Running head: IMPLICIT AND EXPLICIT SEXUAL ATTITUDES

Implicit and Explicit Sexual Attitudes Across Genders and Sexual Orientations Pasquale Anselmi, Alberto Voci, Michelangelo Vianello, Egidio Robusto University of Padua Keywords: Homosexuality, bisexuality, sexual attitude, gender differences, implicit measure. Abstract word count: 114 Main text word count: 2460 References word count: 749 Acknowledgements word count: 16 Table word count: 79 Number of tables: 1 Number of figures: 1

IMPLICIT AND EXPLICIT SEXUAL ATTITUDES - 2 Abstract The article investigates implicit and explicit sexual attitudes held by individuals of different gender and sexual orientation. We found that heterosexual men and women preferred heterosexuals to homosexuals both implicitly and explicitly. Lesbians showed a preference for homosexuals both implicitly and explicitly, whereas gay men showed such a preference only explicitly. Bisexual men preferred heterosexuals to homosexuals only implicitly, whereas bisexual women preferred homosexuals to heterosexuals only explicitly. On the whole, bias in favor of heterosexuals was stronger in the implicit measure than in the explicit measure, and in males than in females. We interpreted these findings in the light of the interplay between selfrelated motives and shared perceptions at a societal level.

IMPLICIT AND EXPLICIT SEXUAL ATTITUDES - 3 Implicit and Explicit Sexual Attitudes Across Genders and Sexual Orientations Sexual prejudice may be seen as the internalization and acceptance of sexual stigma, which is “the negative regard, inferior status, and relative powerlessness that society collectively accords to any nonheterosexual behavior, identity, relationship, or community” (Herek, 2009b, p. 441). This aversive phenomenon may stem from two seeds. First, people tend to perceive their own groups as more positive than outgroups, due to different self-related motives: the need to belong, the quest for a positive self-esteem, the need for certainty and protection (Hewstone, Rubin, & Willis, 2002); when heterosexuals have to judge nonheterosexual individuals, sexual prejudice represents a specific type of ingroup favoritism. Second, at a societal level, shared norms regarding sexuality are generally characterized by the perception of heteronormativity, an internalized view according to which heterosexuality is the only legitimate and expected standard for sexual orientation (Ingraham, 1999; Montgomery & Stewart, 2012). Behavioral consequences of sexual prejudice toward lesbians, gay men and bisexuals are employment or housing discrimination, verbal harassment, and hate crime against their person or property. Most of nonheterosexual individuals have experienced at least once in their life one of these unfair situations (Herek, 2009a). The study of sexual prejudice has therefore important ethical consequences because it informs policies to reduce stigmatizing and discriminatory behaviors. Typically, sexual prejudice has been operationalized and measured through self-report scales. Yet, there are a number of evidences suggesting that self-report measures are heavily influenced by social desirability (see, e.g., Chivers, Seto, Lalumiere, Laan, & Grimbos, 2010). In order to overcome socially desirable answers, “modern” scales of sexual prejudice have been constructed to measure more subtle forms of prejudice (see, e.g., Morrison & Morrison, 2002), but these measures tend to be highly correlated with old-fashioned scales, perhaps because they both refer to introspectively accessible beliefs, affective responses and past behaviors. Alternative measures of sexual prejudice have been developed to overcome the limits of introspection: the so-called implicit measures of attitudes (Wittenbrink & Schwarz, 2007). Implicit measures such as the Implicit Association Test (IAT, Greenwald, McGhee, & Schwartz, 1998) both detect what the individual might explicitly distort and what might not be readily accessible through introspection. Implicit measures of sexual attitude exhibit a weak-to-moderate correlation with explicit measures (Banse, Seise, & Zerbes, 2001; Nosek et al., 2007; Steffens & Buchner, 2003) and – importantly – they have been found to predict gay-related behaviors that are not predicted by explicit measures, such as involvement and immersion in gay culture (Jellison, McConnell, & Gabriel, 2004). Effects of gender and sexual orientation in the expression of sexual attitudes have been described both at an explicit (e.g., Herek, 2009b; see, for a meta-analysis, Petersen & Hyde, 2010) and at an implicit level (e.g., Banse et al., 2001; Steffens, 2005; see, for a meta-analysis, Cullen & Barnes-Holmes, 2008).

