & Cohen, P. (1983). Applied multiple regrssion/correlation analysis for the behavioral sciences (2nd ed.). Hillsdale, NJ: Erlbaum. Conn, S. R., & Rieke, M. L. (1994). The 16PF fifth edition technical manual Champaign, IL: Institute for Pe rsonality and Ability Testing. Costa, P. T., & McCrae, R. R. (1985). The NEO Personality Inventory manual. Odessa, FL: Psychological Assessment Resources. Costa, P. T., & McCrae, R. R. (1992). The NEO Personality Inventory (Revised). Odessa, FL: Psychological Assessment Resources. Cullen, J. M., Wright, L. W., & Alessandri, M. (2002). The personal ity variable openness to experience as it relates to homophobia. Journal of Homosexuality, 42, 119-133. D’Augelli, A. R. (1989). Lesbians’ and gay men’s experiences of discrimination and harassment in a university community. American Journal of Community Psychology, 17, 317-321. Diaz, R. M., Ayala, G., Bein, E., Henne, J., & Marin, V. (2001). The impact of homophobia, poverty, and racism on the ment al health of gay and bisexual latino men: Findings from 3 US cities. American Journal of Public Health, 91, 927-932. DiPlacido, J. (1998). Minority stress among lesbians, gay men, and bisexuals: A consequence of heterosexism and homophobia. In G. M. Herek (Ed.), Stigma and sexual orientation: Understanding pr ejudice against lesbians, gay men, and bisexuals (pp. 138-159). Thousand Oaks, CA: Sage. Eisler, R. M., & Blalock, J. A. (1991). Mascu line gender role stress: Implications for the assessment of men. Clinical Psychology Review, 11, 45-60. Eisler, R. M., Skidmore, J. R. (1987). Masc uline gender role stress: Scale development and component factors in the appr aisal of stressful situations. Behavioral Modification, 11, 123-136. Ekehammar, B., & Akrami, N. (2003). The rela tion between personality and prejudice: A variableand a personcentered approach. European Journal of Personality, 17, 435-448.

PAGE 86

76 Franklin, K. (1998). Unassuming motivations: C onextualizing the narratives of antigay assailants. In G. M. Herek (Ed.), Stigma and sexual orientation: Understanding prejudice against lesbians, gay men, and bisexuals (pp. 39-61). Thousand Oaks, CA: Sage. Franklin, K. (2000). Antigay behaviors among young adults: Prevalence, patterns, and motivators in a noncriminal population. Journal of Interpersonal Violence, 15, 339362. Goldberg, L. R. (1999). A broad-bandwi dth, public domain, personality inventory measuring the lower-level facets of several five-factor models. In I. Mervielde, I. Deary, F. De Fruyt, & F. Ostendorf (Eds.), Personality Psychology in Europe Vol. 7 (pp. 7-28). Tilburg, The Netherland s: Tilburg University Press. Haddock, G., Zanna, M. P., & Esses, V. (1993) Assessing the struct ure of prejudicial attitudes: The case of attitudes toward homosexuals. Journal of Personality and Social Psychology, 65, 1105-1118. Haddock, G., & Zanna, M. P. (1998). Authoritar ianism, values, and the favorability and structure of antigay attitude s. In G. M. Herek (Ed.), Stigma and sexual orientation: Understanding prejudice against le sbians, gay men, and bisexuals (pp. 82-107). Thousand Oaks, CA: Sage. Heaven, P. C. L., & Bucci, S. (2001). Ri ght-wing authoritarianis m, social dominance orientation and personality: An an alysis using the IPIP measure. European Journal of Personality, 15, 49-56. Heaven, P. C. L., & St. Quintin, D. (2002). Pe rsonality factors predic t racial prejudice. Personality and Individual Differences, 34, 625-634. Herek, G. M. (1984). Beyond “homophobia”: A social psychological perspective on attitudes towards le sbians and gay men. Journal of Homosexuality, 10, 1-21. Herek, G. M. (1986). On heterosexual masculin ity: Some psychical consequences of the social construction of gender and sexuality. American Behavioral Scientists, 29, 563-577. Herek, G. M. (1988). Heterosexuals’ attitudes toward lesbians and gay men: Correlates and gender differences. The Journal of Sex Research, 25, 451-477. Herek, G. M. (2000). The psychology of sexual prejudice. Current Directions in Psychological Science, 9(1), 19-22. Herek, G. M., & Capitanio, J. P. (1995). Black heterosexuals’ att itudes toward lesbian and gay men in the United States. The Journal of Sex Research, 32, 95-101. Hill, P. C., & Wood, R. W., Jr., (Eds.). (1999). Measures of Religiosity. Birmingham, AL: Religious Education Press.

PAGE 87

77 Hudson, W. W., & Ricketts, W. A. (1980) A strategy for the measurement of homophobia. Journal of Homosexuality, 5, 357-372. Hunsberger, B. (1995). Religion and prejudice: The role of religious fundamentalism, quest, and right-wing authoritarianism. Journal of Social Issues, 51 113-123. International Personality Item Pool (2001). A scientific collaboration for the development of advanced measures of personality traits and other individual differences. Retrieved October 1, 2005 from http://ipip.ori.org/ John, O. P., Donahue, E., & Kentle, R. L. (1992). The Big Five Inventory: Versions 4a and 54. Technical Report, Institute of Pe rsonality and Social and Research, University of California, Berkley, CA. Johnson, M. E., Brems, C., & Alfod-Keati ng, P., (1997). Personality correlates of homophobia. Journal of Homosexuality, 34, 57-68. Kerns, J. G., & Fine, M. A. (1994). The rela tion between gender and negative attitudes toward gay men and lesbians: Do gender role attitudes mediate this relation? Sex Roles, 31, 297-307. Kite, M. E. (1984). Sex differences in att itudes toward LG persons: A meta-analytic review. Journal of Homosexuality, 10, 69-81. Kite, M. E. (1992). Individual differences in males’ reactions to gay males and lesbians. Journal of Applied Social Psychology, 22, 1222-1239. Kite, M. E., & Deaux, K. (1986). Attitudes toward homosexuality: Assessment and behavioral consequences. Basic and Applied Social Psychology, 7 (2), 137-162. Kite, M. E., & Deaux, K. (1987). Gender belief systems: Homosexuality and the implicit inversion theory. Psychology of Women Quarterly, 11, 83-96. Kite, M. E., & Whitley, B. E., Jr. (1998). Do heterosexual women and men differ in their attitudes towards homosexual ity? A conceptual and met hodological analysis. In G. M. Herek (Ed.), Stigma and sexual orientati on: Understanding prejudice against lesbians, gay men, and bisexuals (pp. 39-61). Thousand Oaks, CA: Sage. Kohn, P.M. (1972). The Authoritarianism-Rebelli on scale: A balanced F Scale with leftwing reversals. Sociometry, 35, 176-189. Kurpius, S. E. R., & Lucart, A. L., (2 000). Military and civilian undergraduates: Attitudes toward women, masculinity, and authoritarianism. Sex Roles, 43, 255265. Laner, M. R., & Laner, R. H. (1979). Persona l style or sexual pr eference: Why gay men are disliked. International Review of Modern Sociology, 9, 215-228.

PAGE 88

78 Laner, M. R., & Laner, R. H. (1980). Sexual preference or personal style? Why lesbians are disliked. Journal of Homosexuality, 5, 339-356. LaMar, L., and Kite, M. E. (1998). Sex differe nces in attitudes towa rd lesbians and gay men: A multidimensional perspective. Journal of Sex Research, 35, 189-196. Lefkowitz, E. S., Shearer, C. L., Gillen, M. M, & Espinosa-Hernandez, M. G. (2003). Measuring female role norms in male and female emerging adults. Unpublished manuscript. MacDonald, A. P., Jr. (1974). Identificat ion and measurement of multidimensional attitudes toward equali ty between the sexes. Journal of Homosexuality, 1, 165-182. Mays, V. M., & Chochran, S. D. (2001). Me ntal health correlates of perceived discrimination among lesbian, gay and bi sexual adults in the United States. American Journal of Public Health, 91, 1869-1876. McCrae, R. R. (1996). Social conseq uences of experimental openness. Psychological Bulletin, 120, 323-337. McCrae, R. R., & Costa, P. T. (2003). Personality in adulthood: A five-factor theory perspective. New York: Guilford Press. McHugh, M. C. & Frieze, I. H. (1997). The measurement of gender-role attitudes: A review and commentary. Psychology of Women Quarterly, 21, 1-16. O’Neil, J. M. (1981). Patterns of gender ro le conflict and strain: Sexism and fear of femininity in men’s lives. The Personnel and Guidance Journal, 60, 203-210. O’Neil, J. M., Helms, B., Gable, R., Davi d, L., & Wrightsman, L. (1986). Gender Role Conflict Scale: College men’s fear of femininity. Sex Roles, 14, 335-350. Patel, S., Long, T. E., McCammon, S. L., & Wuensch, K. L. (1995). Personality and emotional correlates of self-re ported antigay behaviors. Journal of Interpersonal Violence, 10, 354-366. Paulhus, D. L. (1994). Balanced Inventory of Desirabl e Responding: Reference manual for BIDR Version 6. Unpublished manuscript, Department of Psychology, University of British Columbia, Vancouver, BC. Peterson, B. E., Smirles, K. A., & We ntworth, P. A. (1997). Generativity and authoritarianism: Implications for pe rsonality, political involvement, and parenting. Journal of Personality and Social Psychology, 72, 1202-1216. Pleck, J. H., Sonenstein, F. L., & Ku, L. C. (1993). Masculinity id eology: Its impact on adolescent males’ hete rosexual relationships. Journal of Social Issues, 49 (3), 1129.

PAGE 89

79 Pratto, F., Sidanius, J., Stallworth, L. M ., & Malle, B. F., (1994). Social dominance orientation: A personality variable predicting social and political attitudes. Journal of Personality and Social Psychology, 67, 741-763. Rokeach, M. (1956). Political and religi ous dogmatism: An alternative to the authoritarian personality. Psychological Monographs, 70, 1-43. Rokeach, M. (1960). The open and closed mind. New York: Basic Books. Roderick, T., McCammon, S. L., Long, T. E., A llred, L. J. (1998). Be havioral aspects of homonegativity. Journal of Homosexuality, 36, 79-87. Schope, R. D. & Eliason, M. J. (2000). Thinking versus acting: Assessi ng the relationship between heterosexual attitudes a nd behaviors towards homosexuals. Journal of Gay & Lesbian Social Services, 11(4), 69-92. Sidanius, J. & Pratto, F. (1999). Social dominance: An intergroup theory of social hierarch and oppression. Cambridge: Cambridge University Press. Sinn, J. S., (1997). The predictive and discri minate validity of masculinity ideology. Journal of Research in Personality, 31, 117-135. Smith, E.R., Ferree, M. M., & Miller, F. D. (1975). A short scale of attitudes toward feminism. Representative Research in Social Psychology, 6, 51-58 Spence, J. T., & Helmreich, R. L. (1978). Masculinity and femininity: Their psychological dimensions, correlates, and antecedents. Austin: University of Texas Press. Spence, J. T., Helmreich, R. L., & Stapp, J. ( 1973). Ratings of self and peers on sex-role attributes and their relati ons to self-esteem and conceptions of masculinity and femininity. Journal of Personality and Social Psychology, 32, 29-39. Swim, J. K., Ferguson, M. J., & Hyers, L. L. (1999). Avoiding stigma by association: Subtle prejudice against lesbians in the form of social distancing. Basic and Applied Social Psychology, 21 (1), 61-68. Theodore, P. S., & Basow, S. A. (2000) Heterosexual masculinity and homophobia: A reaction to the self? Journal of Homosexuality, 40, 31-48. Thompson, E. H., & Pleck, J. H. (1986). Th e structure of the male role norms. American Behavioral Scientist, 29, 531-534. Thompson, E. H., Pleck, J. H., & Ferrera, D. L. (1992). Men and masculinities: Scales for masculinity ideology and ma sculinity-related constructs. Sex Roles, 27, 573608.

PAGE 90

80 Tokar, D. M., Fischer, A. R., Schaub, M., Moradi, B., (2000). Masculine gender roles and counseling-related variables: Links with and mediation by personality. Journal of Counseling Psychology, 47, 380-393. Whitley, B. E., Jr. (1987) The relationship of sex-role orientation to heterosexuals’ attitudes towards LG persons. Sex Roles, 17, 103-113. Whitley, B. E., Jr. (1999) Right-wing authorita rianism, social dominance orientation, and prejudice. Journal of Personality and Social Psychology, 77, 126-134. Whitley, B. E., Jr. (2001). Gender-role vari ables and attitudes toward homosexuality. Sex Roles, 45, 691-721. Whitley, B. E., Jr., & Egisdttir, S. (2000). The gender belief system, authoritarianism, social dominance orientation, and heterosexuals’ attitudes toward lesbian and gay men. Sex Roles, 42(11/12), 947-967. Whitley, B. E., & Lee, S. E., (2000). The rela tionship of authoritarianism and related constructs to attitude s towards homosexuality. Journal of Applied Social Psychology, 30, 144-170. Wilson, G. D., & Patterson, J. R. (1968) A new measure of conservatism. British Journal of Social and Clinical Psychology, 7, 264-269.

PAGE 91

81 BIOGRAPHICAL SKETCH I was born and raised in Silver Sprin g, Maryland. I graduated magna cum laude with a Bachelor of Science in psychology from the Universi ty of Maryland in 2002. After graduating I spent a year working with autistic children and traveling overseas. In August 2003 I moved to Gainesville to enter into Un iversity of Florida’s Counseling Psychology program. I am currently in my third year of the doctoral program and hope to earn my Ph.D. in a few years.


Permanent Link: http://ufdc.ufl.edu/UFE0013329/00001

Material Information

Title: Mediating Role of Openness in the Links of Right-Wing Authoritarianism, Social Dominance Orientation and Traditional Gender Role Attitudes to Anti-Lesbian and Gay Attitudes and Behaviors
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0013329:00001

Permanent Link: http://ufdc.ufl.edu/UFE0013329/00001

Material Information

Title: Mediating Role of Openness in the Links of Right-Wing Authoritarianism, Social Dominance Orientation and Traditional Gender Role Attitudes to Anti-Lesbian and Gay Attitudes and Behaviors
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0013329:00001


This item has the following downloads:


Full Text












MEDIATING ROLE OF OPENNESS IN THE LINKS OF RIGHT-WING
AUTHORITARIANISM, SOCIAL DOMINANCE ORIENTATION, AND
TRADITIONAL GENDER ROLE ATTITUDES TO ANTI-LESBIAN AND GAY
ATTITUDES AND BEHAVIORS















By

MELINDA BETH GOODMAN


A THESIS PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
MASTER OF SCIENCE

UNIVERSITY OF FLORIDA


2005

































Copyright 2005

by

Melinda Beth Goodman

































This thesis is dedicated to the good people at Starbucks who let me work there for many
hours and whose chai tea helped me through the process. In addition, this document is
dedicated to my cat Calvin who demonstrated his love for me and the literature as noted
by the countless hours spent sitting on my lap and the Journal of Counseling Psychology.















ACKNOWLEDGMENTS

First of all I would like to thank my advisor and committee chair, Dr. Bonnie

Moradi, for all her guidance, support, and encouragement during every step of the way. I

am thankful for the assistance given to me by my committee members, Dr. Martin

Heesacker and Dr. Kendal Broad, and by my research assistants Ashley Stripling and

Korinne Chiu. In addition, I would like to thank my parents for all of their support during

this project and my entire life. I could never have reached this point without their love,

encouragement, and devotion. Finally, I would like to thank my friends from Gainesville

and all of my family and friends from home who have continually supported me despite

my move to Florida.
















TABLE OF CONTENTS



A C K N O W L E D G M E N T S ................................................................................................. iv

L IST O F T A B L E S .......... .......... ..................................................... .... .... ..... vii

LIST OF FIGURES .............. .................... ......................... viii

ABSTRACT .............. .......................................... ix

CHAPTER

1 IN TR O D U C TIO N ......................................................................... .... .. ........

2 REVIEW OF LITERATURE .................................................... ..... .. ............... 11

R ight-W ing A uthoritarianism ........................................................... ... ............ 12
Social D om finance O rientation........................................................... ............... 17
Traditional Gender Role Attitudes................................................ 20
Openness to Experience: An Underlying Construct ..............................................24
Need to Study Anti-Lesbian and Gay Behaviors............................... ...............30
Purpose of Study .................................................................... .. ...... ........ 37

3 M E T H O D S ......................................................... ................ 3 9

P a rtic ip a n ts ............................................................................................................ 3 9
P procedures ............................................................................................... ......... 40
In stru m e n ts ............................................................................................................ 4 0
C riterion V ariables ...........................................................40
P redictor V ariables ...........................................................43

4 R E S U L T S .............................................................................4 9

Gender Differences .................................. ............................ ... .......49
D descriptive Statistics .............................................................50
Test for O order Effects ..................................................................52
Intercorrelations Among Variables of Interest: Hypothesis 1 ..................................53
Tests of Mediation: Hypotheses 2 and 3 .............. ..................... .. .......... 54

5 DISCU SSION ................ ................ ......................... 63



v









L IST O F R E F E R E N C E S ...................................... .................................... ....................74

B IO G R A PH IC A L SK E T C H ...................................................................... ..................81
















LIST OF TABLES

Table p

1 Zero-Order Correlations Among Variables of Interest for Women.........................49

2 Zero-Order Correlations Among Variables of Interest for Men .............................49

3 Summary Statistics Among Variables of Interest for Women..............................50

4 Summary Statistics Among Variables of Interest for Men ...................................51
















LIST OF FIGURES


Figure page

1 Fully saturated model depicting relationships among variables of interest for
w om en ............................................................................................56

2 Trimmed model depicting relationships among variables of interest for women....58

3 Fully saturated model depicting relationships among variables of interest for
m en ................................................................................60

4 Trimmed model depicting relationships among variables of interest for men.........62















Abstract of Thesis Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Master of Science

MEDIATING ROLE OF OPENNESS IN THE LINKS OF RIGHT-WING
AUTHORITARIANISM, SOCIAL DOMINANCE ORIENTATION, AND
TRADITIONAL GENDER ROLE ATTITUDES TO ANTI-LESBIAN AND GAY
ATTITUDES AND BEHAVIORS

By

Melinda Beth Goodman

December 2005

Chair: Bonnie Moradi
Major Department: Psychology

The present study examined concomitantly the links of right-wing authoritarianism

(RWA), social dominance orientation (SDO), and traditional gender role attitudes to anti-

lesbian and gay (LG) attitudes. In addition, the present study examined openness to

experience as a mediator that might explain, partially or fully, the links of the

aforementioned correlates to anti-LG attitudes. Finally, this study examined the extent to

which RWA, SDO, traditional gender role attitudes, and openness are related to anti-LG

behaviors directly or indirectly through anti-LG attitudes. Results showed that RWA and

traditional gender role attitudes were the strongest predictors of anti-LG attitudes and

behaviors. The role of openness as a mediator of the relations of RWA, SDO, and

traditional gender role attitudes to anti-LG attitudes was not supported. Finally, anti-LG

attitudes did indeed partially mediate the relations of RWA, traditional gender role









attitudes, and SDO to anti-LG behaviors. Future directions for research and implications

for practice are discussed.














