pwqu_346.tex Psychology of Women Quarterly, 31 (2007), 125–136. Blackwell Publishing, Inc. Printed in the USA. Copyright C© 2007 Division 35, American Psychological Association. 0361-6843/07 THE F WORD: IS FEMINISM INCOMPATIBLE WITH BEAUTY AND ROMANCE? Laurie A. Rudman and Kimberly Fairchild Rutgers University Three studies examined the predictive utility of heterosexual relationship concerns vis-à-vis support for feminism. Study 1 showed that beauty is perceived to be at odds with feminism, for both genders. The stereotype that feminists are unattractive was robust, but fully accounted for by romance-related attributions. Moreover, more attractive female participants (using self-ratings) showed decreased feminist orientations, compared with less attractive counterparts. Study 2 compared romantic conflict with the lesbian feminist stereotype and found more support for romantic conflict as a negative predictor of support for feminism and women’s civil rights. Study 3 showed that beliefs about an incompatibility between feminism and sexual harmony negatively predicted support for feminism and women’s civil rights. In concert, the findings indicate that a marriage between research on romantic relationships and the factors underlying sexism is overdue for understanding gender inequities. Although the Women’s Movement has made tremendous progress toward gender parity, cultural and psychological barriers remain that may prevent women from capitalizing on the advances that have been made. One barrier to gen- der equity involves orientations toward feminism. Although college-aged adults generally support feminist causes (e.g., passage of the Equal Rights Amendment), they tend to avoid identifying themselves as feminists (e.g., Buschman & Lenart, 1996; Renzetti, 1987; Williams & Wittig, 1997). Equally unsettling is the observation that feminists are eval- uated negatively (Haddock & Zanna, 1994) and are viewed as competent but cold (Fiske, Cuddy, Glick, & Xu, 2002). To the extent that people treat feminism as the F word, or are unfavorably disposed toward those who work for gen- der parity, equal opportunity for women is likely to remain elusive. Thus, it is important to discover the factors that prevent individuals from embracing feminism. What Love Has to Do With It To date, prominent sexism theorists have focused more on power differences between the sexes than on cultural scripts Laurie A. Rudman and Kimberly Fairchild, Department of Psy- chology, Rutgers University. This research was partially supported by Grant BCS-0417335 from the National Science Foundation. We thank Gloria Cowan, Peter Glick, Diana Sanchez, and Jayne Stake for their helpful comments. Address correspondence and reprint requests to: Laurie A. Rudman, Department of Psychology, Tillett Hall, Rutgers Uni- versity, 53 Avenue E, Piscataway, NJ 08854-8040. E-mail: rudman@rci.rutgers.edu that teach men and women how to attract one another. For example, sex differences in status-seeking behaviors have been attributed to gendered role assignments (Cejka & Eagly, 1999; Eagly, 1987), men’s stronger desire to maintain the status quo (Pratto, Stallworth, Sidanius, & Sears, 1997), and men’s appeasement of women (e.g., through benevo- lent sexism) to undermine resistance to gender hegemony (Jackman, 1994). Our goal is not to detract from these co- gent frameworks, but instead to emphasize a likely sus- pect that has been underinvestigated as a factor in gender inequity—heterosexual romance. In this regard, feminist theorists have argued that women place great value in social relationships, often at some sacrifice to their own needs (Chodorow, 1979; Gilligan, 1982). Through peer interactions, women in college may be predominately educated in romance as opposed to learn- ing independence (Holland & Eisenhart, 1990), although there are individual differences in the extent to which women endorse and identify with romantic cultural scripts (Holland, 1992). Within the field of social psychology, Glick and Fiske (1996) broke conceptual ground by emphasizing that benevolent sexism originates in, and is reinforced by, ro- mantic relationships. Because men depend on women for a variety of services, including sexual gratification, emotional intimacy, and domestic labor, it is in their best interests to protect and cherish women. Similarly, women depend on men for economic stability, social prestige, and roman- tic love; as a result, they show benevolence toward their protectors and providers (Glick & Fiske, 1999). Glick and Fiske (1996) stressed that paternalism casts women as won- derful but weak, thereby reinforcing women’s lower status, 125 126 RUDMAN AND