Toward a Greater Understanding of Antigay Prejudice: On the Role of Sexual Orientation and Gender Role Violation

Abstract

Prejudice against gay men and lesbians could be driven by at least two types of expectancy violations: those pertaining to sexual orientation per se, and perceived violations of traditional gender roles (e.g., the fact that gay men are often inferred to be feminine and lesbians to be masculine). However, it is unclear whether one or the other (or both) of these expectancy violations are actually important in driving prejudicial reactions. In a completely crossed design, participants were asked to evaluate a target who varied with respect to biological sex, sexual orientation, and gender role (i.e., whether they were masculine or feminine). In addition, we also examined the contingency of these variables on preexisting individual differences in prejudice toward gay men and lesbians as a whole. Results showed a moderate trend among high prejudice participants to disparage “double violators,” that is, individuals who simultaneously violate expectations about both sexuality and gender roles. Implications for current research and practice are discussed.

Notes

¹Although we did not explicitly assess participants' own sexual orientation, we presumed that the vast majority of our participants were heterosexual, an assumption which seemed reasonable given the existing work on base-rates of sexual orientation (e.g., Sell, Wells, & Wypij, 1995).

²Half of the items on the ATLG refer to gay men, whereas the other half refer to lesbians. Although attitudes toward gay men and lesbians are distinct in theory, in practice these two attitude subscales were extremely highly correlated, r(211) = .84, p < .0001. Moreover, each of these two subscales was equally predictive of gay male and lesbian targets in this study. Hence, although it would be foolish to conclude that there are not distinct factors driving prejudice toward gay men and lesbians, these data provide justification for forming a single index of antigay prejudice.

³As in other prejudice research (e.g., Devine, 1989) we use the terms “high” and “low” merely as heuristically convenient labels, which simply refer to relative levels of self-reported sentiments towards the target group.

⁴Initial correlational analyses among the three judgmental composites (collapsed over target type) revealed a moderate relation between general likeableness and immorality, r(213)=−.31, p<.001. General likeableness was not reliably related to masculinity/femininity, r(213)=.10, ns. There was a tendency for immorality and masculinity/femininity to be related but the magnitude of this relation was trivial, r(213)=.15, p<.05.

Ratings reflect average of ratings of masculinity and reverse-scored femininity; ratings were made on a scale ranging from 0 to 10, with higher numbers indicating higher perceived masculinity.

⁵Analyses also revealed a significant three-way Group Prejudice X Target Sex X Target Behavior interaction, β = − 2.02, t(207) = − 3.27, p < .001. This interaction was difficult to interpret and, more important for our purposes, did not involve target sexual orientation. Because of our main interest in the consequences of sexual orientation (either on its own or in combination with other factors), its implications were not especially relevant for our purposes and it is not considered further.

Ratings based on composite of overall evaluation of the target, desire to meet him or her, and likeableness ratings, all made on a scale from 0 to 10.

Ratings were made on a scale ranging from 0 to 10, with higher numbers indicating higher perceived immorality.

⁶An even more simple, albeit less statistically precise, way of framing this effect is that the two highest immorality ratings were attributed to the two homosexual targets whose behaviors conformed to stereotypic expectations (i.e., the feminine gay man and masculine lesbian, Ms = 3.70 vs. 5.20, respectively). Strictly speaking, however, it is more informative to compare immorality ratings of these two targets relative to the comparably-described heterosexual target, in order to take into account possible differences in how immoral participants considered the behaviors to be.

⁷Skidmore et al. (2006) only used self-identified lesbians and gay men as participants. Nevertheless, these researchers found (as did we) a consistent tendency for participants to derogate other lesbian/gay targets if they violated traditional gender roles than if they did not. This was particularly true when participants judged masculine lesbians, who always elicited more negative reactions than feminine lesbians, and this was true regardless of whether the participant doing the rating was lesbian or gay. This pattern is interesting not only because it parallels our own findings, but also because it suggests that the derogation of certain types of norm-breaking behavior carries over into the lesbian/gay community, although more work is clearly needed in this area. The other focus of their study—the correlates of self-reported anxiety/depression by lesbians and gay men—is obviously important, but is much less relevant to the scope of our study.