CHAPTER 1
INTRODUCTION

A broad range of variables (e.g., age, income, education, gender, and interpersonal

contact) has been shown to correlate with anti-lesbian and gay (LG) attitudes (e.g., Herek

& Capitanio, 1995, Johnson, Brems, & Alford-Keating, 1997). Within this body of

literature, some of the most consistent correlates of anti-LG attitudes have been

personality-related variables. More specifically, right-wing authoritarianism (RWA),

social dominance orientation (SDO), and traditional gender role attitudes have been

linked strongly and consistently to anti-LG attitudes. Although studies that focus on

examining the link of each of these variables and anti-LG attitudes (e.g., Altemeyer &

Hunsberger, 1992; Basow & Johnson, 2000; Whitley & Egisd6ttir, 2000) are informative,

theoretical and empirical scholarship is needed to integrate findings across such studies

and explain the shared links of personality-related variables to anti-LG attitudes. More

specifically, a question that needs to be examined is what, if any, underlying construct or

set of constructs explains the common links of RWA, SDO, and traditional gender role

attitudes to anti-LG attitudes. Conceptual definitions of each of the aforementioned

constructs and prior literature on their links to anti-LG attitudes provide a basis for

exploring this question. Therefore, a brief overview of the literature associated with each

construct and its link to anti-LG attitudes is provided next.

Altemeyer (1981) developed the concept of right-wing authoritarianism (RWA),

based on social learning theory, to address a number of limitations (e.g., lack of

conceptual clarity and falsifiability and psychometric problems in operationalization) in









Adorno, Frenkel-Brunswik, Levinson, and Sanford's (1950) earlier psychodynamic

conception of authoritarianism. Altemeyer (1981) defined RWA as comprised of three

attitudinal clusters. The first component, authoritarian submission, is defined as a "high

degree of submission to authorities who are perceived to be established and legitimate in

the society in which one lives" (p. 148). The second component is authoritarian

aggression and reflects "general aggression, directed at various persons, which is

perceived to be sanctioned by established authorities" (p. 148). Aggressiveness in this

context is the predisposition to do physical, psychological, economic, or social harm to

others. Persons with high levels of RWA tend to be aggressive or support aggressiveness

toward targets that they believe are socially stigmatized. For example, persons high in

RWA might support police violence against suspects in order to obtain confessions. The

final component of RWA is conventionalism defined as a "high degree of adherence to

the social conventions which are perceived to be endorsed by society and its established

authorities" (p. 148). Altemeyer (1981) described "established and legitimate authorities"

to include individuals in society who are usually considered to be legal or moral

authorities such as parents for younger persons, religious officials, heads of government,

police officers, judges, and military officials.

Altemeyer's (1981, 1988) theoretical conceptualization of right-wing

authoritarianism as characterized by high degrees of submission to authorities,

aggressiveness toward out-groups, and adherence to conventions perceived to be

endorsed by society and its authorities, suggests that people with high levels of RWA are

likely to exhibit anti-LG attitudes. More specifically, LG persons (a) are explicitly

condemned by some religious and political leaders, making LG persons an acceptable









target for prejudice and hostility, (b) represent a socially stigmatized out-group relative to

heterosexual women and men, and (c) are perceived as advocates for political change and

challenges to social conventions of traditional gender roles. Indeed, a number of studies

have found significant positive links between RWA and anti-LG attitudes (e.g.,

Altemeyer, 1981, 1998, 2001; Basow & Johnson, 2000; Whitley & Lee, 2000) indicating

that persons with high levels of RWA express more anti-LG attitudes than do persons

with low levels of RWA.

In addition to RWA, social dominance orientation (SDO) also has been linked

consistently with anti-LG attitudes. Pratto, Sidanius, Stallworth, and Malle (1994)

defined SDO as the "extent to which one desires that one's in-group dominate and be

superior to out-groups" (p 724). SDO taps a general attitudinal orientation toward

intergroup relations that reflects a preference for hierarchical rather than egalitarian

relations. More specifically, the desire to maintain the superior position of their in-groups

relative to out-groups motivates people high in SDO to accept hierarchy-legitimizing

myths that denigrate members of out-groups and enforce the status quo of their in-

groups' power position. Thus, persons high in SDO tend to have anti-LG attitudes

because LG persons have lower perceived social status than do heterosexuals (Whitley &

Lee, 2000). The empirical literature supports the expected link between SDO and anti-LG

attitudes. Across a number of studies, significant correlations have been found between

SDO and anti-LG attitudes (e.g., Altemeyer, 1998; Whitley, 1999; Whitley & Lee 2000)

such that higher SDO scores are related to higher levels of anti-LG attitudes.

Although RWA and SDO both include an element of prejudice toward out-groups,

Altemeyer (1988) argued that the two constructs are distinct in that SDO "does not have









the same psychological roots that previous studies have unearthed in right-wing

authoritarians" (p. 61). More specifically, a person high in RWA is very accepting of

traditional values and institutional authorities and is likely to follow the instructions of

those in positions of power and influence. By contrast, an individual high in SDO is not

motivated by a sense of duty or morality (Heaven & Bucci, 2001). Thus, although RWA

and SDO both can explain prejudicial beliefs and behaviors, such as anti-LG attitudes,

they do so in a complementary but not redundant fashion, such that each accounts for

unique variance in a variety of prejudice-related beliefs, behaviors, and traits (Altemeyer,

1998). Indeed, empirical evidence indicates that RWA and SDO are only minimally

correlated with each other (Altemeyer, 1998; Pratto et al., 1994) and the relationship of

SDO to prejudice is not accounted for by a joint relationship with RWA (Altemeyer,

1988; Whitley & Lee, 2000).

The final personality-related construct that has been linked consistently to anti-LG

attitudes is traditional gender role attitudes. Gender role attitudes are "beliefs about the

appropriate role activities for women and men" (McHugh & Frieze, 1997, p. 4).

Individuals who endorse traditional gender roles believe that women should follow

traditional female-role norms and that men should follow traditional male-role norms. A

commonly held stereotype about LG persons is that such persons violate traditional

gender roles, with gay men acting more feminine and lesbian women acting more

masculine than heterosexual men and women should, respectively (Kite & Deaux, 1987).

In addition, LG persons often are perceived to deviate from the cultural expectations of

the roles that men and women should perform in their lives, for example, by not getting

married and having children (Kite, 1984; Whitley, 1987). Thus, heterosexual persons









who hold traditional gender role attitudes are likely to express anti-LG attitudes because

LG persons are stereotyped as having cross-gender traits, roles, and physical

characteristics. Indeed numerous studies measuring constructs related to traditional

gender role attitudes (e.g., Attitudes Toward Women Scale, Sex Role Egalitarianism

Scale) have found that persons who hold more traditional gender role attitudes express

more anti-LG attitudes than do those with less traditional gender role attitudes (Basow &

Johnson, 2000; Herek, 1988; Whitley & Egisd6ttir, 2000).

Thus, conceptual definitions ofRWA, SDO, and traditional gender role attitudes

suggest that each of these variables is related to anti-LG attitudes and empirical evidence

supports such links. In addition to their shared links to anti-LG attitudes, RWA, SDO,

and traditional gender role attitudes each also has been linked conceptually and

empirically to openness to experience, a dimension of personality described in the Five

Factor Model. Openness, in turn, has been shown to be related to anti-LG attitudes. Thus,

openness to experience might be a mediating variable that provides a more parsimonious

explanation of and accounts for the links of RWA, SDO, and traditional gender role

attitudes to anti-LG attitudes.

Within the framework of the Five Factor Model, openness to experience is

conceptualized to reflect active imagination, aesthetic sensitivity, receptiveness to inner

feelings, preference for variety, intellectual curiosity, and independence of judgment

(McCrae & Costa, 2003). Additionally, lack of openness to experience characteristically

involves intolerance for ambiguity and lack of flexibility. A person who is high in

openness is likely to value intellectual matters, is nonconforming, introspective, has a

vivid imagination, and admits that what is right and wrong for one person may not be









applicable in other circumstances. On the other hand, a person scoring low in openness is

likely to be rigid and uncomfortable with complexities, have a narrow range of emotions,

favor conservative values, and judge others in conventional terms (McCrae & Costa,

2003). The characteristics associated with low levels of openness appear to overlap with

the characteristics associated with high levels of RWA, SDO, and traditional gender role

attitudes.

For example, persons high in RWA are expected to reject any thoughts or

experiences that are not sanctioned by authority figures, demonstrating a lack of

flexibility. Characteristically, people high in RWA are rigid, conforming, and have

conservative political values (Altemeyer, 1981, 1988). Indeed, RWA has been shown to

relate significantly and negatively to openness to experience with correlations ranging

from -.30 to -.57 (Altemeyer, 1996; Butler, 2000; Heaven & Bucci, 2001; Peterson,

Smirles, & Wentworth, 1997).

Similarly, a lack of openness to experience is apparent in SDO. Persons high in

SDO accept hierarchy-legitimizing myths in order to promote the superiority of one

group over another (Pratto et al., 1994). Instead of viewing other people as individuals,

persons high in SDO concentrate on the targets' out-group membership. Those high on

SDO characteristically have a lack of sympathy, empathy, and tolerance. The lack of

ability to view the complexities in the world and the narrow range of emotions that

characterize persons high in social dominance are consistent with theoretical definitions

of lack of openness to experience. Indeed, SDO has been shown to correlate significantly

and negatively (r = -.26 to -.28) with openness to experience (Heaven & Bucci, 2001;

Pratto et al., 1994).









Finally, persons who adhere to traditional gender role attitudes expect others to fit

into a relatively stable set of gender roles, traits, and physical attributes (e.g., Kite &

Whitley, 1998). This acceptance of traditional roles demonstrates a lack of tolerance for

ambiguity and variety within the gender spectrum. It is expected that persons who adhere

to traditional gender role attitudes would also be low in openness to experience. Although

the links between openness and adherence to traditional masculine and feminine gender

role norms have not been examined directly, extant data indicate a negative link between

constructs that are related to traditional gender role ideology and openness. For example,

among men, openness to experience was found to be significantly and negatively related

to masculine gender role conflict and masculine gender role stress with correlations of -

.38 and -.21, respectively (Tokar, Fischer, Schaub, & Moradi, 2000). Gender role conflict

is a "psychological state in which gender roles have negative consequences or impact on

the person or others" (O'Neil, 1981, p. 203). Men's gender role stress results from a

failure to meet cultural standards of masculinity (Eisler & Blalock, 1991). Although

men's gender role conflict and stress are not equivalent to traditional gender role

attitudes, there is significant theoretical overlap in that one must endorse traditional

gender roles in order to experience gender role conflict or stress. Indeed, gender role

conflict and stress are significantly correlated with traditional gender role attitudes

(Thompson, 1991, as cited in Thompson, Pleck, & Ferrera, 1992). Thus, it is expected

that traditional gender role attitudes will relate to low openness and the present study will

examine directly this proposition.

Taken together, the theoretical definitions and empirical evidence reviewed suggest

that lack of openness to experience might be a global construct that ties together the links









of RWA, SDO, and traditional gender role attitudes to anti-LG attitudes. Furthermore,

openness might be an underlying construct that explains, partially or fully, the links of the

aforementioned correlates to anti-LG attitudes. In fact, in addition to openness to

experience's links with RWA, SDO, and traditional gender attitudes, openness is also

significantly related to anti-LG attitudes. When examining the Five Factor Model of

personality, Cullen, Wright, and Alessandri (2002) found that openness accounted for

25% of the variance in anti-LG attitudes, and none of the other Five Factor Personality

traits were unique predictors. Due to the correlations of openness to the predictor

variables (i.e., RWA, SDO, traditional gender role attitudes) and anti-LG attitudes, it is

possible that openness will mediate the links of RWA, SDO, and traditional gender role

attitudes to anti-LG attitudes. In other words, RWA, SDO, and traditional gender

attitudes might be related to anti-LG attitudes through their shared link with openness.

Thus, the first general aim of the present study is to examine openness to experience as a

mediator of the links of RWA, SDO, and traditional gender role attitudes to anti-LG

attitudes.

Studying anti-LG attitudes alone, however, is not sufficient. An implicit reason for

studying anti-LG attitudes is that the expression of these attitudes as anti-LG behaviors

might have serious negative interpersonal and intrapersonal consequences for LG persons

and society at large. In fact, the prevalence of anti-LG behaviors (e.g., name-calling,

threats, assaults) is staggering. D'Augelli (1989) found that 50% of the gay and lesbian

college students surveyed reported having overheard disparaging comments, 26%

reported having experienced personal verbal insults, 26% reported having been

threatened with physical violence and 23% reported having been victims of assault.









Furthermore, such negative experiences have been linked to mental health-related

concerns (e.g., suicidal ideation, psychological distress) among LG persons (e.g., Diaz,

Ayala, Bein, Henne, & Marin, 2001; DiPlacido, 1998; Mays & Chochran, 2001). Thus,

there is a need to study anti-LG behaviors in addition to anti-LG attitudes.

Despite the considerable prevalence and potential negative consequences of anti-

LG behaviors and numerous calls for research examining such behaviors (Herek, 2000;

Kite & Whitley, 1998; Whitley, 2001), the majority of extant research has not attended to

predictors or correlates of anti-LG behaviors and has focused mostly on identifying

correlates of anti-LG attitudes. The limited research available shows that although hate

crimes and anti-LG behaviors are inevitably influenced by complex situational factors

(Franklin, 1998), anti-LG attitudes do indeed contribute to anti-LG behaviors. In three

studies that assessed self-reports of both anti-LG attitudes and behaviors, anti-LG

attitudes were a significant predictor of negative behaviors. Correlations between anti-LG

attitudes and behaviors toward LG persons ranged between .25 and .40 (Franklin, 2000;

Patel, Long, McCammon, & Wuensch, 1995; Whitley, 2001).

Research also demonstrates that some of the major correlates of anti-LG attitudes

are also related to anti-LG behaviors. For example, Haddock, Zanna, and Esses (1993)

found that for college students, RWA is related to support of increased budget cuts to an

LG student organization. Furthermore, Franklin (2000) and Whitley (2001) found that

endorsement of traditional gender role attitudes was related positively to anti-LG

behaviors. However, there is no research examining the correlations of SDO and

openness to experience to anti-LG behaviors. Thus, the second general aim of the present

study is to add to the literature by examining the extent to which psychological factors









that have been shown to be related to anti-LG attitudes are also related to anti-LG

behaviors. Also, it is not known if RWA, SDO, traditional gender roles, and openness to

experience will relate directly to anti-LG behaviors, or if their relationship is mediated by

anti-LG attitudes. Thus, the present study will examine whether anti-LG attitudes mediate

the relationships of RWA, SDO, and traditional gender roles and openness to experience

to anti-LG attitudes.

In light of the conceptual and empirical literature reviewed here, the aims of the

present study are to address the following gaps:

1. Examine openness to experience as a mediator of the links of RWA, SDO, and
traditional gender role attitudes to anti-LG attitudes.
2. Examine anti-LG attitudes as a mediator of the links of RWA, SDO, traditional
gender role attitudes, and openness to anti-LG behaviors.














CHAPTER 2
REVIEW OF LITERATURE

A broad range of variables (e.g., age, income, education, gender, and interpersonal

contact) has been shown to correlate with anti-lesbian and gay (LG) attitudes (e.g., Herek

& Capitanio, 1995; Johnson, Brems, & Alford-Keating, 1997). Within this body of

literature, some of the most consistent correlates of anti-LG attitudes have been

personality-related variables. More specifically, right-wing authoritarianism (RWA),

social dominance orientation (SDO), and traditional gender role attitudes have been

linked most strongly and consistently to anti-LG attitudes. Although studies that focus on

examining these relationships (e.g., Altemeyer & Hunsberger, 1992; Basow & Johnson,

2000; Whitley & Egisd6ttir, 2000) are informative, theoretical and empirical scholarship

is needed to integrate these findings and explain the shared links of personality-related

variables to anti-LG attitudes. More specifically, a question that needs to be examined is

what, if any, underlying construct or set of constructs explains the common links of

RWA, SDO, and traditional gender role attitudes to anti-LG attitudes.

In order to provide the groundwork for exploring this question, the present chapter

will review the extant literature on RWA, SDO, and traditional gender role attitudes and

present a fourth variable, openness to experience, that might provide a potentially

integrative and parsimonious explanation of the links of the aforementioned constructs to

anti-LG attitudes. The first three sections of the present chapter reflect critical reviews of

the literature linking RWA, SDO, and traditional gender roles to anti-LG attitudes. In the

fourth section, literature will be reviewed that points to openness to experience as an









underlying construct that might explain the links of the aforementioned variables to anti-

LG attitudes. Next, literature on the links of anti-LG attitudes and their predictors to anti-

LG behaviors will be described. Finally, a conceptual model and the hypotheses that are

examined in the present study will be presented.