FAIRCHILD relative to men. That is, intimate heterosexual partnerships may play a significant role in perpetuating gender stereo- types and status disparities. For example, women may ide- alize men and even pare their ambitions to fit men’s expec- tations to ensure a happy romantic life. However, for the most part, researchers have tended to ignore the influence of sexuality on sexism and the literatures on gender dis- crimination and romantic relationships have developed in parallel. In an exception, Rudman and Heppen (2003) exam- ined the potential costs to women of romantic socialization (e.g., exposure to fairy tales). Using the Implicit Associa- tion Test (Greenwald et al., 2002), they found that, on av- erage, women in their study automatically associated male romantic partners with chivalry and heroism (e.g., Prince Charming, White Knight, hero), suggesting a cognitive link between romance and protection. Moreover, women who showed this link also reported low interest in personal power, including high-paying occupations, advanced ed- ucation, and volunteering for a leadership role. As a re- sult, the authors suggested that women may suffer from a glass-slipper effect such that their personal ambitions may be hobbled by an implicit belief that power might best be gained indirectly, through intimate relationships with men. Because explicit romantic beliefs did not predict women’s interest in power, the authors suggested that romantic so- cialization may have a nonconscious influence on women’s aspirations. In the present research, we sought to investigate factors that undermine feminism’s appeal and, therefore, power for women as a group. The overarching goal was to continue to investigate the influence of sexuality on sexism, but in this case our focus was on heterosexuals’ potential beliefs that feminism conflicts with romance. Among these are stereo- types that feminists are unattractive, have low sex appeal for men, and are likely to be lesbians. In addition, men and women alike may believe that feminism is troubling for intimate relationships. Beliefs that feminism conflicts with romance are likely to undermine gender parity if individuals shy away from feminism as a result. The present research provides a first step toward examining this hypothesis. Feminist Stereotypes Researchers examining feminist stereotypes have observed that feminism and beauty may be perceived as incompati- ble (Goldberg, Gottesdiener, & Abramson, 1975). Goldberg et al. (1975) photographed 30 female undergraduates who reported their feminist attitudes and instructed a subse- quent sample to rate the women on attractiveness. The au- thors found that these ratings did not predict the women’s feminism. Nonetheless, participants were more likely to judge unattractive than attractive women as feminists, and this was true for women and men alike (d = .70 and 1.13, respectively).1 Citing the halo effect for beauty (Dion, Berscheid, & Walster, 1972), the authors interpreted their findings as reflecting negative attitudes toward feminists. However, because gender and participants’ own feminism did not moderate their results, the authors speculated that people may instead have used a stereotype about femi- nists’ attractiveness—one that their results suggested was inaccurate. Follow-up research has often supported the existence of the stereotype that feminists are unattractive (for a review, see Unger, Hilderbrand, & Madar, 1982).2 For the most part, however, researchers have not sought to account for this stereotype by investigating its underlying attributions. In Study 1, we hypothesized that people may judge plain women to be feminists because they have less sex appeal for men or are more likely to be lesbians, compared with pretty women. To the extent that feminists are stigmatized as unattractive, unappealing to men, or lesbians (Swim, Fer- guson, & Hyers, 1999; Williams & Wittig, 1997), women (and men) may devalue feminists and avoid identifying with them. In addition to examining feminist stereotypes, we more directly tested whether feminism is viewed as problematic for heterosexual intimacy. In Study 2, we expected to find low support for feminism among participants who believed that it conflicted with romance (e.g., beliefs that men are less likely to date feminists and that feminists are at risk for relationship conflict). We also compared these beliefs to the lesbian feminist stereotype as predictors of femi- nist identity and attitudes toward feminists and support for women’s civil rights. In Study 3, we examined whether female assertiveness and autonomy, attributes that are in- strumental for gender equality, are