IMPLICIT AND EXPLICIT SEXUAL ATTITUDES - 4 In general, heterosexuals show a preference for heterosexuality compared to homosexuality (Anselmi, Vianello, Voci, & Robusto, 2013; Nosek et al., 2007). This could depend on the fact that the two motivations underlying prejudice point at the same direction: The self-image is protected and enhanced by a positive image of heterosexuality compared to homosexuality, and shared heteronormativity is consistent with this perception. Concerning gender, self-presentation strategies are more often used by heterosexual men, who feel a stronger peer pressure to demonstrate masculinity. As a consequence, heterosexual men are less likely to conceal their negative attitude toward homosexuality because they might use the explicit endorsement of antigay attitudes as a strategy to make their masculinity readily evident to others and to prove that they are “real men” (Glick, Gangl, Gibb, Klumpner, & Weinberg, 2007; Herek, 1986). Consistently, heterosexual men were found to show more negative attitudes toward homosexuals than heterosexual women, both at an explicit (e.g., LaMar & Kite, 1998) and at an implicit level (e.g., Banse et al., 2001). For lesbians and gay men, however, the two motivations underlying social judgments may conflict. On the one hand, the homosexual ingroup could be preferred, as a means to protect the self. On the other hand, the adherence to heteronormativity could lead to biases in favor of heterosexuality. In this case, phenomena of self-stigma are possible, as individuals may internalize a negative view of their own minority group, and thus of themselves (Herek, 2007). This phenomenon is particularly likely to occur among gay men, who may be particularly threatened by the social norms concerning masculinity. Consistently, Herek, Gillis, and Cogan (2009) showed that self-stigma was more pronounced for gay men than for lesbians. Moreover, Banse et al. (2001) found that lesbians showed a preference for homosexuals both implicitly and explicitly, whereas gay men showed such a preference only explicitly. Whereas sexual attitudes of heterosexuals, lesbians and gay men have been investigated extensively (see, e.g., Jellison et al., 2004; Jost, Banaji, & Nosek, 2004; Nosek et al., 2007), little attention has been devoted to attitudes held by bisexuals. Combining data from three studies, Steffens (2005) found that the 33 self-identified bisexuals did not show an implicit preference for heterosexuals or homosexuals. The author did not report disaggregated data for male and female bisexuals. When trying to predict levels of sexual prejudice in bisexual individuals, it has to be considered that the self-definition is neither based on heterosexuality nor on homosexuality, although sexual behaviors are consistent with both orientations. Thus, the perception shared in the society, related on heteronormativity, may not be counterbalanced by ingroup preference phenomena, as the very definition of ingroup and outgroup may be blurred. Accordingly, previous research showed that bisexual individuals, compared to homosexuals, had higher levels of identity confusion and lower levels of community connection (Balsam & Mohr, 2007). Also for bisexuality, the shared social norms concerning masculinity may be particularly aversive for men. Consistently with this prediction, Herek et al. (2009) showed that self-stigma was more present among bisexual individuals than among homosexuals, with bisexual men showing the highest scores.

IMPLICIT AND EXPLICIT SEXUAL ATTITUDES - 5 The article investigates implicit and explicit sexual attitudes held by individuals of different gender and sexual orientation. With respect to heterosexuals, lesbians and gay men, we expect to replicate previous results described in the literature. With respect to bisexuals, we expect to observe biases that are smaller than those observed in heterosexuals, lesbians and gay men, given the absence of ingroup preference phenomena. Such biases are expected to be in favor of heterosexuals, given the shared perception of heteronormativity, and to be stronger in bisexual men, given the influence of self-stigma phenomena on them. In addition, we expect biases in favor of heterosexuals to be stronger in the implicit measure than in the explicit measure, given the influence of self-presentation motives. Method Respondents Respondents were individuals who completed the Sexuality IAT available at https://implicit.harvard.edu/implicit/italy. Among the respondents who provided complete and interpretable data (Nosek, Banaji, & Greenwald, 2002), 2280 reported being heterosexual (for age, M = 26.82  9.08 years; 60.13% were females), 204 being gay (M = 26.59  7.82 years; 35.78% were females), and 210 being bisexual (M = 26.06  8.95 years; 63.81% were females). Materials The Sexuality IAT used the category labels Heterosexuals, Homosexuals, Good and Bad. Four images and four words were used to represent the categories Heterosexuals and Homosexuals. Sixteen words were used to represent the categories Good and Bad. The procedure was the same as in Greenwald, Nosek, and Banaji (2003). The order of the critical blocks was counterbalanced across the respondents. Respondents were asked to rate their preference for heterosexual or homosexual people on a 5-point scale from “I strongly prefer homosexual to heterosexual people” ( 2) to “I strongly prefer heterosexual to homosexual people” (2). Data analysis Two factorial designs were run, in which gender and sexual orientation were the factors. In the first the IAT score D (Greenwald et al., 2003) was the dependent variable. A positive D indicates an implicit preference for heterosexual relative to homosexual people. In the second the dependent variable was the score to the explicit preference scale. One-sample t tests were used to assess the significance of the implicit and explicit biases. Moreover, a mixed factorial design was run, in which gender and sexual orientation varied between subjects and the two measures (implicit and explicit) varied within subjects. To allow comparisons between the two measures, prior to this third analysis the scores were divided by their respective standard deviations. Finally, Pearson’s correlations between the implicit and explicit measures were computed separately for respondents of different gender and sexual orientation, and Fisher’s z tests were used to assess the statistical significance of their differences. Results