Right-Wing Authoritarianism

Adorno, Frenkel-Brunswik, Levinson, and Sanford (1950) first defined the

authoritarian personality as a way to explain the rise of Fascism in the 1930s. The

authoritarian personality style was conceptualized to include nine traits:

Conventionalism, Authoritarian Submission, Authoritarian Aggression, Anti-intraception,

Superstition and Stereotypy, Power and Toughness, Destructiveness and Cynicism,

Projectivity, and exaggerated concern with sexual goings-on (Adorno et al., 1950).

Authoritarianism highlighted the role of psycho-dynamic influences on the individual

that, it was believed, predisposed one to become prejudiced. More specifically, Adorno et

al. (1950) theorize that threatening, forbidding, and status-conscious parents who used

harsh and arbitrary punishment raised authoritarian children. According to this

conceptualization, children repress hostility toward their parents and instead overglorify

their parents and act submissively toward them. Consequently, these children's repressed

aggression toward their parents becomes displaced onto out-groups.

Authoritarianism became a popular variable to study, and a considerable amount of

research was conducted using the F scale, the questionnaire developed by Adorno et al.

(1950) to assess this construct. However, numerous problems were identified with this

theory of authoritarianism and the F scale. First, definitions of the traits of

authoritarianism were thought to be unclear, lack distinction, and reflect some

redundancy (Altemeyer, 1981). Furthermore, the theory was criticized for lacking









testability and falsifiability. For example, if authoritarians report that their parents were

especially harsh, then this supports Adorno et al.'s (1950) theory. If authoritarians report

that their parents were not harsh, however, it can be claimed that they are overglorifing

their parents, still supporting the theory. Lastly, the psychometric properties of scores on

the F scale have been questioned (Altemeyer, 1981). For example, there are no reverse

scored items, making F scale scores susceptible to response set bias. Also many items

were intended to tap several of the nine traits at once, adding empirical ambiguity to the

distinctiveness of the nine traits.

To address some of the limitations of Adorno et al.'s (1950) model, Robert

Altemeyer (1981, 1988, 1996) developed the construct of right-wing authoritarianism

(RWA) based on social learning theory. Altemeyer (1981) defined RWA as comprised of

three attitudinal clusters. The first component, authoritarian submission, is defined as a

"high degree of submission to authorities who are perceived to be established and

legitimate in the society in which one lives" (p. 148). The second component is

authoritarian aggression and reflects a "general aggression, directed at various persons,

which is perceived to be sanctioned by established authorities" (p. 148). Such aggression

is manifested as the predisposition to inflict physical, psychological, economic, or social

harm onto targets perceived as socially stigmatized. The final component of RWA is

conventionalism defined as a "high degree of adherence to the social conventions which

are perceived to be endorsed by society and its established authorities" (p. 148).

Established and legitimate authorities could include parents, religious officials, heads of

government, police officers, judges, and military officials, and other persons or

institutions perceived as legal or moral authorities (Altemeyer, 1981).









In contrast to Adorno et al.'s (1950) model of authoritarianism, Altemeyer's (1981,

1988) theory of RWA relies on a social learning explanation of development. According

to this conceptualization, attitudes are shaped by other people, through direct teaching

and modeling, and through direct and vicarious experiences with the objects of these

attitudes. Thus, a right- wing authoritarian personality can be shaped by the child's

parents, religious leaders, peers, teachers, and through other models in the community.

Additionally, watching television, reading books, and hearing the news can shape the

person's attitudes.

Theoretical conceptualization of RWA suggests that persons with high levels of

RWA are likely to exhibit anti-LG attitudes because LG persons (a) are explicitly

condemned by some religious and political leaders, making LG persons an acceptable

target for prejudice and hostility, (b) represent a socially stigmatized out-group relative to

heterosexual women and men, and (c) are perceived as advocates for political change and

challenges to social conventions of traditional gender roles. Indeed, Robert Altemeyer

has conducted extensive research demonstrating a link between RWA and anti-LG

attitudes.

Altemeyer has measured the relationship between RWA and anti-LG attitudes with

thousands of participants using his Right-Wing Authoritarianism (RWA) scale and

Attitudes toward Homosexuals (ATH) scale (1988). The ATH is a balanced 12-item,

Likert-type scale with scores ranging from -4 to +4, and higher scores indicating more

anti-LG attitudes. ATH scores have been demonstrated to be relatively unidimensional,

with average intercorrelations among items at about .45, producing an alpha internal

consistency reliability estimate of about .90 (Altemeyer, 1988). ATH scores assess









cognitive, affective, and behavioral components of anti-LG attitudes. Sample items

include, "I find the thought of homosexual acts disgusting" and "I won't associate with a

known homosexual if I can help it."

Every year since 1984, Altemeyer has used the RWA and ATH scales to assess

attitudes of University of Manitoba Introductory Psychology students and their parents.

Across these studies, Altemeyer has found a consistent significant relationship between

RWA and anti-LG attitudes with correlations in the .50-.60 range (Altemeyer 1988, 1996,

2001). Based on this research, Altemeyer claimed that RWA "may explain hostility

toward gays and lesbians better than any other personality variables" (1996, p. 26).

The research program of Robert Altemeyer has yielded insight into attitudes

toward LG persons. However, there are particular limitations to these studies. For

instance, Altemeyer sampled primarily Canadian introductory psychology students and

their parents. Findings based on this population may not generalize to Canadian and

North America populations as a whole. In fact, Altemeyer reported that Manitoba,

Canada has recently elected an openly gay mayor (2001), suggesting that attitudes

towards lesbian and gay persons in Manitoba might be more positive that other areas.

Another limitation of Altemeyer's (1981, 1988, 1996, 2001) studies is that he uses the

ATH scale to assess anti-LG attitudes. This scale refers to "homosexuals" rather than to

"lesbians" or "gay men," leaving open the possibility that participants may consider their

attitudes toward only gay men when responding to ATH items (e.g., Haddock, Zanna,

and Esses, 1993). In addition, one question explicitly refers to "gay men" only.

Therefore, it is difficult to generalize results based on this scale to attitudes toward

lesbians.









Due to the lack of research on attitudes toward lesbians, Basow and Johnson (2000)

specifically studied the attitudes of 71 undergraduate women toward lesbians. Using

Herek's (1988) Attitudes Toward Lesbians scale, they found that RWA correlated

strongly with negative attitudes at r = .79, p < .001. Furthermore, a multiple regression

analysis revealed that RWA alone predicted 62% of the variance in attitudes toward

lesbians. This study demonstrated that RWA is an important predictor of attitudes toward

lesbian women.

Other studies also have found similarly strong relationships between RWA and

anti-LG attitudes. Whitely and Lee (2000) conducted a meta-analysis that examined the

relationships of authoritarianism, dogmatism, SDO, and political-economic conservatism

to anti-LG attitudes. They analyzed the results of 36 studies that included a total of

16,218 research participants and provided 51 effect sizes. For authoritarianism, Whitley

and Lee found a mean d of -1.21 and an equivalent r of-.52, p < .001. The d statistic

represents the standardized difference in mean anti-LG attitude scores between

participants scoring high and low on the predictor variable. These values for d and

equivalent r are well above Cohen's (1988) criteria for large effect sizes (i.e., d= 0.8; r =

0.5)

Furthermore, Whitley and Lee (2000) found that authoritarianism had a stronger

association with anti-LG attitudes when measured with the RWA scale, d= -1.19, r = -

.51, p < .001, than with the F scale (based on Adorno et al.'s psychodynamic

conceptualization of authoritarianism) d = -0.96, r = -.43, p < .001. The larger effect

sizes found with the RWA scale could be due to the relatively stronger psychometric

properties of this scale compared to that of the F scale. In addition, all of the studies using









Altemeyer's (1998) RWA scale used Canadian research participants, whereas the studies

using Adorno et al.'s (1950) F scale used United States participants. Therefore,

differences in mean effect sizes found for the two scales could also represent nationality

differences. On the other hand, there was no significant difference in scores based on

whether the ATH, the ATLG, or other measures of anti-LG attitudes was used. Overall,

Whitley and Lee's meta-analysis demonstrates that RWA is indeed related to anti-LG

attitudes.

Social Dominance Orientation

In addition to right-wing authoritarianism (RWA), social dominance orientation

(SDO) has also been linked consistently with prejudice and specifically with anti-lesbian

and gay (LG) attitudes. Pratto, Sidanius, Stallworth, and Malle (1994) defined SDO as

the "extent to which one desires that one's in-group dominate and be superior to out-

groups" (p 724). The desire to maintain the superior position of their in-groups motivates

people high in SDO to denigrate members of out-groups, to oppose equality-enhancing

social programs such as affirmative actions, and to discriminate against members of out-

groups in order to enforce the status quo of the in-groups' power position. Thus, persons

high in SDO tend to have anti-LG attitudes because LG persons have lower social status

than do heterosexuals (Whitley & Lee, 2000) and because LG persons are perceived as

advocates for social equality (Whitley, 1999).

Empirical research supports the proposed relationship between SDO and anti-LG

attitudes. In addition to examining RWA, the meta-analysis described earlier by Whitley

and Lee (2000) also examined the relationship between SDO and anti-LG attitudes.

Based on 9 effect sizes, Whitley and Lee found that SDO was related to anti-LG attitudes

with a mean d= -.76 and an equivalent r = -.35, p < .001, indicating effect sizes within









the medium to large range (Cohen, 1988). Effect sizes did not vary significantly whether

SDO was assessed with Pratto et al.'s (1994) SDO scale or a measure developed by

Altemeyer (1988). Results of Whitley and Lee's meta-analysis demonstrate that SDO, no

matter which measure is used to assess it, is an important correlate of anti-LG attitudes.

Although RWA and SDO are both related to anti-LG attitudes, Altemeyer (1988)

argued that the two constructs are distinct in that SDO "does not have the same

psychological roots that previous studies have unearthed in right-wing authoritarians" (p.

61). RWA is conceptualized as an intragroup phenomenon, whereas SDO is

conceptualized as an intergroup phenomenon (Altemeyer, 1998; Pratto et al., 1994).

More specifically, RWA focuses on submission to in-group authority figures independent

of whether they advocate intergroup dominance, whereas SDO focuses on dominance

over out-groups independent of the views of the in-group authorities. As such, a person

high in RWA is very accepting of traditional values and institutional authorities and is

likely to follow the instructions of those in positions of power and influence. By contrast,

an individual high in SDO is not motivated by a sense of duty to authorities or morality

(Heaven & Bucci, 2001). Thus, although RWA and SDO frequently explain prejudicial

beliefs and behaviors, such as anti-LG attitudes, they do so in a complementary but not

redundant fashion. such that each accounts for unique variance in a variety of beliefs,

behaviors, and traits (Altemeyer, 1998).

Indeed, empirical evidence indicates that RWA and SDO are only minimally

correlated (Altemeyer, 1998; Pratto et al., 1994) and the relationship of SDO to prejudice

is not accounted for by a joint relationship with RWA. For example, even after

controlling for the effects of RWA, SDO still explained variance in ethnocentrism,









racism, and nationalism (Pratto et al. 1994). Furthermore, across several studies reviewed

and conducted by Altemeyer (1998), RWA and SDO were only minimally correlated (r =

.22), yet when entered into a regression equation with numerous other personality

variables, together accounted for over 50% of the variance in a generalized prejudice

index (comprised of prejudice against African Americans, women, and LG persons).

Altemeyer (1998) stated that RWA and SDO each "explained different segments of the

overall prejudice," implying that both RWA and SDO accounted for unique variance in

the equation, yet he did not describe the specific data that served as the basis for this

statement.

Whitley and Lee (2000)'s findings, however, provided more detail and were

consistent with Altemeyer's (1998) conclusion that RWA and SDO each accounted for

unique variance in anti-LG attitudes. More specifically, Whitley and Lee conducted a

follow up study to the meta-analysis described above, in which they surveyed 316

Introductory Psychology students. The participants completed Herek's (1988) Attitudes

Toward Lesbians and Gay Men (ATLG) scale, Altemeyer's (1998) RWA scale,

Rokeach's (1956) Dogmatism scale, Pratto et al.'s (1994) SDO scale, and Pratto et al.'s

(1994) three-item scale for political-economic conservatism. These authors found that

RWA and SDO were only slightly correlated with each other (r = .28). Of all the

variables studied, RWA correlated most strongly with negative attitudes toward lesbians

(r = -.58) and negative attitudes toward gay men (r = -.54). SDO was the second

strongest correlate of negative attitudes toward lesbians (r = -.39) and negative attitudes

toward gay men (r = -.44) (all ps < .001). Furthermore, with RWA and SDO controlled as

covariates, political-economic conservatism did not significantly add any unique variance









to attitudes toward lesbians and gay men. Additionally, with RWA and SDO controlled

as covariates, dogmatism added less than 2% unique variance to attitudes toward

lesbians, F (1, 312) = 8.29, p < .01, and less than 1% of the variance in attitudes toward

gay men, F (1, 312) = 4.46, p < .05. In another equation that included all four attitudinal

variables as predictors, RWA and SDO each accounted for unique variance in attitudes

toward lesbians (B = -.42 and 8 = -.18) and gay men (B = -.43 and 8 = -.18). Taken

together, these results suggest that RWA and SDO are two primary correlates of anti-LG

attitudes. Each variable was related to anti-LG attitudes with the other controlled and they

jointly accounted for most of the relationships of dogmatism and conservatism to anti-LG

attitudes.

Traditional Gender Role Attitudes

A third personality-related construct that has been demonstrated consistently to

have a significant relationship with anti-lesbian and gay (LG) attitudes is traditional

gender role attitudes. Gender role attitudes are "beliefs about the appropriate role

activities for women and men" (McHugh & Frieze, 1997, p. 4). Those endorsing

traditional gender roles have the attitude that women should abide by traditional female-

role norms and that men should abide by traditional male-role norms. Heterosexual

persons who endorse traditional gender roles hold anti-LG attitudes because LG persons

are stereotypically perceived as having cross-gender traits, roles, and physical

characteristics.

Empirical evidence suggests that a commonly held stereotype is that LG persons

violate traditional gender roles, with gay men acting more feminine and lesbian women

acting more masculine than heterosexual men and women should, respectively. Kite and

Deaux (1987) examined gender stereotypes about LG persons with 206 participants from









introductory psychology courses. Participants rated lesbian women, gay men,

heterosexual women, and heterosexual men on how likely each group was to possess

masculine and feminine traits (e.g., self-confident, kind), role behaviors (e.g., financial

provider, cooks the meals), physical characteristics (e.g., sturdy, graceful), and

occupations (e.g., construction worker, telephone operator). Results were analyzed using

multivariate analysis of variance and a significant interaction was found between sexual

orientation of target and sex of target, multivariate F (8, 182) = 20.60, p < .001. This

interaction demonstrated that participants accepted the implicit gender inversion theory

for LG persons in that gay men were rated lower on masculine qualities, yet higher on

feminine qualities than were heterosexual men, and that lesbian women were rated lower

on feminine qualities, yet higher on masculine qualities than were heterosexual women.

This was evident in all categories (i.e., traits, role behaviors, physical characteristics,

occupations). These findings demonstrate that college students stereotypically describe

LG persons in cross-gender ways.

In general, people who act in ways consistent with the opposite gender are not

viewed positively (e.g., Laner & Laner, 1979, 1980). Therefore, heterosexual persons

might hold anti-LG attitudes because LG persons are, as previously shown,

stereotypically viewed as deviating from the cultural expectations of the roles that men

and women should perform in their lives. These anti-LG attitudes should be particularly

strong among people who hold traditional gender role attitudes because homosexuality

poses a threat to their world view that places "masculine" and "feminine" at opposite

ends of a bipolar dimension (Kite & Whitley, 1998) and stereotypical views of LG

persons place them at inappropriate ends of the gender dichotomy.









Indeed, research using various indicators of traditional gender role attitudes has

found significant correlations with anti-LG attitudes. Whitley (2001) conducted a meta-

analysis to examine the relationship between traditional gender role attitudes and anti-LG

attitudes. This meta-analysis included 42 studies with a total of 9629 research participants

that yielded 50 effect sizes for the relationship between gender-role beliefs and attitudes

toward LG persons. Whitley (2001) included seven types of measures for gender role

attitudes: Attitudes Toward Women Scale (ATW; Spence & Helmreich, 1978), Attitudes

Toward Feminism (FEM) scale (Smith, Ferree, & Miller, 1975), a measure of attitudes

toward sexual equality (MacDonald, 1974), measures of old-fashioned sexism (i.e., belief

in traditional gender stereotypes or approval of unequal treatment of men and women),

measures of endorsement of traditional male role norms, and measures that assessed

attitudes toward both male and female roles. The mean overall effect size indicated that

people who endorsed traditional gender role attitudes held more anti-LG attitudes, d= -

1.11, r = -.48, Z= -102.03, p <.001, Qw(49) = 1041.87, p < .001, with effect size in the

large range (Cohen, 1988).

Whitley (2001) also compared the magnitude of the anti-LG attitudes-gender role

ideology link across different operationalizations of traditional gender role attitudes.

These comparisons revealed that endorsement of traditional gender roles and gender

hierarchies, rather than awareness of and sensitivity to gender discrimination or gender-

related self-concept, are key correlates of anti-LG attitudes. More specifically, the mean

effect size for studies that assessed endorsement of traditional gender hierarchies (i.e., old

fashioned sexism) was similar to that of studies that assessed endorsement of male role

norms, but was significantly greater than that of studies that assessed perceptions of









discrimination against women (i.e., modern sexism) x2 (1) = 119.11, p = .001. With

regard to gender-related self-concept, the mean effect size for the relationship of

femininity/expressiveness to anti-LG attitudes was not statistically significant. Although

the mean effect size of the relationship between masculinity/instrumentality and anti-LG

attitudes was significant, it was quite small (Cohen, 1988) and therefore of little practical

significance, d= .07, r = .03, Z= 3.83,p <.001, Qw(19) = 179.62,p < .001.