perceived as promot- ing sexual conflict. For example, people who endorsed per- ceptions that men perform better sexually when they are in charge, or that women ought not to compete against men in order to be loved, should also show low enthusiasm for feminism. In sum, our studies investigated a perceived conflict between feminism and heterosexual attraction in- volving beauty, romance, and sex, with the expectation that relationship concerns would negatively covary with femi- nist identity and attitudes and support for women’s civil rights. Although our focus was on women, we included men in each investigation. Because men’s cooperation in women’s quest for gender parity is important for its success, it is equally important to investigate factors that predict men’s enthusiasm for feminism. A priori, we expected men and women to similarly support our focal hypotheses because both genders are invested in harmonious romantic relation- ships. Moreover, they have both shown evidence of femi- nist stereotypes (e.g., Fiske et al., 2002; Goldberg et al., 1975; Swim et al., 1999). However, as in past research, we expected men to show lower feminist identity and pro- feminist attitudes compared to women (e.g., Haddock & Zanna, 1994). Feminism and Romance 127 STUDY 1 In Study 1, we examined predicted linkages between fem- inism, beauty, and romance for heterosexual women and men. In line with Goldberg et al. (1975), we expected fe- male targets to be perceived as feminists to the extent they were judged as unattractive. However, we sought to explain this negative relationship by assessing beliefs about targets’ (a) sex appeal and (b) lesbianism. If the link between per- ceived beauty and feminism is accounted for by beliefs that unattractive targets are likely to be unpopular with men and/or to be lesbians, results would suggest that feminists are stigmatized as “unsexy” and thereby underscore the importance of romance and sexuality vis-à-vis judgments of female targets’ feminism. We also tested feminist iden- tity as a moderator of reactions to female targets. Although Goldberg et al. (1975) found no evidence for feminism as a moderator, contemporary feminists might not stereotype plain female targets as feminists. Method Participants Ninety-nine heterosexual volunteers (61 women, 38 men) participated in exchange for partial credit toward their In- troductory Psychology research participation requirement. Of these, 39 (44%) were European American, 25 (28%) were Asian American, 8 (9%) were African American, 7 (6%) were Latino/a, and the remainder reported another ethnic identity. Of the original sample (N = 111), 4 partic- ipants were excluded because they failed to follow instruc- tions, 2 because of technical difficulties, and 6 because they indicated they were lesbian or bisexual (all female). Study Materials Stimulus photos. Eight black and white pictures of women (four pretty, four plain) were taken from a Web site displaying 1974 high school graduation photos.3 We used our own judgment to determine pretty and plain targets; a manipulation check from our participants confirmed our decisions (see the Results section). All photos were head- shots of women facing forward, looking at the camera. The two sets of pictures were yoked on hair coloring and facial expression (two blondes, two brunettes and two smiling, two nonsmiling per category). All of the women were of normal weight, had medium-length hair, and wore similar clothing (sweaters). Target ratings. Participants rated each of the eight tar- gets on four items, using 7-point scales ranging from 1 (strongly disagree) to 7 (strongly agree). Two of the items were: “In my opinion, she is attractive” and “She was prob- ably popular (dating-wise) in high school.” In addition, fol- lowing Goldberg et al. (1975), we told people that we knew what had happened to these women, and we asked them to respond to two other items: “She likely became a lesbian” and “She probably grew up to become a feminist.” These four ratings were reliable across the four pretty women (α > .72 for all) and the four plain women (α > .65 for all). Therefore, we averaged them to form four indexes for both pretty and plain targets (collapsed across photos). We labeled these the attractive index, the sex appeal index (de- rived from dating popularity), the lesbian index, and the feminist index. It is important to note that these ratings were administered in randomized order for each of the eight targets. Self-ratings. Participants responded to three items us- ing 7-point scales ranging from 1 (strongly disagree) to 7 (strongly agree). The items were: “I consider myself to be a feminist”; “I