IMPLICIT AND EXPLICIT SEXUAL ATTITUDES - 6 Table 1 shows the implicit and explicit measures computed on respondents of different gender and sexual orientation. In the implicit measurement, bias towards heterosexuals was stronger in males than in females (Mmale = .41  .44, Mfemale = .28  .46; F(1, 2688) = 50.36, p < .001), and decreased when moving from heterosexual to gay respondents (Mhetero = .39  .43, Mbisex = .13  .49, Mgay =  .10  .41; F(2, 2688) = 163.59, p < .001, all post-hoc tests were significant at p < .001). The interaction between gender and sexual orientation was not significant (F(2, 2688) = 2.29, p = .10). Heterosexual and lesbian females held a preference for the ingroup, whose strength, compared by taking into account the scores D in absolute terms, did not differ (z|hetero|-|gay| = 1.77, p = .08). Also heterosexual males preferred the ingroup, whereas gay males did not. At an implicit level, bisexual males preferred heterosexuals to homosexuals, whereas bisexual females did not prefer either of them. ________________________ Insert Table 1 about here ________________________ In the explicit measurement too, the bias towards heterosexuals was stronger in males than in females (Mmale = .54  .89, Mfemale = .19  .72; F(1, 2688) = 41.21, p < .001), and decreased when moving from heterosexual to gay respondents (Mhetero = .45  .78, Mbisex =  .13  .61, Mgay =  .43  .80; F(2, 2688) = 184.22, p < .001, all post-hoc tests were significant at p < .001). The interaction between gender and sexual orientation was not significant (F(2, 2688) = 2.60, p = .08). In general, heterosexuals and gays of both genders held a preference for the ingroup. However, it is worth noting that, whereas the ingroup preference observed in heterosexual males was stronger than that observed in gay males (z|hetero|-|gay| = 4.53, p < .001), the ingroup preference observed in heterosexual females was weaker than that observed in lesbians (z|hetero|-|gay| =  2.86, p < .01). At an explicit level, bisexual females preferred homosexuals to heterosexuals, whereas bisexual males did not prefer either of them. Figure 1 depicts the implicit (solid lines) and explicit (dashed lines) measures observed on male (squares) and female (circles) respondents of different sexual orientation (the measures were divided by their standard deviations). There was a significant interaction between gender, sexual orientation, and the type of measure (F(2, 2688) = 5.45, p < .01). On the whole, the implicit measure yielded a stronger bias towards heterosexuals than the explicit measure (Mimplicit = .74, Mexplicit = .41; F(1, 2688) = 77.58, p < .001). This held for all groups, except for lesbians (Cohen’s d = .23) and, partly, for heterosexual men: Their difference between implicit and explicit measures was reliable, due to sample size, although Cohen’s d indicates a moderate effect (.34; see Figure 1). In general, males showed a stronger bias towards heterosexuals than females both implicitly and explicitly, with one exception: Gay males and lesbians did not show any difference on the explicit measure.