Overall, these results indicate that modern sexism and gender-role self-concept are

not closely linked to anti-LG attitudes. On the other hand, traditional gender role attitudes

defined to "represent people's ideas of the proper roles for men and women in society

and of behavioral norms for men and women" (Whitley, 2001, p. 692) and typically

assessed with measures such as the Masculine Role Norms Scale (Thompson & Pleck,

1986) and the ATW (Spence & Helmreich, 1978) are closely linked with anti-LG

attitudes. Indeed, Whitley's (2001) results indicated that traditional gender role attitudes

accounted for a substantial (23%) amount of variance in anti-LG attitudes. The

magnitude of this relationship was similar for attitudes toward lesbians and attitudes

toward gay men. Finally, the relatively high consistency of mean effect sizes across

different measures of traditional gender role attitudes and anti-LG attitudes attests to the

robustness of the relationship (Whitley, 2001).

Kite and Whitley (1998) noted that the relationship between traditional gender role

attitudes and anti-LG attitudes could reflect a relationship between attitudes toward LG

persons and a generalized conservative belief system. As such, it could be that the

relationship between traditional gender role attitudes and anti-LG attitudes could be

explained by variables such as RWA and SDO. Whitley and Egisd6ttir (2000) examined









this possibility with 266 introductory psychology students. Each participant completed

the ATLG (Herek, 1988), AWS (Spence, Helmreich & Stapp, 1973), RWA (Altemeyer,

1988) and SDO (Pratto et al., 1994) scales. In this study, gender role attitudes accounted

for an additional 4% of the variance in attitudes toward lesbians, F (1, 249) = 17.22, p <

.001, and an additional 8% of the variance in attitudes toward gay men, F (1, 249) =

36.66, p < .001, beyond variance accounted for by RWA and SDO. Therefore, gender

role attitudes appear to account for variance in attitudes toward LG persons beyond that

accounted for by RWA and SDO.

Overall, studies on the link of traditional gender role attitudes to anti-LG attitudes

demonstrate that heterosexual persons stereotypically assume that lesbians and gay men

violate traditional gender role norms. Thus, persons who endorse traditional gender role

attitudes hold more anti-LG attitudes. This significant relationship has been demonstrated

using various measures of traditional gender role attitudes. However, the key aspect of

gender role-related variables that is related to anti-LG attitudes appears to be traditional

gender role attitudes and not gender role self-concept or modern sexism. Finally, the

relationship between traditional gender role attitudes and anti-LG attitudes is not

accounted for fully by a joint relationship with either RWA or SDO. Taken together,

these findings support the inclusion of traditional gender role attitudes in the present

study.

Openness to Experience: An Underlying Construct

The literature reviewed thus far indicates that conceptual definitions of right-wing-

authoritarianism (RWA), social dominance orientation (SDO), and traditional gender role

attitudes link each of these variables to anti-lesbian and gay (LG) attitudes and empirical

evidence supports such links. In addition to their links to anti-LG attitudes, RWA, SDO,









and traditional gender role attitudes, each have also been linked conceptually and

empirically to openness to experience, a dimension of personality described in the Five

Factor Model. Openness, in turn, has been shown to be related to anti-LG attitudes. Thus,

openness to experience might be a mediating variable that provides a more parsimonious

explanation and accounts for the links of RWA, SDO, and traditional gender roles to anti-

LG attitudes.

Within the framework of the Five Factor Model, openness to experience is

conceptualized to reflect active imagination, aesthetic sensitivity, receptiveness to inner

feelings, preference for variety, intellectual curiosity, and independence of judgment

(McCrae & Costa, 2003). A person who is high in openness values intellectual matters, is

nonconforming, introspective, has a vivid imagination, and admits that what is right and

wrong for one person may not be applicable in other circumstances. On the other hand, a

person scoring low in openness is likely to be rigid, have intolerance for ambiguity, have

a narrow range of emotions, favor conservative values, judge others in conventional

terms, and is uncomfortable with complexities. These characteristics associated with low

levels of openness overlap with the characteristics associated with high levels of RWA,

SDO, and traditional gender role attitudes.

First, a lack of openness is evident in the traditional conception of authoritarianism

and in the more modem definition ofRWA. Adomo et al. (1950), who first defined

authoritarianism, argued that persons high in this variable are closed-minded, mentally

rigid, and disinterested in imaginative pursuits. These traits are also evident in

Altemeyer's (1981) conceptualization of RWA. More specifically, Altemeyer conceived

RWA as being comprised of conventionalism or a "high degree of adherence to the social









conventions which are perceived to be endorsed by society and its established

authorities" (p. 148). Thus, persons high in RWA are likely to reject thoughts or

experiences that are not sanctioned by proper authorities.

Indeed, RWA has consistently been shown to relate significantly and negatively to

openness to experience, with correlations ranging from -.30 to -.57 (Altemeyer, 1996;

Butler, 2000; Heaven & Bucci, 2001; Peterson, Smirles, & Wentworth, 1997). For

example, Butler (2000) administered the full 240-item NEO Personality Inventory (PI;

Costa & McCrae, 1992) and the RWA scale (Altemeyer, 1996) to 76 participants

recruited from psychology courses. In this sample, RWA was significantly correlated

with openness, r = -.48, p < .001, but not with any of the other personality dimensions

(i.e., extroversion, neuroticism, agreeableness, and conscientiousness). Furthermore,

RWA was significantly correlated with five of the six facets of openness. The exception

was the fantasy facet, but even this relationship was in the predicted direction and fantasy

did not correlate strongly with any of the other openness facets in this sample. Consistent

with theoretical conceptions of RWA, RWA scores were most strongly correlated with

openness to values, followed closely by feelings and actions. Butler (2001) claimed

"given that the mean correlation of RWA and the openness facets (r = -.30), compares

favorably to the magnitude of the mean intercorrelations among the facets themselves (r

= .22), we may be justified in viewing authoritarianism as an additional facet of openness

to experience" (p. 8). Furthermore, McCrae (1996) suggested that openness versus

closedness "is an indispensable element" (p. 327) in explaining an authoritarian

personality.









Similarly, a lack of openness to experience is apparent in SDO. Persons high in

SDO accept hierarchy-legitimizing myths in order to promote the superiority of one

group over another (Pratto et al., 1994). Instead of viewing other people as individuals,

persons high in SDO concentrate on targets' out-group membership and characteristically

have a lack of sympathy, empathy, and tolerance for members of out-groups. These

characteristics, along with the lack of ability to view the complexities in the world and

the narrow range of emotions evident in persons high in SDO, are consistent with a lack

of openness to experience. Consistent with this theoretical expectation, research has

shown that SDO is correlated significantly and negatively with openness. Heaven and

Bucci (2001) used the International Personality Item Pool (IPIP; Goldberg, 1999) with a

sample of psychology students and found a correlation of r = -.26, p < .001 between SDO

and openness. Additionally Pratto et al. (1994) used John, Donahue, and Kentle's (1992)

Big-Five Personality Inventory and, with a sample of college students, found a

correlation ofr = .28, p < .01 between SDO and openness.

Finally, there is a theoretical link between an adherence to traditional gender role

attitudes and a lack of openness to experience. Persons who adhere to traditional gender

role attitudes expect others to fit into a relatively stable set of gender roles, traits, and

physical attributes (e.g., Kite & Whitley, 1998). This acceptance of traditional roles

demonstrates a lack of tolerance for ambiguity and variety within the gender spectrum.

Additionally, it represents traditional and conforming values. Therefore persons who

adhere to traditional gender role attitudes are expected to have low levels of openness to

experience. Although there is a theoretical connection between traditional gender role

attitudes and openness, previous research has not examined this link directly.









Nevertheless, extant findings indicate a negative link between openness and constructs

that are linked with traditional gender role attitudes. For example, Tokar, Fischer, Schaub

& Moradi (2000) administered to a sample of male university students the Gender Role

Conflict scale (O'Neil, Helms, Gable, David & Wrightsman, 1986), the Masculine

Gender Role Stress scale (Eisler & Skidmore, 1987), and the 60 item NEO Personality

Inventory (Costa & McCrae, 1985). These authors found that masculine gender role

conflict and gender role stress both were significantly and negatively related to openness

to experience with correlations of-.38 and -.21, respectively. Gender role conflict is a

"psychological state in which gender roles have negative consequences or impact on the

person or others" (O'Neil, 1981, p. 203). Men's gender role stress results from a failure

to meet cultural standards of masculinity (Eisler & Blalock, 1991). Thus, although men's

gender role conflict and stress are somewhat different from traditional gender role

attitudes, there is significant theoretical overlap in that one must endorse traditional

gender roles in order to experience gender role conflict or stress. Indeed, both gender role

conflict and stress are correlated significantly with traditional gender role attitudes

(Thompson, 1991, as cited in Thompson, Pleck, & Ferrera, 1992). Thus, it is expected

that traditional gender role attitudes will relate to low openness and the present study will

examine directly this proposition.

It follows from the theoretical conceptualization ofRWA, SDO, and traditional

gender role attitudes and from empirical evidence, that lack of openness to experience

might be a global construct that underlies these various correlates of anti-LG attitudes.

Furthermore, openness might be an underlying construct that explains partially or fully,

the links of the aforementioned correlates to anti-LG attitudes. In fact, in addition to its









conceptual and empirical links to RWA, SDO, and traditional gender attitudes, openness

is also significantly related to prejudicial attitudes in general and more specifically to

anti-LG attitudes. For example, Ekehammar and Akrami (2003) examined the Five

Factor Model and a generalized prejudiced factor with a sample of 156 Swedish, non-

psychology university students. The generalized prejudice factor included anti-LG

attitudes in addition to modern and classical racism, sexism, and attitudes toward

intellectually disabled individuals. The participants took a Swedish translation of the

NEO-PI (Costa & McCrae, 1985) and Scandinavian versions of instruments assessing the

prejudice variables. In this study, openness was significantly related to scores on the

generalized prejudice factor, r = -0.45, p < .001. In another study, Cullen, Wright, and

Alessandri (2002) focused more specifically on the link between anti-LG attitudes and

the Five Factor Model dimensions of personality with a sample of 123 university

students. Participants completed the Neuroticism, Extroversion, and Openness scales of

the NEO-PI (Costa & McCrae, 1985) and the ATLG (Herek, 1984). In a regression

equation that included openness, neuroticism, extroversion, gender, and contact with LG

persons, 25% of the total variance in anti-LG attitudes was accounted for, and openness

was the only personality variable that accounted for significant unique variance.

Since openness is correlated with both the predictor variables and the outcome

variable of anti-LG attitudes, it is hypothesized that openness will mediate the links of

RWA, SDO, and traditional gender role attitudes to anti-LG attitudes. In other words, it is

predicted that RWA, SDO and traditional gender attitudes have indirect links to anti-LG

attitudes through openness. According to Baron and Kenny (1986) for a variable to be

considered a mediator, there must be a significant relationship (a) between the predictor









and the mediator, (b) between the mediator and the criterion, and (c) and the predictor

and the criterion. The literature reviewed thus far fits with this definition and statistical

requirements. More specifically, in the previous literature, openness to experience

(mediator) has been shown to be correlated significantly with RWA, SDO, and traditional

gender role attitudes (predictors) as well as with anti-LG attitudes (criterion). The present

study will provide an empirical test of this mediational relationship.

Need to Study Anti-Lesbian and Gay Behaviors

Although much of extant literature has focused on the correlates of anti lesbian/gay

attitudes, an implicit reason for studying anti-LG attitudes is that the expression of these

attitudes as anti-LG behaviors might have serious negative interpersonal and

intrapersonal consequences for LG persons and society at large. In fact, the prevalence of

anti-LG behaviors (e.g., name-calling, threats, assaults) is staggering. D'Augelli (1989)

found that 50% of the gay and lesbian college students surveyed reported having

overheard disparaging comments, 26% reported having experienced personal verbal

insults, 26% reported having been threatened with physical violence, and 23% reported

having been victims of assault. Furthermore, such negative experiences have been linked

to mental health related concerns (e.g., suicidal ideation, psychological distress) among

lesbian and gay persons (e.g., Diaz et al, 2001; DiPlacido, 1998; Mays & Chochran,

2001). Thus there is a need to study anti-LG behaviors in addition to anti-LG attitudes.

Despite the considerable prevalence and potential negative consequences of anti-

LG behaviors and numerous calls for research examining such behaviors (Kite &

Whitley, 1998; Herek, 2000; Whitley, 2001), the majority of extant research has not

attended to predictors or correlates of anti-LG behaviors and has focused mostly on

identifying correlates of anti-LG attitudes. The limited research available shows that









although hate crimes and anti-LG behaviors are inevitably influenced by complex

situational factors (Franklin, 1998), anti-LG attitudes do indeed contribute to anti-LG

behaviors. A review of extant literature on the link between anti-LG attitudes and anti-LG

behaviors will now be discussed.

Franklin (2000) conducted empirical research into prevalence rates and motivations

for anti-LG harassment and violence with a noncriminal sample. Her sample included

489 students at six community colleges in five counties in Northern California. The

sample was racially and economically diverse with 40% White, 30% Asian, 14% Latino,

and 12% African American, and 59% reporting that their father's highest educational

degree was a high school degree. Approximately 93% of the participants identified as

heterosexual. The participants completed an Antigay Behaviors Inventory developed for

this study. This 89-item inventory assessed (a) frequencies of various types of anti-LG

behaviors, (b) descriptions of specific incidents, (c) the stated motivations for such

behaviors, and (d) motivations reported by nonassailants for not engaging in anti-LG

assaults. In addition, participants completed the ATLG (Herek, 1988), a measure of social

drinking, and the Attitude Function Inventory (Herek, 1986), which measures four

distinct social and psychological functions served by attitudes toward LG persons.

Results of this study indicated that large percentages of the participants admitted

engaging in anti-LG behaviors, having friends who had engaged in such behaviors, and

witnessing anti-LG incidents. In fact, 10% (n = 49) of the sample reported physically

assaulting or threatening people whom they believed were lesbian or gay and 32% (n =

158) reported calling LG persons by insulting names (but not physically assaulting them).









Additionally, of the 321 respondents who denied any verbal or physical harassment, 23%

(n =83) reported witnessing such incidents.

As part of the Antigay Behaviors Inventory, individuals who admitted to engaging

in anti-LG behavior were asked to rate 19 motivations on a 4-point Likert scale. The

motivations included such items as "I did it to have fun" and "I did it because

homosexuals disgust me." Franklin (2000) examined the structure of these motivations

using exploratory factor analytic procedures. She identified four factors labeled peer

dynamics, antigay ideology, thrill seeking, and self-defense, which accounted for 64% of

the variance in the data. Although peer dynamics (35%) accounted for the greatest

percentage of variance, antigay ideology accounted for the second highest amount of

variance at 15%. This factor reflected anti-LG attitudes such as disgust, hatred, religious

and moral values, and the belief that LG persons spread AIDS. The emergence of this

factor demonstrates that anti-LG attitudes were a motivation for anti-LG behaviors.

Furthermore, in this sample, a significant relationship between anti-LG attitudes

and anti-LG behaviors was found. The mean attitude score for participants who reported

anti-LG behaviors (M= 25.3, SD = 6.92) was significantly greater than that for

nonassailants (M= 21.5, SD = 7.24), t(484) = .25,p < .001. In addition, Franklin

separated the sample into those who had committed anti-LG behaviors and those who had

not. A point-biserial correlation revealed a positive relation between committing anti-LG

behaviors and anti-LG attitudes as measured by the ATLG, r = .25, p < .001.

Other studies have provided further support for the link between anti-LG attitudes

and behaviors. For example, Patel, Long, McCammon, and Wuensch (1995) found a

relationship between anti-LG attitudes and behaviors with 102 male college students.









Approximately half of the participants were enlisted marines enrolled in college classes

on two military bases, whereas the rest were civilians enrolled in classes at a university.

Participants completed a modified version of the Index of Attitudes Toward Homophobia

(IAH; Hudson & Ricketts, 1980). This scale was developed to measure fear, disgust,

anger, discomfort, and aversion toward LG people. Patel et al. (1995) created the Self-

Report of Behavior Scale (SBS) for this study to measure past negative behavior toward

gay men in a variety of situations. Similar to Franklin's (2000) findings, anti-LG

behaviors were not uncommon. In this sample of men, 5% reported getting into a

"physical fight with someone who I thought was making moves on me," 34% verbally

threatened someone who had "checked me out," and 42% moved away from someone

who was believed to be gay. Anti-LG attitudes were significantly correlated with anti-LG

behaviors, r = .40, p < .01. A multiple regression analysis was conducted predicting anti-

LG behaviors with anti-LG attitudes and personality variables (social adjustment,

emotional stability, defensiveness) measured with the Minnesota Muliphasic Personality

Inventory (MMPI). In this equation, anti-LG attitudes accounted for significant unique

variance in anti-LG behaviors (8 = .38, p < .001).

A limitation of the Patel et al. (1995) study is that all of the participants were male

and that attitudes and behaviors were assessed toward only gay men. Whitley (2001)

addressed this limitation by assessing attitudes and behaviors toward lesbians and gay

men with a sample of 227 female and 207 male introductory psychology students. Anti-

LG attitudes were assessed using a composite score of an affective measure and the

Beliefs About Lesbians and Beliefs About Gay Men scales (LaMar & Kite, 1998).

Whitley (2001) assessed anti-LG behaviors using a revised version of Patel et al.'s (1995)









Self Report of Behavior Scale (SRBS-R). The scale was revised to include questions

about lesbians in addition to questions about gay men. In this study, anti-LG attitudes and

anti-LG behaviors were significantly and moderately correlated r = .39.

In addition to research on the self-report of anti-LG behaviors, experimental

research has demonstrated that anti-LG attitudes are related to actual anti-LG behaviors,

measured in a variety of ways. Bernat, Calhoun, Adams, and Zeichner (2001) found that

in a sample of college men, those who held more anti-LG attitudes as measured by the

Homophobia Scale (HS) were more aggressive toward an LG target. In this experiment,

physical aggression was measured by the number of shocks given to a fictitious opponent

during a competitive reaction time task. Additionally, Kite (1992) found that college men

who held more anti-LG attitudes as measured by the Homosexuality Attitude Scale

(HAS) were less likely to choose to meet a partner perceived to be lesbian or gay. In

another study, Kite and Deaux (1986) found that when male college student participants

had an interpersonal interaction with an LG target, those with anti-LG attitudes formed

more negative impressions and recalled less information about the target than did

participants who did not hold negative attitudes. A limitation of the previously described

studies is that they only assessed anti-LG behaviors for male participants. A study that

addressed this limitation examined proposed budget cuts for student organizations.