would feel proud if someone called me a feminist”; and “I would feel offended if someone called me a feminist” (reverse scored). These were averaged to form the feminist identity index (α = .82). To measure feminist attitudes, we administered a thermometer on which peo- ple responded to “How do you feel toward feminists” on a scale ranging from 1 (very cold) to 10 (very warm). Partici- pants rated their own attractiveness using two items (“Peo- ple often tell me I am attractive” and “I consider myself to be attractive”) on scales ranging from 1 (strongly disagree) to 7 (strongly agree). These items were strongly related, r(97) = .85, p < .001, and were averaged to form self- rated attractiveness. Participants also indicated whether they were exclusively heterosexual (used for screening pur- poses in each study) by responding 1 (no), 2 (not sure), or 3 (yes). Procedure Participants were escorted by an experimenter to individual booths equipped with a desktop PC; the experimenter ad- ministered instructions and started the computer program. Measures were administered in the following order: tar- get ratings, attitude scales, feminist identity, self-reported attractiveness, and demographics (gender, race, and het- erosexuality). Within each measure, items were randomly presented. Following completion of these measures, par- ticipants were thanked and debriefed. Results Gender Differences in Self-Ratings Table 1 shows descriptive statistics for Study 1’s variables separately by gender. Women scored higher on the feminist identity index than did men, resulting in a large effect size for this gender difference. In addition, compared with men, women reported more favorable attitudes towards femi- nists, with moderately large effect sizes for these differ- ences. Finally, no significant gender differences emerged in self-rated attractiveness. 128 RUDMAN AND FAIRCHILD Table 1 Descriptive Statistics as a Function of Participant Gender (Study 1) Women Men Sex difference M SD M SD t d Self-ratings Feminist identity 3.91 1.38 2.40 1.31 5.36∗∗ .98 Self-attractiveness 4.99 1.30 4.98 1.22 .02 .00 Feminist attitude 5.87 2.20 4.50 2.23 3.01∗∗ .59 Pretty target ratings Attractive 5.52 .74 5.86 .73 1.59 −.45 Sex appeal 5.87 .72 5.86 .74 .03 .01 Feminist 3.30 1.09 3.11 1.01 .88 .18 Lesbian 2.36 .97 2.66 .95 1.53 −.30 Plain target ratings Attractive 2.14 .82 1.91 .78 1.40 .28 Sex appeal 2.29 .78 2.25 .84 .25 .05 Feminist 4.55 1.11 4.12 1.13 1.86 .38 Lesbian 4.17 1.21 4.03 1.19 .57 .12 Note. Means are based on 61 women and 38 men. The effect size (Cohen’s d) for gender differences is based on the pooled standard de- viation. Positive d scores indicate women scored higher than men. By convention, small, moderate, and large effect sizes correspond to .20, .50, and .80, respectively (Cohen, 1988). ∗∗p < .01. Target Ratings We conducted four separate multivariate analyses of vari- ance (MANOVAs) analyzing participants’ ratings of pretty and plain targets’ attractiveness, sex appeal, lesbianism, and feminism. We used participant gender as a between- participants variable. We used feminist identity as a covari- ate in these analyses to examine possible interaction effects. Results revealed that, for ratings of attractiveness, sex ap- peal, and lesbianism, each MANOVA showed only a main effect for target attractiveness, attractiveness: F(1, 95) = 254.61, p < .001, sex appeal: F(1, 95) = 144.61, p < .001, and lesbianism: F(1, 95) = 37.84, p < .001, with pretty tar- gets rated as more attractive, more sexually appealing, and less likely to be lesbians, compared with plain targets, by both genders (see Table 1). The remaining effects were not significant. For feminism, results again showed a main effect for tar- get, F(1, 95) = 32.29, p < .001, with plain targets rated as more likely to be feminists than pretty targets by both genders. However, there was also a significant Target × Gender × Feminist Identity interaction, F(1, 95) = 5.05, p < .01. Among women, feminist identity was positively and significantly related to judging the feminism of pretty targets, r(59) = .42, p < .01, but not the feminism of plain targets, r(59) = −.11, ns. Among men, feminist iden- tity was not significantly linked to judging the feminism of either pretty targets, r(37) = .17, ns, or plain targets, r(37) = −.07, ns. The remaining effects in this analysis were not significant. In sum, these results revealed strong effects for target attractiveness on measures of perceived attractiveness, sex appeal, lesbianism, and feminism, which were not moder- ated by gender or, for the most part, feminist identity. How- ever, for female participants, feminist