IMPLICIT AND EXPLICIT SEXUAL ATTITUDES - 7 ________________________ Insert Figure 1 about here ________________________ Weak to moderate positive correlations between the implicit and the explicit measures were observed across gender and sexual orientation of respondents (for males, rhetero = .30; rgay = .15; rbisex = .33; for females, rhetero = .25; rgay = .30; rbisex = .20). The differences between the correlations were not statistically significant (p  .09). Discussion The article investigated implicit and explicit sexual attitudes held by individuals of different gender and sexual orientation. We found that heterosexual men and women preferred the ingroup at both the implicit and the explicit level, and that the former held a stronger preference than the latter. Lesbians showed an ingroup preference both implicitly and explicitly, whereas gay men showed such a preference only explicitly. The aforementioned results are consistent with the literature (Banse et al., 2001; LaMar & Kite, 1998). In addition, we found that bisexual men preferred heterosexuals to homosexuals implicitly, whereas bisexual women preferred homosexuals to heterosexuals explicitly. On the whole, the implicit measure yielded stronger biases in favor of heterosexuals that the explicit measure. Some comments concerning our results are in order. At the explicit level, men expressed a bias in favor of heterosexuals more strongly than women did (notice that the difference between the implicit and the explicit measure is smaller in heterosexual men than in women). It is possible that heterosexual men perceived homosexuals as more threatening than heterosexual women did. Gay men showed no preference at the implicit level, whereas the ingroup preference that they held explicitly was smaller than that held by heterosexual men. The ingroup preference held by lesbians at the implicit level was not different from that held by heterosexual women, whereas their ingroup preference held at the explicit level was even stronger. It is likely that the influence of heteronormativity was stronger in gay men than in lesbians, possibly due to the additional influence of self-stigma phenomena, that have been shown to be particularly unfavorable for gay men (Herek et al., 2009). On the whole, the biases held by bisexuals were smaller than those held by heterosexuals and gays. This is consistent with the fact that their preference for heterosexuals or homosexuals cannot be strengthened by ingroup preference phenomena, as the identification of heterosexuals or homosexuals as ingroup or outgroup is less clear. Heteronormativity and self-stigma phenomena might have had a combined influence on the preference for heterosexuals to homosexuals showed by bisexual men at the implicit level. Regardless of gender and sexual orientation, the implicit measure yielded more bias in favor of heterosexuals than the explicit measure. Heteronormativity might have had a stronger effect at the implicit level than at the explicit level. This may have caused the preference for homosexuals, observed in gay men and bisexual women at the explicit level, to disappear at the

IMPLICIT AND EXPLICIT SEXUAL ATTITUDES - 8 implicit level, and the preference for heterosexuals to be observed in bisexual men implicitly, but not explicitly. A limit of the present study is that, because of the method used for data collection, respondents to the Sexuality IAT were self-selected. Strengths of the study are the investigation of sexual attitudes held by bisexuals, and the number of nonheterosexual individuals, which is larger than it has been in most previous research. Acknowledgements The authors thank the principal investigators of Project Implicit for their permission to report data.

IMPLICIT AND EXPLICIT SEXUAL ATTITUDES - 9 References Anselmi, P., Vianello, M., Voci, A., & Robusto, E. (2013). Implicit sexual attitude of heterosexual, gay and bisexual individuals: Disentangling the contribution of specific associations to the overall measure. Manuscript submitted for publication. Balsam, K. F., & Mohr, J. J. (2007). Adaptation to sexual orientation stigma: A comparison of bisexual and lesbian/gay adults. Journal of Counseling Psychology, 54(3), 306−319. doi: 10.1037/0022-0167.54.3.306 Banse, R., Seise, J., & Zerbes, N. (2001). Implicit attitudes towards homosexuality: Reliability, validity, and controllability of the IAT. Zeitschrift für Experimentelle Psychologie, 48(2), 145−60. doi: 10.1026//0949-3946.48.2.145 Chivers, M. L., Seto, M. C., Lalumiere, M. L., Laan, E., & Grimbos, T. (2010). Agreement of self-reported and genital measures of sexual arousal in men and women: A meta-analysis. Archives of sexual behavior, 39(1), 5−56. doi: 10.1007/s10508-009-9556-9 Cullen, C., & Barnes-Holmes, D. (2008). Implicit pride and prejudice: A heterosexual phenomenon? In M. A. Morrison &T. G. Morrison (Eds.), The psychology of modern prejudice (pp. 195–223). New York, NY: Nova Science. Glick, P., Gangl, C., Gibb, S., Klumpner, S., & Weinberg, E. (2007). Defensive reactions to masculinity threat: More negative affect toward effeminate (but not masculine) gay men. Sex Roles, 57, 55–59. Greenwald, A. G., McGhee, D. E., & Schwartz, J. L. K. (1998). Measuring individual differences in implicit cognition: The Implicit Association Test. Journal of Personality and Social Psychology, 74(6), 1464-1480. doi: 10.1037/0022-3514.74.6.1464 Greenwald, A. G., Nosek, B. A., & Banaji, M. R. (2003). Understanding and using the Implicit Association Test: I. An improved scoring algorithm. Journal of Personality and Social Psychology, 85(2), 197−216. doi: 10.1037/0022-3514.85.2.197 Herek, G. M. (1986). On heterosexual masculinity: Some psychical consequences of the social construction of gender and sexuality. American Behavioral Scientist, 29(5), 563−577. doi: 10.1177/000276486029005005 Herek, G. M. (2007). Confronting sexual stigma and prejudice: Theory and practice. Journal of Social Issues, 63, 905−925. doi: 10.1111/j.1540-4560.2007.00544.x Herek, G. M. (2009a). Hate crimes and stigma-related experiences among sexual minority adults in the United States: prevalence estimates from a national probability sample. Journal of Interpersonal Violence, 24, 54–74. doi: 10.1177/0886260508316477 Herek, G. M. (2009b). Sexual Prejudice. In T. Nelson (Ed.), Handbook of prejudice, stereotyping, and discrimination (pp. 441–468). New York, NY: Taylor & Francis. Herek, G. M., Gillis, J. R., & Cogan, J. C. (2009). Internalized stigma among sexual minority adults: Insights from a social psychological perspective. Journal of Counseling Psychology, 56(1), 32−43. doi: 10.1037/a0014672 Hewstone, M., Rubin, M., & Willis, H. (2002). Intergroup bias. Annual Review of Psychology, 53, 575−604. doi: 10.1146/annurev.psych.53.100901.135109