Haddock and Zanna (1998) found that male and female students with anti-LG attitudes

proposed greater cuts to an LG student organization (mean percentage funding reduction

= 42%) than did participants with relatively neutral (25%) or favorable attitudes (27%).

Additionally, Swim, Ferguson, and Hyers (1999) found that female college students

engaged in greater social distancing (by altering their stated opinions to disagree with the









target) from a lesbian confederate than from a heterosexual woman confederate.

Furthermore the participants with greater anti-LG attitudes, as measured by the HAS,

socially distanced themselves in more ways (i.e., by altering their opinion to agree with

the majority as opposed to making independent responses which differed from all

participants) than those participants with less anti-LG attitudes. Overall, these

experiments demonstrate that for both men and women anti-LG attitudes are positively

related to anti-LG behaviors.

Just as there is evidence linking anti-LG attitudes to anti-LG behaviors, there is

also some support for the notion that the predictors of anti-LG attitudes (i.e., right-wing

authoritarianism (RWA), social dominance orientation (SDO), and traditional gender role

attitudes are related to anti-LG behaviors. Support for the link between RWA and anti-

LG behaviors can be inferred from Schope and Ellison's (2000) study on religious

fundamentalism. These authors found, in a sample of 129 college student participants,

that those who self identified as having "very traditional or fundamentalist" religious

beliefs were more likely than their less religiously fundamentalist counterparts to have

engaged in anti-LG behaviors. More specifically, one third of the religious

fundamentalist participants reported that they had threatened a lesbian or gay person

within the last year, compared to only 4% of participants who identified their religion as

"somewhat traditional or conservative" or liberal. While the religious fundamentalists

comprised only 12% of the sample, they accounted for 50% of those who had threatened

gay persons. Although RWA was not directly examined in this study, RWA has been

shown (a) to be related strongly (r = .68) to religious fundamentalism (Altemeyer &

Hunsberger, 1992) and (b) to account for the link of religious fundamentalism to anti-LG









(Hunsberger, 1995). Therefore, evidence linking religious fundamentalism to anti-LG

behaviors suggests that RWA is likely to be related to anti-LG behaviors as well.

Furthermore, direct support for the link between RWA and anti-LG behaviors is

provided by an experiment by Haddock, Zanna, and Esses (1993). In this study, 63

college student participants completed the RWA scale. Two weeks later these participants

returned and were told that the undergraduate psychology society was soliciting views of

psychology students for the student government about budget cuts for student groups. A

median split was conducted for RWA scores and an ANOVA revealed that those with

high RWA scores proposed significantly greater cuts (37%) to an LG student

organization than those with low RWA scores (25%), F(1, 53) = 6.05, p < .05. This study

demonstrates that RWA is indeed related to anti-LG behaviors.

Unlike research on the link of RWA to anti-LG behaviors, no empirical research to

date has examined the relationship between SDO and anti-LG behaviors. However, SDO

has a strong relationship with anti-LG attitudes. Furthermore, SDO is based on a theory

of group conflict, which postulates that people will engage in unofficial terror, which is

"violence or threat of violence perpetrated by private individuals from dominant groups

against member of subordinate groups" (Sidanius & Pratto, 1999, p. 42). Thus it is

expected that persons high in SDO will report more anti-LG behaviors. This study will be

the first to empirically examine this relationship.

Lastly, extant data suggest that traditional gender role attitudes are related to anti-

LG behaviors. Franklin's (2000) study provided some support for the relationship

between traditional gender role attitudes and anti-LG behaviors. Her participants

completed the Male Role Attitudes Scale (MRAS; Pleck, Sonenstein, & Ku, 1993), which









assesses participants' perceptions about the importance of men fulfilling masculinity

standards. Franklin found that male assailants scored significantly higher on masculinity

ideology than did their nonassailant counterparts, t(202) = 3.24, p = .001. Also, for men,

there was a significant correlation between masculine ideology and anti-LG behaviors, r

= .21, p <. 01. On the other hand, there was not a significant difference between female

assailants and nonassailants on masculine ideology, and for women, the relationship

between masculine ideology and anti-LG behaviors was non-significant.

MRAS scores only had adequate reliability with a coefficient alpha of .68 for the

entire sample, male and female. This could be one reason why a significant relationship

between masculine ideology and anti-LG behaviors was not found among the female

participants. In fact, when Whitley (2001) examined traditional gender role attitudes

about women and men (using AWS and MRNS scores, which had better psychometric

properties than did MRAS scores in Franklin's sample), he found that scores on the AWS

(r = .49, p < .001) and MRNS (r = .45, p < .001) were both correlated with anti-LG

behaviors for women and men participants. In summary, traditional masculine ideology, a

component of gender roles attitudes was shown to be related to anti-LG behaviors for

men. Furthermore, gender role attitudes assessed more broadly were also related

significantly to anti-LG behaviors for both women and men.

Purpose of Study

Previous research has demonstrated that right-wing authoritarianism (RWA), social

dominance orientation (SDO), and traditional gender role attitudes are important

predictors of anti-LG attitudes. Each of these variables is also related conceptually and

empirically to openness to experience. Openness in turn is related to anti-LG attitudes.

Therefore, openness to experience may serve as a mediator of the relationships of RWA,









SDO, and traditional gender role attitudes to anti-LG attitudes. Furthermore, research has

also shown that anti-LG attitudes, RWA, and traditional gender roles are predictors of

anti-LG behaviors. It is expected that SDO and openness to experience will also be

related to anti-LG behaviors. However, it is not known if RWA, SDO, traditional gender

roles, and openness to experience will relate directly to anti-LG behaviors, or if their

relationship is mediated by anti-LG attitudes. Thus, the present study will examine if

anti-LG attitudes mediate the relationships of RWA, SDO, traditional gender role

attitudes, and openness to anti-LG behaviors. The model tested in the current study is

presented in Figure 1 and examines the following hypotheses:

1. RWA, SDO, and traditional gender role attitudes will be significantly
and positively related to anti-LG attitudes and behaviors and openness
to experience will be significantly and negatively related to anti-LG
attitudes and behaviors. Support for these links is a prerequisite for
examining corresponding links in Hypothesis 2 and 3.
2. Openness to experience will either partially or fully mediate the
relationships of RWA, SDO, and traditional gender role attitudes to
anti-LG attitudes.
3. Finally, anti-LG attitudes will either partially or fully mediate the
relationships of RWA, SDO, traditional gender role attitudes, and
openness to experience to anti-LG behaviors.














CHAPTER 3
METHODS

Participants

A total of 238 undergraduate students were recruited at a large southeastern

university from the Psychology 100 participant pool and from two other undergraduate

psychology courses where extra credit was granted. Given the present study's focus on

heterosexual persons' attitudes toward lesbian and gay persons, two participants who

identified as bisexual were not included in analyses, because it is expected that such

persons' attitudes and behaviors toward LG individuals could stem from different

constructs than those focused on in the present study. In addition, 4 participants were not

included in analyses because they did not complete significant portions of the survey

packet, and 1 participant was excluded because she or he did not indicate gender.

Therefore analyses were conducted using data from 231 participants; 97 men and 134

women.

Participants ranged in age from 17 to 27 years (M 19.48; Mdn = 19; SD =

1.51). Eighty-nine percent of the sample identified as exclusively heterosexual and 11%

as mostly heterosexual. With regard to race/ethnicity, 58% of the sample identified as

White, 16% Latina/Hispanic, 11% African American/Black, 6% Asian American/Pacific

Islander, 0.4% Native American, and 10% multiracial or other. Overall, 50% of the

participants were in their first year of college, 23% in their second year, 17% in their

third year, and 10% in their fourth year. Fifty-two percent of the participants reported that

they were single, 37% were in a dating relationship, and 11% were married or in a









committed relationship. In terms of social class, 44% of the sample identified as middle

class, 34% upper middle class, 15% working class, 4% lower class, and 3% upper class.

With respect to religious identification, 32% identified as Catholic, 23% Christian, 11%

Baptist, 8% Methodist, 5% Jewish, 4% Presbyterian, 1% Muslim, .5% Buddhist, 3%

Agnostic, and 3% Atheist or no religious orientation. Furthermore, 31% of the sample

reported that they never attend religious services, 29% attend less than once a month, 9%

once a month, 9% twice a month, 14% once a week, and 9% attend services more than

once a week.

Procedures

Participants completed a survey packet that included the instruments described

below as well as a demographic questionnaire. Before beginning the survey, participants

received written informed consent, and after completing the survey, they received written

debriefing. The order of the instruments in the survey packet was counterbalanced to

control for order effects.

Instruments

Criterion Variables

Attitudes toward LG persons were assessed with the Attitudes Toward Lesbians

and GayMen (ATLG) scale. Herek's (1988) scale is one of the more commonly used

measures of anti-LG attitudes and assesses the affective component of anti-LG attitudes

referred to as "condemnation-tolerance." The ATLG is a 5-point Likert scale ranging

from 1 = disagree strongly to 5 = agree strongly. The scale has 20 items; 10 items assess

attitudes toward lesbians (ATL), and 10 items assess attitudes toward gay men (ATG).

Sample items include "Homosexual behavior between two men is just plain wrong" and

"The growing number of lesbians indicate a decline in American morals." With this









scale, separate scores for attitudes toward lesbian persons and attitudes toward gay

persons can be computed by averaging or adding ATL or ATG items, respectively.

However, given that the focus of the present study is on anti-LG attitudes in general, and

that ATL and ATG scores are highly correlated (r = .88), overall ATLG scores were used

in the present study. Appropriate items were reverse scored and item ratings were

averaged to yield an overall ATLG score with higher scores indicating more negative

attitudes. Construct validity for ATLG scores has been demonstrated through consistently

high correlations with variables associated conceptually with anti-LG attitudes such as

dogmatism, conservative political ideology, and lack of personal contact with LG persons

(Herek, 1988; Whitley & Lee, 2000). ATLG scores have yielded satisfactory levels of

internal consistency. More specifically, for scores on the combined scale, attitudes toward

lesbians subscale, and attitudes toward gay men subscale, Cronbach's alphas were .95,

.90, and .90 respectively with Herek's (1988) sample of 405 college students. In the

sample for the present study, the Cronbach's alpha for ATL scores was .90 for men, .92

for women, and .91 for the combined sample. ATG scores yielded a Cronbach's alpha of

.93 for men, women, and combined samples. Finally, overall ATLG scores (used in the

analyses) yielded a Cronbach's alpha of .95 for men, .96 for women, and a .96 for the

combined sample.

Behaviors toward LG persons were assessed with the Self-Report of Behavior

Scale-Revised (SRBS-R). The SRBS-R is a 30 item, 5-point Likert-type scale with item

ratings ranging from 1 = never to 5 = always. Item ratings were averaged to yield an

overall scale score and higher scores indicate more negative behaviors. The scale was

originally created by Patel, Long, McCammon, and Wuensch (1995) and then revised by









Roderick, McCammon, Long, and Allred (1998). It was designed to measure the extent to

which respondents report having engaged in anti-LG behaviors in the past. A factor

analysis demonstrated that the SRBS-R consisted of two factors that reflected passive

avoidance, or moving-away-from behaviors and active aggression, or moving-against

behaviors. Although the original version of the SRBS-R refers to "a gay person" as the

target of the behaviors, the current study asked participants to report their behaviors

separately for lesbians and gay men as used by Whitley (2001). Sample items include: "I

have spread negative talk about someone because she was a lesbian," and "I have

participated in playing jokes on someone because I suspected he was gay." In addition,

validity is demonstrated by the relationship between SRBS-R scores and negative

affective responses to gay persons as measured by the Index of Attitudes Toward

Homosexuals scale (IAH; Roderick et al., 1998). Furthermore, SRBS-R scores are related

to Scale 9 of the Minnesota Muliphasic Personality Inventory (MMPI), which measures

overactivity, emotional liability, hostility, aggressive outbursts, and thrill seeking (Patel

et al., 1995). In the current sample, for both men and women, SRBS-R scores were not

significantly related to scores on the Balanced Inventory of Desirable Responding

(BIDR; Paulhus, 1994). SRBS-R scores have good internal consistency, with Cronbach's

alphas of .88 for the lesbian scale and .93 for the gay male scale (Whitley, 2001). Whitley

reported that anti-lesbian behaviors and anti-gay behaviors were strongly correlated (r =

.41) and loaded strongly on a single factor in a factor analysis of attitudinal and behavior

measures. Thus, Whitley (2001) argued that the two subscales should be combined into a

single scale score labeled anti-LG behavior, as was done in the present study. The









obtained alpha internal consistency reliability estimate for SRBS-R scores with the

current sample was .93 for men, .89 for women, and .91 for the combined sample.

Predictor Variables

Right-wing authoritarianism was measured with the Right-Wing

Authoritarianism (RWA) scale. Developed by Altemeyer (1981, 1988), this scale assesses

authoritarian submission, authoritarian aggression, and conventionalism. It is a 30-item,

9-point Likert-type scale, with item ratings ranging from -4 = very strongly disagree to 4

= very strongly agree, as recommended by Altemeyer (1998). Item ratings were

transposed to a 1-9 scale and then averaged to yield an overall score, with higher scores

indicating higher levels of RWA. Sample items include: "Obedience and respect for

authority are the most important virtues children should learn," and "What our country

really needs is a strong, determined leader who will crush evil and take us back to our

true path." One item assesses attitudes toward LG persons. Consistent with prior research

(Whitley & Egisdottir, 2000; Whitley & Lee, 2000), this item was not scored in the

present study to avoid an inflated correlation with the ATLG due to item overlap. With

regard to validity, Altemeyer (1981) demonstrated that RWA scores were correlated

positively with scores on other measures of authoritarian personality including the F scale

(Adorno et al., 1950), the Dogmatism scale (Rokeach, 1960), the Conservatism scale

(Wilson & Paterson, 1968), the Balanced F scale (Kohn, 1972), and the Authoritarianism

Rebellion scale. Additionally, RWA scores were significantly correlated with scores on

theoretically related variables such as orientation to established authority and the law,

acceptance of the law as the basis of morality, and punitiveness toward sanctioned targets

(Altemeyer, 1981, 1988, 1996). In terms of reliability, RWA scores have yielded









Cronbach's alphas ranging from .81 to .95 (Hill & Wood, 1999). In the current sample,

the obtained alphas were .95 for the men, women, and combined samples.

Social dominance orientation was assessed with the Social Dominance

Orientation (SDO) scale. Pratto et al. (1994) developed this scale to measure

respondents' preference for inequality among social groups. The SDO is a 14 item, 7-

point Likert-type scale and participants rate their responses on a range from 1 = very

negative to 7 = very positive. A scale score was computed by averaging item ratings and

higher scores indicate greater levels of SDO. Sample items include: "It's OK if some

groups have more of a chance in life than others," and "We would have fewer problems if

we treated people more equally." With regard to validity, SDO scores have been shown

to correlate positively with cultural elitism, ethnic prejudice, sexism, political-economic

conservatism, and a desire to end affirmative action, thus demonstrating a high degree of

construct validity (Pratto et al., 1994). Additionally, SDO scores have yielded high

reliability estimates with an average Cronbach's alpha of .90 across multiple samples

(Pratto et al., 1994). In the current sample, alpha was .87 for men, .87 for women, and .88

for the combined sample.

Traditional gender role attitudes was measured with two scales, the Anti-

Femininity subscale of the Male Role Norms Scale (MRNS) and the Anti-Masculinity

subscale of the Female Role Norms Scale (FRNS). These scales assess the degree to

which participants endorse traditional gender roles, that women should abide by

traditional female-role norms and that men should abide by traditional male-role norms.

Thompson and Pleck (1986) developed the Anti-Femininity subscale to assess the degree

to which participants endorsed the traditional male role norm that men should avoid









anything feminine. The scale consists of 7 items and uses a 7-point Likert-type scale,

with item ratings ranging from 1 = very strongly disagree to 7 = very strongly agree. A

scale score was computed by averaging item ratings and higher scores indicate more

traditional masculine ideology. Sample items include: "It bothers me when a man does

anything that I consider 'feminine'," and "I might find it a little silly or embarrassing if a

male friend of mine cried over a sad love scene in a movie." Although this scale was

originally normed on men, it has since been used with women (e.g., Kurpius & Lucart,

2000; Whitley, 2001). With samples of men, the validity of MRNS scores is supported by

their relationship to holding an adversarial view of sexual relationships, lower levels of

self-disclosure, and a greater number of reported sexual encounters (Sinn, 1997). With a

sample that included women and men, MRNS scores were related positively to AWS

scores and endorsement of traditional gender roles for oneself (Whitley, 2001).

Furthermore, MRNS scores were minimally correlated with the masculinity and

femininity scales of the Bem Sex Role Inventory (Bem, 1974), demonstrating that the

MRNS assesses a construct that is unique from trait based masculinity and femininity

(Sinn, 1997). In addition, MRNS scores have good internal consistency with a

Cronbach's alpha of .88 for a male sample (Sinn, 1997). Of the two studies that used the

MRNS with women, one did not report reliability information (Kurpius & Lucart, 2000).

The other reported an alpha of .87 for a combined sample of men and women and for the

total MRNS score including the Toughness and Status Seeking subscales. With the

current sample, alpha for MRNS scores was .80 for men, .77 for women, and .82 for the

combined sample.









The Anti-Masculinity scale assesses the degree to which respondents believe that

women should abide by traditional female role norms by avoiding anything masculine

(Lefkowitz, Shearer, Gillen, & Espinosa-Hernandez, 2003). The scale consists of 7 items

and uses a 7-point Likert-type scale with item ratings ranging from 1 = strongly disagree

to 7 = strongly agree. An overall score was computed by averaging the item ratings and

higher scores indicate more traditional gender role attitudes. Sample items include: "A

woman whose hobbies are fishing, fixing cars, and watching sports probably wouldn't

appeal to me," and "Unless she was really desperate, I would probably advise a woman to

keep looking rather than accept a job as a construction worker." In this sample, scores on

the Anti-Masculinity scale were significantly correlated with scores on the Anti-

Femininity scale (r = .55 for women; r = .71 for men), and in other samples to beliefs

about males' and females' roles in the family such as marital and childrearing roles, and

endorsement of a sexual double standard (Lefkowitz et al., 2003). Anti-Masculinity

scores yielded Cronbach's alpha of .79 with a sample of males and females (Lefkowitz et

al., 2003). In the current sample, FRNS scores yielded an alpha of .66 for men, .78 for

women, and .75 for the combined sample.