identity was positively linked to judging pretty targets as feminists, whereas fem- inist identity did not relate to reactions to plain targets for either gender. Thus, there was weak evidence to support the notion that contemporary feminists might be more resistant to the stereotype that feminists are unattractive. Accounting for the Unattractive Feminist Stereotype Our primary goal was to test our hypothesis that perceived sex appeal and/or lesbianism could account for the negative link between attractiveness and feminism. The top diagram (.66**) -.33* (-.42**) -.11 (-.36*) Sex appeal Attractive Feminist .03 (-.22*) .55** (.58**) -.14 (-.27**) Lesbian Attractive Feminist (.66**) -.37** (-.35**) (-.22*) Sex appeal Attractive Lesbian Fig. 1. Study 1 regression analyses testing mediators of the re- lationship between pretty women’s perceived attractiveness and feminism (top and center diagrams) and pretty women’s attractive- ness and lesbianism (bottom diagram). Coefficients in parentheses are bivariate correlations. A dashed arrow indicates full mediation. ∗p < .05. ∗∗p < .01. Feminism and Romance 129 (.77**) -.37** (-.32**) .07 (-.22*) Sex appeal Attractive Feminist -.11 (-.36**) .70** (.68**) .03 (-.22*) Lesbian Attractive Feminist (.77**) -.33* (-.41**) (-.36**) Sex appeal Attractive Lesbian Fig. 2. Study 1 regression analyses testing mediators of the rela- tionship between plain women’s perceived attractiveness and fem- inism (top and center diagrams) and plain women’s attractiveness and lesbianism (bottom diagram). Coefficients in parentheses are bivariate correlations. A dashed arrow indicates full mediation. ∗p < .05. ∗∗p < .01. in Figures 1 and 2 shows the results of the sex appeal mediational analysis, respectively, for pretty and plain women (Baron & Kenny, 1986). As can be seen, the relation- ship between attractiveness and feminism was negative in each case, but wholly accounted for by the sex appeal index. A Sobel’s (1982) test confirmed sex appeal as a significant mediator for pretty women, z = 2.54, p < .05, and for plain women, z = 2.43, p < .05. In the second diagram in Fig- ures 1 and 2, sex appeal was replaced by the lesbian index. As can be seen, the lesbian index was also a successful mediator for pretty women, z = 2.29, p < .05, and for plain women, z = 3.47, p < .001. Thus, although pretty women were rated as more attractive than plain women, there was sufficient variability in these ratings to support a negative relation- ship between beauty and feminism for each target group and to support perceived sex appeal and homosexuality as mediators of that relationship. As a check on their unique ability to predict ratings of targets’ feminism, we simultaneously regressed the femi- nist index on ratings of targets’ attractiveness, sex appeal, and lesbianism. Results for pretty women supported both lesbianism, β = .50, p < .001, and sex appeal, β = −.25, p < .05, as unique predictors; attractiveness was expectedly nonsignificant, β = .01, ns (overall R2 = .39). Results for plain women supported lesbianism, β = .67, p < .001, but not sex appeal, β = −.14, ns, or attractiveness, β = −.13, ns (overall R2 = .48). Thus, for plain women, targets’ presumed lesbianism was the sole unique predictor of their feminism, underscoring a perceived conflict between feminism and heterosexual relationships. Because lesbianism played a central role in targets’ per- ceived feminism, a final set of analyses sought to account for targets’ lesbianism ratings. The hypothesis was that unattractive female targets might be viewed as lesbians be- cause men do not pursue them as romantic partners. If so, the sex appeal index should account for the negative link between targets’ perceived attractiveness and their sexual orientation. The bottom diagrams in Figures 1 and 2 support this hypothesis for pretty and plain women, respectively. In each case, after accounting for sex appeal, the negative link between attractiveness and lesbianism was reduced to non- significance. A Sobel’s test confirmed that sex appeal was a significant mediator for pretty women, z = 2.74, p < .01 (overall R2 = .17), and for plain women, z = 2.20, p < .05 (overall R2 = .23). Are Attractive Women Unlikely to Be Feminists? Goldberg et al. (1975) found the unattractive feminist stereotype to be inaccurate (i.e., attractive women were just as likely as plain women to endorse feminism), but the Women’s Movement was young in 1975. In the present research, we correlated self-rated attractiveness with femi- nist identity and attitudes, separately for women and men. As shown in Table 2, for women, the more they reported Table 2 Correlations as a Function of Participant Gender (Study 1) Women Men Feminist Feminist Feminist Feminist identity attitude identity attitude Self-ratings Self-attractiveness −.34∗∗ −.41∗∗ −.08 .25 Feminist attitude .58∗∗ .53∗∗ Note. Correlations are based on 61 women and 38 men. ∗∗p < .01. 