IMPLICIT AND EXPLICIT SEXUAL ATTITUDES - 10 Ingraham, C. (1999). White weddings: Romancing heterosexuality in popular culture. New York, NY: Routledge, Chapman & Hall. Jellison, W. A., McConnell, A. R., & Gabriel, S. (2004). Implicit and explicit measures of sexual orientation attitudes: Ingroup preferences and related behaviors and beliefs among gay and straight men. Personality and Social Psychology Bulletin, 30, 629–642. doi: 10.1177/0146167203262076 Jost, J. T., Banaji, M. R., & Nosek, B. A. (2004). A decade of system justification theory: Accumulated evidence of conscious and unconscious bolstering of the status quo. Political Psychology, 25(6), 881−919. doi: 10.1111/j.1467-9221.2004.00402.x LaMar, L., & Kite, M. (1998). Sex differences in attitudes toward gay men and lesbians: A multi-dimensional perspective. The Journal of Sex Research, 35(2), 189−196. doi: 10.1080/00224499809551932 Montgomery, S. A., & Stewart, A. J. (2012). Privileged allies in lesbian and gay rights activism: Gender, generation, and resistance to heteronormativity. Journal of Social Issues, 68(1), 162−177. doi: 10.1111/j.1540-4560.2012.01742.x Morrison, M.A, & Morrison, T. G. (2002). Development and validation of a scale measuring modern prejudice toward gay men and lesbian women. Journal of Homosexuality, 43(2), 15–37. doi: 10.1300/J082v43n02_02 Nosek, B. A., Banaji, M. R., & Greenwald, A. G. (2002). Harvesting implicit group attitudes and beliefs from a demonstration website. Group Dynamics: Theory, Research, and Practice, 6, 101115. doi: 10.1037/1089-2699.6.1.101 Nosek, B. A., Smyth, F. L., Hansen, J. J., Devos, T., Lindner, N. M., Ranganath, K. A., … Banaji, M. R. (2007). Pervasiveness and correlates of implicit attitudes and stereotypes. European Review of Social Psychology, 18: 3688. doi: 10.1080/10463280701489053 Petersen, J. L., & Hyde, J. S. (2010). A meta-analytic review of research on gender differences in sexuality: 1993 to 2007. Psychological Bulletin, 136, 21–38. doi: 10.1037/a0017504 Steffens, M. C. (2005). Implicit and explicit attitudes towards lesbians and gay men. Journal of Homosexuality, 49, 39–65. doi: 10.1300/J082v49n02_03 Steffens, M. C., & Buchner, A. (2003). Implicit Association Test: Separating transsituationally stable and variable components of attitudes toward gay men. Experimental Psychology, 50, 33–48. doi:10.1026//1618-3169.50.1.33 Wittenbrink, B., & Schwarz, N. (2007). Implicit measures of attitudes. New York, NY: Guilford Press.

IMPLICIT AND EXPLICIT SEXUAL ATTITUDES - 11 Table 1 Implicit and Explicit Measures Computed on Respondents of Different Gender and Sexual Orientation Implicit measure Explicit measure Respondent M SD M SD Heterosexuals Male .48*** .40 .71*** .82 Female .34*** .43 .28*** .70 Bisexuals Male .30*** .46 .11 .72 Female .04 .48  .27*** .49 Gays Male  .03 .38  .37*** .80 Female  .25*** .42  .55*** .79 Note. Differences from 0. *** p < .001.

IMPLICIT AND EXPLICIT SEXUAL ATTITUDES - 12 Figure caption Figure 1. Implicit and explicit measures observed on respondents of different gender and sexual orientation, divided by their standard deviations. Bars represent 95% confidence intervals. All comparisons reported in the right side are significant (p <.001), except K-L (p = .32) and F-L (p = .13).

IMPLICIT AND EXPLICIT SEXUAL ATTITUDES - 13