Since there is theoretical evidence that an overall endorsement of traditional gender

role attitudes for the roles of women and men is related to anti-LG attitudes and

behaviors, in this study an overall composite score of traditional gender role attitudes was

created by computing the average of Anti-Femininity and Anti-Masculinity item ratings.

Evidence of reliability for the composite traditional gender role attitudes scores was

obtained by examining item-total correlations and the alpha internal consistency

reliability estimate for the sample. With the current sample, the composite scale scores









yielded a Cronbach's alpha of .82 for men, .86 for women, and .87 for the combined

sample. The range of corrected item-total correlation was .05-.61 for men (M= .45), .33-

.70 for women (M = .48), and .29-.64 for the combined sample (M= .53). The item with

the lowest corrected item-total correlation of .05 for men was "I might find it a little silly

or embarrassing if a female friend of mine shouted at the television during a football

game." With this item removed, the range of corrected item-total correlations would have

been .21-.61, but overall alpha would not have improved substantially (alpha = .84 if the

item is deleted). Furthermore, this item had a corrected item-total correlation of .41 for

women and .29 for the combined sample. Therefore, this item was retained in computing

overall traditional gender role attitudes scores.

Based on theoretical conceptualizations and past research, positive correlations

between traditional gender role attitudes and RWA and SDO as well as non-significant or

small correlations with social desirability would support the construct validity of scores

on the composite measure. As such, validity of composite traditional gender role attitudes

scores was supported by significant positive correlations with RWA (r = .39 for men and

.44 for women) and SDO (r = .29 for men and .40 for women) scores (convergent

validity) and non-significant correlations with the Balanced Inventory of Desirable

Responding (BIDR; Paulhus, 1994) for men and the combined sample and a negative

correlation for women (r = -.19).

Openness to experience was measured with the International Personality Item

Pool (IPIP) openness scale. The IPIP was designed by Goldberg (1999) and is available

through the website: http://ipip.ori.org. Like the NEO-PI-R (Costa & McCrae, 1992), the

IPIP scales assess the personality dimensions of the Five Factor Model (International









Personality Item Pool, 2001). The 20-item openness to experience subscale was used in

the present study. Respondents rated how accurately each item described them on a five-

point scale from 1=very inaccurate to 5=very accurate. Sample items include: "Have

difficulty understanding abstract ideas" and "Enjoy hearing new ideas." IPIP Openness

scores correlate highly (r = .83) with the corresponding Openness scores on the NEO-PI-

R (International Personality Item Pool, 2001). In addition, subscale scores of the IPIP's

openness scale correlate with corresponding subscale scores of the Sixteen Personality

Factors Questionnaire (16PF; Conn & Rieke, 1994); more specifically, Intellect scores

correlate to Reasoning at r = .51, Imagination scores correlate with Abstractedness at r =

.65, and Emotionality scores correlate with Tension at r = .59 (International Personality

Item Pool, 2001). Openness scores yielded a Cronbach's alpha of .89 in previous research

(International Personality Item Pool, 2001) and with the current sample, alphas were .87

for men, .88 for women, and .88 for the combined sample.















CHAPTER 4
RESULTS

Preliminary analyses were conducted to explore potential gender differences,

examine descriptive statistics, and test potential order effects. Tests of hypotheses were

conducted following these preliminary analyses.

Gender Differences

Examination of correlation matrices for variables of interest by gender revealed a

few differences in the pattern of intercorrelations for women and men (see Tables 1 and

2).

Table 1. Zero-Order Correlations Among Variables of Interest for Women
Variables 1 2 3 4 5
1. Negative Attitudes Toward LG
Persons
2. Negative Behaviors toward LG .53**
persons
3. Right-Wing Authoritarianism .79** .34**
4. Social Dominance Orientation .19* .22* .36**
5. Traditional Gender Role Attitudes .46** .47** .44** .40**
6. Openness to Experience -.23** -.23** -.35** -.34** -.52**
Note. *p < .05. **p <.001.


Table 2. Zero-Order Correlations Among Variables of Interest for Men
Variables 1 2 3 4 5
1. Negative Attitudes Toward LG Persons
2. Negative Behaviors Toward LG persons .47**
3. Right-Wing Authoritarianism .79** .34**
4. Social Dominance Orientation .17 .24* .08
5. Traditional Gender Role Attitudes .50** .50** .37** .29**
6. Openness to Experience -.37** -.37** -.45** -.16 -.32**
Note. *p < .05. **p <.001.









Previous research also has found gender differences in levels of negative attitudes

towards lesbian and gay (LG) persons (Kerns & Fine, 1994; Kite & Whitley, 1996;

LaMar & Kite, 1998) and in relations among predictors of negative attitudes towards LG

persons (Basow & Johnson, 2000). Due to the pattern of gender difference in the present

data and prior literature, all analyses were conducted separately for women and men in

order to ensure that combining all participants into one sample did not mask unique

patterns of findings for women and men.

Descriptive Statistics

Although the present sample's scores on some of the variables of interest

indicated some restriction in range (see Tables 3 and 4), means and standard deviations

for the present sample were generally comparable to those obtained using the same

instruments in previous samples of undergraduate students.

Table 3. Summary Statistics Among Variables of Interest for Women
Variables Possible Sample M SD oc
Range Range
1. Negative Attitudes Toward LG 1.00-5.00 1.00-4.80 2.25 .95 .96
Persons
2. Negative Behaviors toward LG 1.00-5.00 1.00-2.90 1.23 .28 .89
persons
3. Right-Wing Authoritarianism 1.00-9.00 1.03-6.55 3.82 1.53 .96
4. Social Dominance Orientation 1.00-7.00 1.00-4.64 2.34 .86 .87
5. Traditional Gender Role Attitudes 1.00-7.00 1.00-5.00 2.53 .92 .86
6. Openness to Experience 1.00-5.00 2.45-4.95 4.01 .55 .88
Note. Higher scores indicate higher levels of the construct assessed.

More specifically, levels of anti-LG attitudes were generally close to the mid range

of possible scores for both women and men and the current sample's means and standard

deviations for anti-LG attitudes (women: M= 2.25, SD = .95; men: M = 2.59, SD = .91)

were comparable to those reported by Kerns and Fine (1994) in a sample of

undergraduate women (M= 2.88, SD = .80) and men (M= 3.26, SD = .99). Relatively









few anti-LG behaviors were reported in the current sample by women (M = 1.23, SD =

.28) and men (M= 1.42, SD = .42). These low scores are comparable to the scores

reported by Whitley (2002) with an undergraduate university sample of women (M=

1.25, SD = .28) and men (M= 1.55, SD = .56). Scores for right-wing authoritarianism

(RWA) were below the possible mid-point demonstrating lower levels of the variable. In

addition, this sample's scores for RWA (women: M= 3.82, SD = 1.53; men: 3.92, SD =

1.40), were lower than those reported by Whitley and Egisd6ttir (2000) for undergraduate

women (M= 5.28, SD = 1.00) and men (M= 5.03, SD 1.18), but comparable to those

reported by Basow and Johnson (2000) for undergraduate women (M = 3.95, no SD

reported) and those reported by Butler (2000) for undergraduate men and women (M=

4.10, SD = 1.00). Like RWA scores, social dominance orientation (SDO) scores also

were below the possible mid-point demonstrating lower levels of this variable.

Nevertheless, the current sample's means and standard deviations for SDO scores

(women: M= 2.34, SD = .86; men: M= 2.99, SD = .99) were comparable to those

reported by Whitley (1999) with a sample of undergraduate women (M= 2.79, SD = .96)

and men (M= 3.46, SD = 1.17).

Table 4. Summary Statistics Among Variables of Interest for Men
Variables Possible Sample M SD oc
Range Range
1. Negative Attitudes Toward LG 1.00-5.00 1.00-4.78 2.59 .91 .95
Persons
2. Negative Behaviors toward LG 1.00-5.00 1.00-3.10 1.42 .42 .93
persons
3. Right-Wing Authoritarianism 1.00-9.00 1.00-7.45 3.92 1.40 .95
4. Social Dominance Orientation 1.00-7.00 1.00-5.57 2.99 .99 .87
5. Traditional Gender Role Attitudes 1.00-7.00 1.00-5.57 3.29 .86 .82
6. Openness to Experience 1.00-5.00 2.45-4.95 3.75 .56 .87
Note. Higher scores indicate higher levels of the construct assessed.









Since the traditional gender role attitudes measure used in the present study was a

combination of previously used scales, the present sample's scores on overall traditional

gender role attitudes (women: M 2.53, SD = .92; men: M= 3.28, SD = .86) could not be

compared directly to that of prior samples. However, this sample's scores on the

Antifeminity subscale (women: M= 2.80, SD = 1.04; men: M= 3.76; SD = 1.14), which

comprised half of the items in the composite measure, were comparable to Kurpius and

Lucart's (2000) scores for a sample of college students (women: M= 2.67, SD = 1.02;

men: M = 3.36, SD = .80). Past research has not reported means and standard deviations

for the Antimasculinity subscale of the Female Role Norms Scale, which comprised the

second half of the traditional gender role attitudes scale. Lastly, with respect to openness

to experience, scores were generally close to the mid range of possible scores for both

women and men. The current sample's means and standard deviations for openness to

experience (women: M = 4.01, SD = .55; men: M= 3.75, SD = .56) were comparable to

those reported by Chan, Rounds and Drasgow (2000) for openness with a combined

sample of college women and men (M= 3.79, SD = .55). Overall, the pattern of similarity

of descriptive statistics for the present sample, and that of previous samples of

undergraduate students, suggests that any restriction in range of variables of interest

found in the present sample likely reflects the range of attitudes among undergraduate

students rather than pointing to sample specific idiosyncrasies.

Test for Order Effects

To test for order effects across the two orders of the survey, a MANOVA was

conducted with survey order as the independent variable and the variables of interest (i.e.,

anti-LG attitudes, anti-LG behaviors, RWA, SDO, traditional gender role attitudes, and

openness to experience) as dependent variables. The overall model was not significant (F









[3, 229] = .77, p > .05, r,= .20), demonstrating that survey order did not significantly

impact participants' responses.

Intercorrelations Among Variables of Interest: Hypothesis 1

Zero-order correlations were computed to test hypothesized relations among

variables of interest. More specifically, Hypothesis 1 stated that (a) RWA, SDO, and

traditional gender role attitudes will be significantly and positively related to anti-LG

attitudes and behaviors and (b) openness to experience will be significantly and

negatively related to anti-LG attitudes and behaviors. Analyses were conducted

separately for women and men participants.

Intercorrelations for women. For women, Hypothesis 1 was fully supported (see

Table 1). As expected, anti-LG attitudes were correlated positively with RWA (r = .79, p

< .001), SDO (r = .19,p < .05), and traditional gender role attitudes (r = .46, p < .001),

and negatively with openness to experience (r = -.23,p < .001). Additionally, anti-LG

behaviors were correlated positively with RWA (r = .34, p < .001), SDO (r = .22, p <

.05), and traditional gender role attitudes (r = .47, p < .001), and negatively with

openness to experience (r = -.23, p < .01). Furthermore, openness to experience was

negatively correlated with RWA (r = -.35, p < .001), SDO (r = -.34, p < .001), and

traditional gender role attitudes (r = -.52, p < .001).

Intercorrelations for men. For the men, Hypothesis 1 was partially supported

(see Table 2). Anti-LG attitudes were correlated positively with RWA (r = .79, p < .001)

and traditional gender role attitudes (r = .50, p < .001), and negatively with openness to

experience (r = -.37, p < .001). Additionally, anti-LG behaviors were correlated

positively with RWA (r = .34, p < .01), SDO (r = .24, p < .05), and traditional gender









role attitudes (r = .50, p < .001), and negatively with openness to experience (r = -.37, p

< .001). Furthermore, openness to experience was correlated negatively with RWA (r = -

.45,p < .001) and traditional gender role attitudes (r = -.32,p < .01). However,

inconsistent with Hypothesis 1, anti-LG attitudes were not correlated significantly with

SDO, and SDO was not correlated significantly with openness to experience.

Tests of Mediation: Hypotheses 2 and 3

Hypothesis 2 stated that openness to experience would either partially or fully

mediate the relations of RWA, SDO, and traditional gender role attitudes to anti-LG

attitudes. In addition, Hypothesis 3 stated that anti-LG attitudes will either partially or

fully mediate the relations of RWA, SDO, traditional gender role attitudes, and openness

to anti-LG behaviors. To test the mediations proposed in these hypotheses, Baron and

Kenny's (1986) procedures were followed. According to these authors, for a variable to

be considered as a mediator, significant relations must exist between (a) the predictor and

the mediator, (b) the mediator and the criterion, and (c) the predictor and criterion.

According to Baron and Kenny (1986), if preconditions are satisfied, a variable acts as a

mediator to the extent that it accounts for the relationship between the predictor and

criterion. In order to test the significance of mediations, Amos 5.0 (Arbuckle, 2003) was

used to conduct path analyses of a fully saturated model in which all direct and indirect

paths were estimated (see the models presented in Figures 1 and 3). To test the

significance of mediation, appropriate standardized path coefficients were multiplied to

compute indirect effects, a procedure recommended by Cohen and Cohen (1983), and

Sobel's formula (1982) was used to determine whether or not the indirect effects were

significantly different from zero. Again, separate analyses were conducted for women

and men. Maximum likelihood estimation was utilized with the covariance matrix of the









variables of interest as input. Given that the models tested were fully saturated, values for

the Goodness of Fit Index (GFI), Incremental Fit Index (IFI), Comparative Fit Index

(CFI), and the Normed Fit Index (NFI) all were 1.0. The model for women accounted for

67% of the variance in anti-LG attitudes, 38% of the variance in anti-LG behaviors, and

30% of the variance in openness to experience. The model for men accounted for 67% of

the variance in anti-LG attitudes, 35% of the variance in anti-LG behaviors, and 24% of

the variance in openness to experience.

Tests of mediation with women. Hypothesis 2 proposed that openness would

mediate the links of RWA and traditional gender role attitudes to anti-LG attitudes. For

Hypothesis 2, preconditions for mediation were satisfied for the role of openness as a

mediator of the link of traditional gender roles to anti-LG attitudes, but not for openness

as a mediator of RWA and SDO to anti-LG attitudes (see Figure 1 for path coefficients).

That is, zero order correlations indicated that traditional gender role attitudes (predictor

variable) were correlated significantly with anti-LG attitudes (criterion). Furthermore, the

path model indicated that traditional gender role attitudes (predictor variable) were

related significantly and uniquely to openness (potential mediator), which in turn was

related significantly and uniquely with anti-LG attitudes (criterion).

Sobel's test indicated, however, that the indirect link of traditional gender role

attitude to anti-LG attitudes through openness was not significant. Given that in the full

model preconditions for mediation were not met for RWA or SDO (i.e., neither variable

was related uniquely to openness) and the indirect effect was not significant for

traditional gender role attitudes, openness did not act as a mediator and thus Hypothesis 2

was not supported.

































Figure 1. Fully saturated model depicting relationships among variables of
interest for women. Values reflect standardized coefficients. *p < .05. **p <
.001.

Hypothesis 3 proposed that anti-LG attitudes would fully or partially mediate the

links of RWA, SDO, traditional gender role attitudes, and openness to anti-LG behaviors.

The preconditions for mediation were satisfied for the role of anti-LG attitudes as a

mediator of the links of RWA, SDO, and traditional gender role attitudes to anti-LG

behaviors, but not for anti-LG attitudes as a mediator of openness to anti-LG behaviors

(see Figure 1 for path coefficients). That is, zero order correlations indicated that RWA,

SDO, and traditional gender role attitudes, but not openness (predictor variables) were

correlated significantly with anti-LG behaviors (criterion). Furthermore, the path model

indicated that RWA, SDO and traditional gender role attitudes (predictor variables) were

related significantly and uniquely to anti-LG attitudes (potential mediator), which in turn

was related to anti-LG behaviors (criterion).









Multiplication of appropriate standardized path coefficients and Sobel's formula

indicated that consistent with Hypothesis 3, through anti-LG attitudes, a significant

indirect link of .50 (.78 x .64; z = 5.00, p < .001) was obtained from RWA to anti-LG

behaviors. Since there was also a significant direct relationship between RWA and anti-

LG behaviors, anti-LG attitudes acted as a partial mediator. Additionally, through anti-

LG attitudes, a significant indirect link of -.09 (-.14 x .64; z = -2.25, p < .05) was

obtained from SDO to anti-LG behaviors. Since there was not a significant direct

relationship between SDO and anti-LG behaviors, anti-LG attitudes acted as a full

mediator. Also, through anti-LG attitudes, a significant indirect link of .15 (.24 x .64; z =

3.04, p < .01) was obtained from traditional gender role attitudes to anti-LG behaviors.

Since there was a significant direct relationship between traditional gender role attitudes

and anti-LG behaviors, anti-LG attitudes acted as a partial mediator. Therefore,

Hypothesis 3 was partially supported in that anti-LG attitudes partially mediated the links

of RWA and traditional gender role attitudes to anti-LG behaviors and fully mediated the

links of SDO to anti-LG behaviors, but anti-LG attitudes did not mediate the link of

openness to anti-LG behaviors.

In the model for women, there were three significant unique links that were in the

unexpected directions. There was a positive significant unique link from openness to anti-

LG attitudes (P= .12, p < .05), a significant negative unique link from SDO to anti-LG

attitudes (P = -.14, p < .05), and a significant negative unique link from RWA to anti-LG

behaviors ( = -.33,p < .01).



































Figure 2. Trimmed model depicting relationships among variables of interest
for women. Values reflect standardized coefficients. All paths depicted are
significant at p < .05.