130 RUDMAN AND FAIRCHILD being attractive, the less they (a) identified with feminism and (b) showed positive attitudes toward feminists. This pat- tern was unexpected, given Goldberg et al.’s (1975) findings, but they suggest that, in the present sample, self-reported attractiveness was somewhat incompatible with feminism. By contrast, and not surprisingly, these relationships were not shown for men. Discussion Study 1 supported our central hypotheses. As expected, there was evidence that the unattractive feminist stereo- type is robust. However, we found that perceptions of fe- male targets’ sex appeal and lesbianism fully mediated the negative link between targets’ perceived attractiveness and feminism. Moreover, the negative link between targets’ per- ceived attractiveness and lesbianism was fully mediated by their perceived sex appeal. In concert, these findings sug- gest the importance of beauty and romance when predict- ing judgments of female targets’ feminism and sexuality, and they shed light on why feminisim has become a stig- matized label. Women who subscribe to the view that fem- inists are unattractive or sexually unappealing may distance themselves to preserve their own romantic appeal. More- over, they may well fear that others will stereotype them as lesbians if they embrace an overt feminist identity (Swim et al., 1999). Moreover, men who perceive feminists to be lesbians might view them as unlikely to fulfill their sexual needs, which could lead to negative attitudes toward femi- nist women (Glick & Fiske, 1996). In sum, Study 1’s central findings suggest that the unattractive feminist stereotype can be fully explained by romance-related attributions, thus extending Goldberg et al.’s (1975) analysis. Further, and in contrast to Goldberg et al. (1975), feminist identity moderated stereotype use, but only to a limited degree. For women, feminist iden- tity predicted judging pretty (but not plain) targets as likely feminists; men’s feminist identity did not predict these rat- ings for either plain or pretty targets. Finally, and again in contrast to Goldberg et al. (1975), female participants who rated themselves as attractive showed low enthusiasm for feminism on the identity and attitude measures. That is, the present sample showed some evidence that the unattractive feminist stereotype contains a kernel of truth. However, we caution against overinterpreting this unexpected find- ing. It is possible that feminists resist the beauty myth (Wolff, 1991; cf. Rubin, Nemeroff, & Russo, 2004) by not spending a great deal of time on their appearance or that they surround themselves with people unlikely to com- ment on their attractiveness because they are more inter- ested in achievement than personal appearance. By con- trast, women who are not feminists may be more committed to romantic ideals and may report themselves to be more attractive to be consistent with this ideal. They may even request attractiveness feedback from others frequently and remember it easily for this reason. Thus, the present find- ings are suspect, and future research should use indepen- dent raters’ assessment of attractiveness, following Golberg et al. (1975). STUDY 2 Because heterosexual men and women rely on one another for love and sexual gratification, it is not surprising that feminism might be shunned if it is perceived to be at odds with these basic human needs. In Study 2, we extended our analysis by employing a romantic conflict index, designed to assess concerns that feminism is troubling for roman- tic relationships. For example, men may be unwilling to date feminists if they are more interested in women who would comply with traditional roles (Glick & Fiske, 1996). Moreover, female feminists have often been maligned in the media as radical man-haters, which could lead to the perception that they are more resentful toward men than nonfeminists (Bell & Klein, 1996; Misciagno, 1997). If fem- inism and romance are perceived to be at odds, participants who score high on the romantic conflict index should score low on the feminist identity and attitude measures. They might also not show support for women’s civil rights—an outcome that has serious implications for gender parity. Thus, we added a measure of support for these rights in Study 2. Finally, we assessed the stereotype that feminists are les- bians because Study 1 suggested a negative link between beauty and homosexuality that stigmatizes feminists. It was possible that the lesbian stereotype would prove to be an important predictor of participants’ own feminist orienta- tions (e.g., participants might dislike feminists because they do not like lesbians, or seek to avoid the stigma of being identified with lesbians; Swim et al., 1999). Although we expected the lesbian stereotype to dampen enthusiasm for feminism, Study 2 allowed us to compare this effect to the romantic conflict index, which we hypothesized would also be an important predictor of feminist identity and attitudes and support for women’s civil rights. Method Participants Heterosexual volunteers (N = 236; 121 women, 115 men) participated in exchange for partial credit toward their In- troductory Psychology research participation requirement. Of these, 127 (54%) were European American, 85 (36%) were Asian American, 26 (11%) were African American, 18 (19%) were Latino/a, and the remainder reported another ethnic identity. Of the original sample (N = 241), 5 par- ticipants were excluded because they indicated they were lesbian, gay, or bisexual (2 women, 3 men). Feminism and Romance 131 Measures Feminist orientations. Following Study 1, participants indicated “How do you feel toward feminists?” on a thermometer scale ranging from 1 (very cold) to 10 (very warm). They also indicated their feminist identity using Study 1’s index (α = .78). To assess support for women’s civil rights, participants rated their agreement with four items on scales ranging from 1 (strongly disagree) to 10 (strongly agree). The items were: “The U.S. should pass the ERA (Equal Rights Amendment) to ensure equality for women”; “Women should have the same career opportuni- ties that men have”; “Women should not have to put up with sexual harassment”; and “Women and men should have the same sexual freedoms.” Responses were averaged to form the civil rights index (α = .73). Romantic conflict. To assess perceived conflict be- tween feminism and romance, we administered four items on scales ranging from 1 (strongly disagree) to 10 (strongly agree). The items were: “Most men would probably not want to date a feminist”; “Feminism and romance do not mix well for women”; “Feminism can cause women to resent men”; and “Feminism can add stress to relationships with men.” These were averaged to form the romantic conflict index (α = .80). Lesbian stereotype. Using the identical scales, partici- pants rated their agreement with four items that were av- eraged to form the lesbian stereotype index (α = .89). The items were: “Most feminists are probably lesbians”; “The leaders of the feminist movement tend to be lesbians”; “Lesbians have taken over the feminist movement”; and “Feminism focuses too much on lesbian issues.” Procedure Participants were escorted to individual booths equipped with a desktop PC by an experimenter, who then adminis- tered instructions and started the computer program. Par- ticipants first completed the measures in the order de- scribed above. Following this, they completed demographic measures (i.e., gender, ethnicity, age, and sexual orien- tation). Within each measure, items were randomly pre- sented. Following completion of these measures, partici- pants were thanked and debriefed. Results and Discussion Gender Differences Table 3 shows Study 2’s descriptive statistics separately by gender. Not surprisingly, women scored higher than men on the feminist identity, attitude, and civil rights indices. Effect sizes were large for each of the variables (d > .50 for all). Although we did not predict gender differences on the romantic conflict and lesbian stereotype indices, Table 3 shows that men scored higher than women on both mea- Table 3 Descriptive Statistics as a Function of Participant Gender (Study 2) Women Men Sex Difference Measure M SD M SD t d Feminist identity 3.65 .91 3.06 .87 5.04∗∗ .63 Feminist attitude 5.97 2.17 4.72 2.06 4.07∗∗ .57 Civil rights 9.15 1.19 8.31 1.67 4.37∗∗ .56 Romantic conflict 4.98 1.92 6.02 1.90 4.20∗∗ −.52 Lesbian stereotype 4.03 2.11 4.72 2.06 2.53∗ −.32 Note. The effect size (Cohen’s d) for gender differences is based on the pooled standard deviation. Positive scores indicate women scored higher than men. By convention, small, moderate, and large effect sizes corre- spond to .20, .50, and .80, respectively (Cohen, 1988). ∗p < .05. ∗∗p < .01. sures. Thus, men were particularly likely to view feminism as being in conflict with heterosexual relationships. Perceived Conflict Between Feminism and Romance Our primary hypothesis was that heterosexual women and men would show resistance to feminism to the extent