Finally, the fit of a fully saturated model was compared to that of an alternative

trimmed model that eliminated the non-significant direct paths (a) from RWA to

openness, (b) from SDO to openness, (c) from SDO to anti-LG behaviors, and (d) from

openness to anti-LG behaviors (see Figure 2 for the trimmed model). The goodness of fit

indices for this model were above the acceptable cut offs and nearly identical to those

obtained from the original model (GFI = .98; IFI = .99; CFI = .99; NFI = .98). The

trimmed model explained 68% of the variance in anti-LG attitudes, 38% of the variance

in anti-LG behaviors, and 27% of the variance in openness. As indicated in Figures 1 and

2, the amounts of variance accounted for in endogenous variables as well as significance









decisions and magnitudes of paths coefficients were similar across the fully saturated and

trimmed models. Specifically, in the trimmed model the paths from traditional gender

role attitudes to openness increased from -.41 to -.52, from RWA to anti-LG attitudes

decreased from .78 to .77, from traditional gender role attitudes to anti-LG behaviors

increased from .27 to .32, and from RWA to anti-LG behaviors decreased from -.33 to -

.27. Thus, compared with the fully saturated model, the trimmed model appears to be

more parsimonious but equally appropriate in explaining the relationships among the

variables of interest.

Tests of mediation with men. To test the mediations proposed in Hypotheses 2

and 3 for men in the sample, Baron and Kenny's (1986)'s procedures, multiplication of

appropriate standard path coefficients and Sobel's formula were used as outlined above.

Hypothesis 2 stated that openness to experience would either partially or fully mediate

the relationships of RWA, SDO, and traditional gender role attitudes to anti-LG attitudes.

The precondition that openness (potential mediator) must be uniquely related to anti-LG

attitudes (criterion) was not met (see Figure 3), thus Hypothesis 2 was not supported.

With regard to Hypothesis 3, preconditions for mediation were satisfied for the

role of anti-LG attitudes as a mediator of the links of RWA, and traditional gender role

attitudes to anti-LG behaviors, but not for anti-LG attitudes as a mediator of SDO and

openness to anti-LG behaviors (see Figure 1 for path coefficients). That is, zero order

correlations indicated that RWA and traditional gender role attitudes (predictor variables)

were correlated significantly with anti-LG behaviors (criterion). Furthermore, the path

model indicated that RWA and traditional gender role attitudes, but not SDO or openness

(predictor variables) were related significantly and uniquely to anti-LG attitudes









(potential mediator), which in turn was related significantly and uniquely with anti-LG

behaviors (criterion).


Figure 3. Fully saturated model depicting relationships among variables of interest for
men. Values reflect standardized coefficients. *p < .05. **p < .001.

Consistent with Hypothesis 3, through anti-LG attitudes, a significant indirect link

of .24 (.71 x .34; z = 2.31,p < .05) was obtained between RWA and anti-LG behaviors.

Because no significant direct link existed between reported RWA and anti-LG behaviors,

anti-LG attitudes acted as a full mediator of this link. However, through anti-LG attitudes

there was not a significant indirect link between traditional gender role attitudes or

openness and anti-LG behaviors. Therefore, Hypothesis 3 was partially supported in that









anti-LG attitudes fully mediated the link of RWA to anti-LG behaviors, but not the links

of SDO, traditional gender role attitudes, or openness to anti-LG behaviors.

Finally, the fit of the fully saturated model was compared to that of an alternative

trimmed model that eliminated the non-significant direct paths (a) from SDO and

traditional gender role attitudes to openness, (b) from SDO and openness to anti-LG

attitudes, and (c) from RWA and SDO to anti-LG behaviors (see Figure 4 for the trimmed

model). The goodness of fit indices for this model were above the acceptable cut offs and

nearly identical to those obtained from the original model (GFI= .98; IFI = 1.00; CFI =

1.00; NFI = .97). The trimmed model explained 67% of the variance in anti-LG attitudes,

33% of the variance in anti-LG behaviors, and 20% of the variance in openness. As

indicated in Figures 3 and 4, the amounts of variance accounted for in endogenous

variables as well as significance decisions and magnitudes of paths coefficients were

similar across the fully saturated and trimmed models. Specifically, in the trimmed model

the paths from RWA to openness increased from -.39 to -.45, from traditional gender role

attitudes to anti-LG attitudes increased from .24 to .25, from RWA to anti-LG attitudes

decreased from .71 to .70, from traditional gender role attitudes to anti-LG behaviors

increased from .29 to .32, from openness to anti-LG behaviors decreased from -.20 to -

.17, and from anti-LG attitudes to anti-LG behaviors decreased from .34 to .25. Thus,

compared with the fully saturated model, the trimmed model appears to be more

parsimonious, but equally appropriate in explaining the relationships among the variables

of interest.
































Figure 4. Trimmed model depicting relationships among variables of interest
for men. Values reflect standardized coefficients. All paths depicted are
significant at p < .05.














CHAPTER 5
DISCUSSION

Numerous studies have examined separately the relations of right-wing

authoritarianism (RWA), social dominance orientation (SDO), traditional gender role

attitudes, and openness to experience with anti-lesbian and gay (LG) attitudes. However,

no empirical study has examined this set of variables together to determine the relative

strength of their associations with anti-LG attitudes. The present study addressed this gap

and also extended the literature by examining the relations of this set of variables with

self-reported anti-LG behaviors in addition to anti-LG attitudes. Furthermore, this study

provided the first test of openness to experience as a potential mediator of the links of

RWA, SDO, and traditional gender role attitudes to anti-LG attitudes, and the first test of

anti-LG attitudes as a mediator of the links of RWA, SDO, traditional gender role

attitudes, and openness to anti-LG behaviors. By examining concomitantly the relations

among RWA, SDO, traditional gender role attitudes, openness, anti-LG attitudes, and

anti-LG behaviors, the present findings build on prior research to identify the strongest

correlates (among individual difference variables) of anti-LG attitudes and behaviors.

Identifying the most critical correlates will be important in informing future research and

interventions that aim to reduce anti-LG attitudes and behaviors.

In order to reduce anti-LG attitudes and behaviors, it is important to identify the

variables that are most strongly related to these constructs. Of the individual difference

variables examined in the present study, RWA emerged as the strongest unique correlate

of anti-LG attitudes for women and men and also had a substantial relation with anti-LG









behaviors. More specifically, RWA had large zero-order correlations with anti-LG

attitudes (r = .79) for both women and men. Furthermore, in path models, once

interrelations among all exogenous variables were accounted for, RWA again had the

strongest relation with anti-LG attitudes for both women (3 = .78) and men (3 = .71).

Additionally, RWA had a significant indirect relation to anti-LG behaviors through anti-

LG attitudes for both women and men. For men, anti-LG attitudes fully mediated the link

of RWA to anti-LG behaviors. That is through anti-LG attitudes, a significant indirect

link of .24 was obtained between RWA and anti-LG behaviors. For women, anti-LG

attitudes partially mediated the link of RWA to anti-LG behaviors. That is, through anti-

LG attitudes, a significant indirect link of .50 was obtained between RWA and anti-LG

behaviors.

Thus, overall, the role of RWA was strong and as expected. An unexpected relation

also was found, however, between RWA and anti-LG behaviors for women. Although

there was a significant positive zero-order correlation between RWA and anti-LG

behaviors (r = .34), a significant negative unique relation (P = -.33) emerged between

RWA and anti-LG behaviors when all variables of interest were entered into the path

model. Since anti-LG attitudes partially mediated the link of RWA to anti-LG behaviors,

a portion of the positive link of RWA to anti-LG behaviors occurred through anti-LG

attitudes, but the additional unique direct link of RWA to anti-LG behaviors was negative

for women.

One possible explanation for the negative unique link of RWA with anti-LG

behaviors (after other variables were accounted for) emerges from the theoretical

conceptualization of RWA. Theoretically, RWA is comprised of three attitudinal clusters









including authoritarian submission, authoritarian aggression, and conventionalism

(Altemeyer, 1981). Altemeyer (1981) theorized that aggression is manifested as the

predisposition to inflict physical, psychological, economic, or social harm. However,

most studies of RWA have examined its link with prejudicial attitudes but not with

prejudicial aggressive behaviors (e.g., Altemeyer, 1998, 2001; Heaven & St. Quintin,

2002; Whitley, 1999). Therefore, there is a lack of clarity about the link between RWA

and aggressive behaviors, although such a link would be consistent with the

conceptualization of the second attitudinal cluster of authoritarian aggression. Additional

empirical research is needed to determine the degree to which persons with high levels of

RWA are also aggressive. Attention should be paid to the type of aggression exhibited by

persons high in RWA, and the range of situational variables that can impact aggressive

behaviors. The present study suggests that after accounting for the indirect link of RWA

to anti-LG behaviors through anti-LG attitudes, for women, RWA negatively relates to

anti-LG behaviors. Perhaps this is because RWA also reflects an adherence to authorities

and conventionalism that does not coincide with aggressive or even criminal behaviors

such as writing graffiti, yelling insulting comments, and hitting an LG person.

Furthermore, the significant negative relationship with anti-LG behaviors was only found

with women and not men. Perhaps there are gender differences in the expression of high

RWA, with women acting less physically aggressive than men because physical

aggression may contradict the conventional view of women, and adherence to

conventional roles is an important part of RWA. Thus, future empirical research is

needed to examine potential gender differences in the expression of the aggressive

component of RWA.









The second variable that was demonstrated to have an important relation with both

anti-LG attitudes and behaviors was traditional gender role attitudes. Next to RWA, this

variable had the strongest links with anti-LG attitudes in both zero order correlations for

women (r = .46) and men (r = .50) and in the path model (P = .24) for both women and

men. Furthermore, traditional gender role attitudes had the strongest links with anti-LG

behaviors for both women and men. More specifically, traditional gender role attitudes

had large zero order correlations with anti-LG behaviors for both women (r = .47) and

men (r = .50). Furthermore, for women, traditional gender role attitudes had a significant

indirect link (through the partial mediating role of anti-LG attitudes) of. 15 and additional

unique direct link of .27 with anti-LG behaviors. For men, there was no significant

mediation through anti-LG attitudes, but traditional gender role attitudes had a unique

direct link of .29 with anti-LG behaviors. Thus, overall, traditional gender role attitudes

emerged as having significant unique relations with anti-LG attitudes and behaviors for

both women and men. These findings point to the important unique role of traditional

gender role attitudes in shaping anti-LG attitudes and behaviors.

In addition to the role of RWA and traditional gender role attitudes in anti-LG

attitudes and behaviors, the present findings provide further support for the connection

between anti-LG attitudes and behaviors. More specifically, the path analysis showed that

the relation between anti-LG attitudes and self-reported behaviors was strong for women

(P = .64) and moderate for men (P = .34). Perhaps situational variables play a larger role

in shaping men's anti-LG behaviors, whereas attitudes seem fairly strongly related to

self-reported behaviors for women. Also, Franklin (2000) found that the gender gap in

anti-LG behaviors grew wider as the specific behaviors became more violent, with









women reporting much lower levels of more violent behaviors than men. Perhaps as anti-

LG behaviors become more violent, the relation between anti-LG attitudes and behaviors

decreases, particularly for men because other individual difference and situational

variables play a more important role in shaping violent behavior. Future research could

explore if the relation between anti-LG attitudes and behaviors is different for persons

engaging in non-violent versus violent anti-LG behaviors. Future research could also

examine the magnitude of gender difference in the relation of anti-LG attitudes across

various levels of violence of those behaviors.

The role of anti-LG attitudes in shaping the links of individual difference variables

to anti-LG behaviors was supported. More specifically, for women, anti-LG attitudes

partially mediated the links of RWA (indirect link of .49) and traditional gender role

attitudes (indirect link of. 14) to anti-LG behaviors and fully mediated the link of SDO

(indirect link of -.09) to anti-LG behaviors. Similarly, for men, anti-LG attitudes fully

mediated the link of RWA (indirect link of .25) to anti-LG behaviors. This shows that the

strongest correlates of anti-LG behaviors had indirect links to anti-LG behaviors through

anti-LG attitudes. This pattern of findings demonstrates the importance of anti-LG

attitudes in understanding anti-LG behaviors. Longitudinal and experimental studies can

build on the present findings and evaluate the direction and causal role of anti-LG

attitudes in shaping anti-LG behaviors. Future research is needed to examine whether

lowering anti-LG attitudes does indeed lead to a reduction of anti-LG behaviors.

One important pattern in the present findings is the apparent minimal role of

openness to experience in anti-LG attitudes and behaviors. In relation to anti-LG attitudes

and behaviors, openness had small zero-order correlations for women with anti-LG









attitudes (r = -.23) and anti-LG behaviors (r = -.23) and moderate zero-order correlations

for men with anti-LG attitudes (r = -.37) and anti-LG behaviors(r = -.37). However, once

all predictor variables were entered into the path model, the relation of openness to anti-

LG attitudes was not significant for men, and was minimal and in the opposite direction

than expected for women (P = .12). Perhaps, once the shared variance of openness with

the other variables is taken into account, the positive relation with anti-LG attitudes could

reflect a greater openness and willingness to report negative attitudes. Similarly, in the

path model, openness was not significantly uniquely related with anti-LG behaviors for

women, but was uniquely related to anti-LG behaviors (P = -.20) for men (in the expected

direction). Finally, the role of openness as a mediator of the relations of RWA, SDO, and

traditional gender role attitudes to anti-LG attitudes was not supported. In fact, not only

did openness not act as a mediator, after accounting for the roles of RWA, SDO, and

traditional gender role attitudes, the links of openness to anti-LG attitudes and behaviors

was non-significant or minimal for both women and men. Thus, in the context of the

other individual difference variables examined in the present study, openness to

experience did not appear to play an important unique role in anti-LG attitudes or

behaviors.

Similar to openness, after accounting for the role of RWA and traditional gender

role attitudes, the links of SDO to anti-LG attitudes and behaviors was minimal. For men

there was no significant zero-order correlation between SDO and anti-LG attitudes and

there was a small positive relation between SDO and anti-LG behaviors (r = .24). This

small relation disappeared, however, once all of the variables were entered into the path

model. For women, there were small significant zero-order correlations between SDO









and anti-LG attitudes (r = .19) and behaviors (r = .22). Once in the path model, SDO and

anti-LG attitudes (P = -.14) had a small relation in the unexpected direction. Additionally,

for women, anti-LG attitudes mediated the relation between SDO and anti-LG behaviors

and a small indirect effect of -.09 was found (in an unexpected direction). It was expected

that persons with high levels of SDO would report greater anti-LG attitudes and

behaviors because they would view LG persons as an out-group. These relations were

indeed demonstrated through the positive zero order correlations. In the context of

variance accounted for in anti-LG attitudes and behaviors by RWA, traditional gender

role attitudes, and openness, however, SDO was related negatively to the remaining

variance in anti-LG attitudes and behaviors. Perhaps, the observed positive correlations

of SDO with anti-LG attitudes and behavior, in part, reflected the correlation with SDO

with RWA, traditional gender role attitudes, and openness. Overall, the link of SDO with

anti-LG attitudes and behaviors was minimal, especially when other individual difference

variables were considered. Thus, the present findings suggest that future empirical

research and interventions should be aimed at RWA and traditional gender role attitudes,

which demonstrated the strongest and most consistent relations with anti-LG attitudes

and behaviors.

An additional important pattern in the present findings is that there were some

differences in the pattern and magnitude of intercorrelations and mediational relations for

women and men. Previous research has consistently found gender differences in the

levels of anti-LG attitudes and behaviors; with men holding more negative attitudes and

reporting more negative behaviors than do women (e.g., Herek, 2002; Kite, 1984). In

addition, Basow and Johnson (2000) found a different pattern in relations among









predictors of anti-LG attitudes for their sample of college women than was found by

Theodore and Basow (2000) for their sample of college men. Furthermore, Whitley and

Egisd6ttir (2000) found that gender of participants had a unique link to anti-LG attitudes

in their path model that also examined RWA, SDO, and attitudes toward women.

However, studies that have examined anti-LG attitudes and behaviors typically have not

reported their findings separately by gender (e.g., Roderick et al., 1998; Whitley & Lee,

2000). Thus, it is not clear whether the gender differences in the pattern of interrelations

found in the present study also existed in previous samples. Future researchers should

explicitly explore the pattern of interrelations by gender and report their findings

separately for women and men in order to evaluate the replicability of the present

findings and ultimately tailor interventions most appropriately to each gender if needed.

The present findings must be interpreted in light of a number of limitations. For

example, the current sample consisted of students, most of whom were in their first year

of college, White, and identified as middle or upper middle class. Thus, results from the

present study may not generalize to the general population, which includes greater

diversity in terms of age, level of education, and social class. Furthermore, given that the

present sample did not target convicted criminals, the results may not generalize to

persons who have been convicted of hate crimes against LG persons. Therefore, future

research is needed to examine the replicability of the present results in a variety of

samples in order to capture the broader diversity of the general population. Such research

can also explore the extent to which the pattern of gender differences in interrelations

found in the present study will generalize to other samples.









Another limitation of the present study reflects needs for advancement in the state

of the art of measuring anti-LG behaviors. Although the present study utilized the most

appropriate measure of anti-LG behaviors available in the literature, such measurement

can be improved by addressing a number of limitations. First, although SRBS-R scores

were not correlated significantly with social desirability, it is possible that participants

did not report the full extent of negative behaviors that they have engaged in, or that they

are not consciously aware of all of their negative behaviors. For example, a person could

act rudely to an LG person by creating more personal distance or avoiding eye contact,

yet not be conscious of such actions. In addition, the fact that the SRBS-R assessed only

overt negative behaviors and crimes (e.g., verbal harassment, writing graffiti, hitting)

may have contributed to the restriction in range and floor effect of SRBS-R scores (see

Tables 3 and 4), with women scoring a mean of 1.28 and men scoring a mean of 1.42,

when the scale midpoint is 2.50. One way to address these limitations is to integrate

assessment of a continuum of behaviors from overtly negative to overtly positive (which

might include items assessing helpful behaviors such as attending a pride parade,

befriending an LG person, and supporting legislation for increased LG rights). Using a

more inclusive measure of hostile and affirming behaviors would allow for a fuller

exploration of the links of RWA, SDO, traditional gender role attitudes, openness, and

anti-LG attitudes to behaviors toward LG persons.

Another important area for advancing understanding of behaviors toward LG

persons is to build on the present findings regarding the roles of individual difference

variables by integrating attention to situational variables, such as contact with LG persons

and youthful thrill seeking, which also have been linked to anti-LG behaviors (Franklin,









1998; Schope, 2000). In the present study, individual difference variables accounted for a

much greater percentage of the variance of anti-LG attitudes than behaviors. Thus,

attention to other potential correlates of anti-LG behaviors is warranted and data from

such studies can inform the development and evaluation of interventions aimed at

reducing anti-LG behaviors.

By advancing scientific understanding of the predictors of negative attitudes and

behaviors towards LG persons, the present study can serve as groundwork for developing

theoretically and empirically based interventions that aim to reduce anti-LG attitudes and

behaviors. This study points to the importance of RWA and traditional gender role

attitudes in predicting anti-LG attitudes and behaviors of college women and men. The

present data raise questions, however, about the roles of SDO and openness in anti-LG

attitudes and behaviors. Based on these results, interventions aiming to prevent anti-LG

attitudes and behaviors could focus limited time and resources specifically on modifying

RWA and traditional gender role attitudes. Since persons with high levels of RWA are

very submissive to authority figures, a possible intervention could involve legitimate

authority figures (e.g., university presidents, religious leaders, and elected officials)

voicing affirming-LG attitudes and behaviors. Additionally, to combat traditional gender

role attitudes, interventions could focus on explaining the range of possible gender

expressions and restrictions that traditional gender role attitudes place on individuals. If

these interventions are successful at lowering RWA and traditional gender role attitudes,

then it can also be assessed if anti-LG attitudes and behaviors are also reduced.

Evaluating the impact of such interventions would also test the causal direction of

relations among RWA, traditional gender role attitudes, anti-LG attitudes, and anti-LG






73


behaviors. Thus, the present findings can provide the basis for such research and

ultimately inform prevention and educational interventions aimed at reducing anti-LG

attitudes and behaviors.
















LIST OF REFERENCES

Adorno, T. W., Frenkel-Brunswik, E., Levinson, D. J., & Sanford, R. N. (1950). The
authoritarian personality. New York: Harper & Row.

Altemeyer, B. (1981). Right-wing authoritarianism. Winnipeg, Canada: University of
Manitoba Press.

Altemeyer, B. (1988). Enemies offreedom: Understanding right-wing authoritarianism.
San Francisco: Jossey-Bass.

Altemeyer, B. (1996). The authoritarian specter. Cambridge, MA: Harvard University
Press.

Altemeyer, B. (1998). The other "authoritarian personality." Advances in Experimental
Social Psychology, 30, 47-91.

Altemeyer, B. (2001). Changes in attitudes towards homosexuals. Journal of
Homosexuality, 42(2), 63-75.

Altemeyer, B., & Hunsberger, B. (1992). Authoritarianism, religious fundamentalism,
quest, and prejudice. The International Journal for the Psychology ofReligions,
2(2), 113-133.

Arbuckle, J. L. (2003). Amos (Version 5.0) [Computer software]. Chicago, IL:
SmallWaters Corporation.

Baron, M. R. & Kenny, D. A. (1986). The moderator-mediator variable distinction in
social psychological research: Conceptual, strategic, and statistical considerations.
Journal ofPersonality and Social Psychology, 51, 1173-1182.

Basow, S. A., & Johnson, K. (2000). Predictors of homophobia in female college
students. Sex Role, 42(5/6), 391-403.

Bem, S. L. (1974). The measurement of psychological androgyny. Journal of Consulting
and Clinical Psychology, 42, 155-162.

Beat, J. A., Calhoun, K. S., Adams, H. E., & Zeichner, A. (2001). Homophobia and
physical aggression toward homosexual and heterosexual individuals. Journal of
Abnormal Psychology, 110(1), 179-187.









Butler, J. C. (2000). Personality and emotional correlates of right-wing authoritarianism.
Social Behavior and Personality, 28(1), 1-14.

Chan, K., Rounds, J., & Drasgow, F. (2000). The relation between vocational interests
and the motivation to lead. Journal of Vocational Behavior, 57, 226-245.

Cohen, J. (1988). Statistical power analysisfor the behavioral sciences (2nd ed.).
Hillsdale, NJ: Lawrence Earlbaum Associates.

Cohen, J., & Cohen, P. (1983). Applied multiple regrssion/correlation analysis for the
behavioral sciences (2nd ed.). Hillsdale, NJ: Erlbaum.

Conn, S. R., & Rieke, M. L. (1994). The 16PFfifth edition technical manual.
Champaign, IL: Institute for Personality and Ability Testing.

Costa, P. T., & McCrae, R. R. (1985). The NEO Personality Inventory manual. Odessa,
FL: Psychological Assessment Resources.

Costa, P. T., & McCrae, R. R. (1992). The NEO Personality Inventory (Revised). Odessa,
FL: Psychological Assessment Resources.

Cullen, J. M., Wright, L. W., & Alessandri, M. (2002). The personality variable openness
to experience as it relates to homophobia. Journal ofHomosexuality, 42, 119-133.

D'Augelli, A. R. (1989). Lesbians' and gay men's experiences of discrimination and
harassment in a university community. American Journal of Community
Psychology, 17, 317-321.

Diaz, R. M., Ayala, G., Bein, E., Henne, J., & Marin, V. (2001). The impact of
homophobia, poverty, and racism on the mental health of gay and bisexual latino
men: Findings from 3 US cities. American Journal of Public Health, 91, 927-932.

DiPlacido, J. (1998). Minority stress among lesbians, gay men, and bisexuals: A
consequence of heterosexism and homophobia. In G. M. Herek (Ed.), Stigma and
sexual orientation: Understanding prejudice against lesbians, gay men, and
bisexuals (pp. 138-159). Thousand Oaks, CA: Sage.

Eisler, R. M., & Blalock, J. A. (1991). Masculine gender role stress: Implications for the
assessment of men. Clinical Psychology Review, 11, 45-60.

Eisler, R. M., Skidmore, J. R. (1987). Masculine gender role stress: Scale development
and component factors in the appraisal of stressful situations. Behavioral
Modification, 11, 123-136.

Ekehammar, B., & Akrami, N. (2003). The relation between personality and prejudice: A
variable- and a person-centered approach. European Journal ofPersonality, 17,
435-448.









Franklin, K. (1998). Unassuming motivations: Conextualizing the narratives of antigay
assailants. In G. M. Herek (Ed.), Stigma and sexual orientation: Understanding
prejudice against lesbians, gay men, and bisexuals (pp. 39-61). Thousand Oaks,
CA: Sage.

Franklin, K. (2000). Antigay behaviors among young adults: Prevalence, patterns, and
motivators in a noncriminal population. Journal ofInterpersonal Violence, 15, 339-
362.

Goldberg, L. R. (1999). A broad-bandwidth, public domain, personality inventory
measuring the lower-level facets of several five-factor models. In I. Mervielde, I.
Deary, F. De Fruyt, & F. Ostendorf (Eds.), Personality Psychology in Europe, Vol.
7 (pp. 7-28). Tilburg, The Netherlands: Tilburg University Press.

Haddock, G., Zanna, M. P., & Esses, V. (1993). Assessing the structure of prejudicial
attitudes: The case of attitudes toward homosexuals. Journal ofPersonality and
Social Psychology, 65, 1105-1118.

Haddock, G., & Zanna, M. P. (1998). Authoritarianism, values, and the favorability and
structure of antigay attitudes. In G. M. Herek (Ed.), Stigma and sexual orientation:
Understanding prejudice against lesbians, gay men, and bisexuals (pp. 82-107).
Thousand Oaks, CA: Sage.

Heaven, P. C. L., & Bucci, S. (2001). Right-wing authoritarianism, social dominance
orientation and personality: An analysis using the IPIP measure. European Journal
ofPersonality, 15, 49-56.

Heaven, P. C. L., & St. Quintin, D. (2002). Personality factors predict racial prejudice.
Personality and Individual Differences, 34, 625-634.

Herek, G. M. (1984). Beyond "homophobia": A social psychological perspective on
attitudes towards lesbians and gay men. Journal ofHomosexuality, 10, 1-21.

Herek, G. M. (1986). On heterosexual masculinity: Some psychical consequences of the
social construction of gender and sexuality. American Behavioral Scientists, 29,
563-577.

Herek, G. M. (1988). Heterosexuals' attitudes toward lesbians and gay men: Correlates
and gender differences. The Journal of Sex Research, 25, 451-477.

Herek, G. M. (2000). The psychology of sexual prejudice. Current Directions in
Psychological Science, 9(1), 19-22.

Herek, G. M., & Capitanio, J. P. (1995). Black heterosexuals' attitudes toward lesbian
and gay men in the United States. The Journal of Sex Research, 32, 95-101.

Hill, P. C., & Wood, R. W., Jr., (Eds.). (1999). Measures ofReligiosity. Birmingham,
AL: Religious Education Press.









Hudson, W. W., & Ricketts, W. A. (1980). A strategy for the measurement of
homophobia. Journal ofHomosexuality, 5, 357-372.

Hunsberger, B. (1995). Religion and prejudice: The role of religious fundamentalism,
quest, and right-wing authoritarianism. Journal of Social Issues, 51, 113-123.

International Personality Item Pool (2001). A scientific collaboration for the development
of advanced measures ofpersonality traits and other individual differences.
Retrieved October 1, 2005 from http://ipip.ori.org/

John, O. P., Donahue, E., & Kentle, R. L. (1992). The Big Five Inventory: Versions 4a
and 54. Technical Report, Institute of Personality and Social and Research,
University of California, Berkley, CA.

Johnson, M. E., Brems, C., & Alfod-Keating, P., (1997). Personality correlates of
homophobia. Journal ofHomosexuality, 34, 57-68.

Kerns, J. G., & Fine, M. A. (1994). The relation between gender and negative attitudes
toward gay men and lesbians: Do gender role attitudes mediate this relation? Sex
Roles, 31, 297-307.

Kite, M. E. (1984). Sex differences in attitudes toward LG persons: A meta-analytic
review. Journal ofHomosexuality, 10, 69-81.

Kite, M. E. (1992). Individual differences in males' reactions to gay males and lesbians.
Journal of Applied Social Psychology, 22, 1222-1239.

Kite, M. E., & Deaux, K. (1986). Attitudes toward homosexuality: Assessment and
behavioral consequences. Basic and Applied Social Psychology, 7(2), 137-162.

Kite, M. E., & Deaux, K. (1987). Gender belief systems: Homosexuality and the implicit
inversion theory. Psychology of Women Quarterly, 11, 83-96.

Kite, M. E., & Whitley, B. E., Jr. (1998). Do heterosexual women and men differ in their
attitudes towards homosexuality? A conceptual and methodological analysis. In G.
M. Herek (Ed.), Stigma and sexual orientation: Understanding prejudice against
lesbians, gay men, and bisexuals (pp. 39-61). Thousand Oaks, CA: Sage.

Kohn, P.M. (1972). The Authoritarianism-Rebellion scale: A balanced F Scale with left-
wing reversals. Sociometry, 35, 176-189.

Kurpius, S. E. R., & Lucart, A. L., (2000). Military and civilian undergraduates:
Attitudes toward women, masculinity, and authoritarianism. Sex Roles, 43, 255-
265.

Laner, M. R., & Laner, R. H. (1979). Personal style or sexual preference: Why gay men
are disliked. International Review of Modern Sociology, 9, 215-228.









Laner, M. R., & Laner, R. H. (1980). Sexual preference or personal style? Why lesbians
are disliked. Journal ofHomosexuality, 5, 339-356.

LaMar, L., and Kite, M. E. (1998). Sex differences in attitudes toward lesbians and gay
men: A multidimensional perspective. Journal of Sex Research, 35, 189-196.

Lefkowitz, E. S., Shearer, C. L., Gillen, M. M, & Espinosa-Hernandez, M. G. (2003).
Measuring female role norms in male and female emerging adults. Unpublished
manuscript.

MacDonald, A. P., Jr. (1974). Identification and measurement of multidimensional
attitudes toward equality between the sexes. Journal ofHomosexuality, 1, 165-182.

Mays, V. M., & Chochran, S. D. (2001). Mental health correlates of perceived
discrimination among lesbian, gay and bisexual adults in the United States.
American Journal of Public Health, 91, 1869-1876.

McCrae, R. R. (1996). Social consequences of experimental openness. Psychological
Bulletin, 120, 323-337.

McCrae, R. R., & Costa, P. T. (2003). Personality in adulthood: A five-factor theory
perspective. New York: Guilford Press.

McHugh, M. C. & Frieze, I. H. (1997). The measurement of gender-role attitudes: A
review and commentary. Psychology of Women Quarterly, 21, 1-16.

O'Neil, J. M. (1981). Patterns of gender role conflict and strain: Sexism and fear of
femininity in men's lives. The Personnel and Guidance Journal, 60, 203-210.

O'Neil, J. M., Helms, B., Gable, R., David, L., & Wrightsman, L. (1986). Gender Role
Conflict Scale: College men's fear of femininity. Sex Roles, 14, 335-350.

Patel, S., Long, T. E., McCammon, S. L., & Wuensch, K. L. (1995). Personality and
emotional correlates of self-reported antigay behaviors. Journal ofInterpersonal
Violence, 10, 354-366.

Paulhus, D. L. (1994). Balanced Inventory of Desirable Responding: Reference manual
for BIDR Version 6. Unpublished manuscript, Department of Psychology,
University of British Columbia, Vancouver, BC.

Peterson, B. E., Smirles, K. A., & Wentworth, P. A. (1997). Generativity and
authoritarianism: Implications for personality, political involvement, and
parenting. Journal ofPersonality and Social Psychology, 72, 1202-1216.

Pleck, J. H., Sonenstein, F. L., & Ku, L. C. (1993). Masculinity ideology: Its impact on
adolescent males' heterosexual relationships. Journal of Social Issues, 49(3), 11-
29.









Pratto, F., Sidanius, J., Stallworth, L. M., & Malle, B. F., (1994). Social dominance
orientation: A personality variable predicting social and political attitudes. Journal
ofPersonality and Social Psychology, 67, 741-763.

Rokeach, M. (1956). Political and religious dogmatism: An alternative to the
authoritarian personality. Psychological Monographs, 70, 1-43.

Rokeach, M. (1960). The open and closed mind. New York: Basic Books.

Roderick, T., McCammon, S. L., Long, T. E., Allred, L. J. (1998). Behavioral aspects of
homonegativity. Journal ofHomosexuality, 36, 79-87.

Schope, R. D. & Eliason, M. J. (2000). Thinking versus acting: Assessing the relationship
between heterosexual attitudes and behaviors towards homosexuals. Journal of Gay
& Lesbian Social Services, 11(4), 69-92.

Sidanius, J. & Pratto, F. (1999). Social dominance: An intergroup theory of social
hierarch and oppression. Cambridge: Cambridge University Press.

Sinn, J. S., (1997). The predictive and discriminate validity of masculinity ideology.
Journal ofResearch in Personality, 31, 117-135.

Smith, E.R., Ferree, M. M., & Miller, F. D. (1975). A short scale of attitudes toward
feminism. Representative Research in Social Psychology, 6, 51-58

Spence, J. T., & Helmreich, R. L. (1978). Masculinity andfemininity: Their
psychological dimensions, correlates, and antecedents. Austin: University of Texas
Press.

Spence, J. T., Helmreich, R. L., & Stapp, J. (1973). Ratings of self and peers on sex-role
attributes and their relations to self-esteem and conceptions of masculinity and
femininity. Journal ofPersonality and Social Psychology, 32, 29-39.

Swim, J. K., Ferguson, M. J., & Hyers, L. L. (1999). Avoiding stigma by association:
Subtle prejudice against lesbians in the form of social distancing. Basic and
Applied Social Psychology, 21(1), 61-68.

Theodore, P. S., & Basow, S. A. (2000). Heterosexual masculinity and homophobia: A
reaction to the self? Journal ofHomosexuality, 40, 31-48.

Thompson, E. H., & Pleck, J. H. (1986). The structure of the male role norms. American
Behavioral Scientist, 29, 531-534.

Thompson, E. H., Pleck, J. H., & Ferrera, D. L. (1992). Men and masculinities: Scales
for masculinity ideology and masculinity-related constructs. Sex Roles, 27, 573-
608.









Tokar, D. M., Fischer, A. R., Schaub, M., Moradi, B., (2000). Masculine gender roles and
counseling-related variables: Links with and mediation by personality. Journal of
Counseling Psychology, 47, 380-393.

Whitley, B. E., Jr. (1987) The relationship of sex-role orientation to heterosexuals'
attitudes towards LG persons. Sex Roles, 17, 103-113.

Whitley, B. E., Jr. (1999) Right-wing authoritarianism, social dominance orientation, and
prejudice. Journal ofPersonality and Social Psychology, 77, 126-134.

Whitley, B. E., Jr. (2001). Gender-role variables and attitudes toward homosexuality. Sex
Roles, 45, 691-721.

Whitley, B. E., Jr., & Egisd6ttir, S. (2000). The gender belief system, authoritarianism,
social dominance orientation, and heterosexuals' attitudes toward lesbian and gay
men. Sex Roles, 42(11/12), 947-967.

Whitley, B. E., & Lee, S. E., (2000). The relationship of authoritarianism and related
constructs to attitudes towards homosexuality. Journal ofApplied Social
Psychology, 30, 144-170.

Wilson, G. D., & Patterson, J. R. (1968). A new measure of conservatism. British Journal
of Social and Clinical Psychology, 7, 264-269.















BIOGRAPHICAL SKETCH

I was born and raised in Silver Spring, Maryland. I graduated magna cum laude

with a Bachelor of Science in psychology from the University of Maryland in 2002. After

graduating I spent a year working with autistic children and traveling overseas. In August

2003 I moved to Gainesville to enter into University of Florida's Counseling Psychology

program. I am currently in my third year of the doctoral program and hope to earn my

Ph.D. in a few years.