they believed it conflicted with romance. We also sought to com- pare romantic conflict and the lesbian stereotype as predic- tors of feminist identity and attitudes. Not surprisingly, the lesbian stereotype and romantic conflict indexes covaried for both women and men (r = .57 and .54, respectively, p < .001), suggesting that the stereotype accounts for the perception that feminism conflicts with romance (or vice versa). Nonetheless, we examined their separate relation- ships with feminist orientations. Table 4 shows the results separately by gender. As can be seen, the pattern of relationships was sim- ilar for both genders. First, romantic conflict negatively correlated with feminist identity and attitude and support for women’s civil rights. Thus, participants who believed that feminism is troubling for romance also showed less Table 4 Correlations as a Function of Participant Gender (Study 2) Romantic Lesbian Civil conflict stereotype rights Women Feminist identity −.32∗∗ −.28∗∗ .20∗ Feminist attitude −.31∗∗ −.27∗∗ .25∗∗ Civil rights −.19∗ −.05 — Men Feminist identity −.47∗∗ −.24∗ .21∗ Feminist attitude −.49∗∗ −.42∗∗ .27∗∗ Civil rights −.24∗ −.26∗∗ — Note. Correlations are based on 121 women and 115 men. ∗p < .05. ∗∗p < .01. 132 RUDMAN AND FAIRCHILD enthusiasm for feminists and policies that support women. These results support our hypothesis that romance-related concerns about feminism may be an important barrier to gender equity. Table 4 also shows similar relationships for the lesbian stereotype index, with one exception. Men (but not women) who stereotyped feminists as lesbians also showed less support for women’s civil rights, compared with men who did not endorse this stereotype. The next analysis checked on the ability of the roman- tic conflict and lesbian stereotype measures to contribute unique variance to feminist identity and feminist attitude scores. Because feminist identity and attitude scores were highly related for both genders, r > .54 for both, p < .001, we averaged their standardized scores to form a single index of feminist orientations. We then regressed feminist orien- tations on the romantic conflict and lesbian stereotype in- dexes. For men, results showed that romantic conflict, but not the lesbian stereotype, remained a predictor of feminist orientations, β = −.49, p < .001, and β = −.10, p = .22, respectively (overall R2 = .32). The results for women were similar (romantic conflict, β = −.21, p < .05; lesbian stereo- type, β = −.16, p = .18; overall R2 = .15). Thus, concerns about the influence of feminism on romantic relationships uniquely contributed to weak feminist identity and attitude for men and women alike. Similar analyses favored romantic conflict over the les- bian stereotype as a predictor of support for women’s civil rights. For women, romantic conflict was a unique predic- tor, whereas the lesbian stereotype was nonsignificant, β = −.24, p < .05, and β = −.09, p = .44, respectively (over- all R2 = .20). For men, romantic conflict was a marginally significant predictor, whereas the lesbian stereotype was nonsignificant, β = −.19, p < .08, and β = −.13, p = .22, respectively (overall R2 = .29). In sum, Study 2 showed that people who viewed femi- nism as troubling for romance were unlikely to show fem- inist orientations or to support women’s civil rights, even after accounting for the lesbian stereotype. Thus, Study 2 continued to support the hypothesis that romantic con- cerns about feminism have unfavorable implications for the progress of the Women’s Movement. Moreover, Study 2 (unexpectedly) suggested that these concerns are more im- portant than the lesbian stereotype when determining par- ticipants’ own feminism. STUDY 3 In Study 3, we examined whether feminism might be viewed as incompatible with satisfying sexual relations. Specifically, we hypothesized that female assertiveness and autonomy, mechanisms that promote gender equality, might be seen as undermining heterosexual intimacy, particularly among men. Gender socialization includes learning sexual scripts that dictate greater dominance for men, whereas women’s roles are more passive and submissive (Impett & Peplau, 2003; Sanchez, Crocker, & Boike, 2005). As a result, men are more likely to initiate and determine the nature of sex- ual contact (Browning, Hatfield, Kessler, & Levine, 2000). They also exhibit a more dominant sexual drive compared to women (Baumeister, Catanese, & Vohs, 2001). Thus, tra- ditional sexual scripts are at odds with female assertiveness and autonomy, attributes that enhance women’s ability to compete successfully in performance settings. To the ex- tent that individuals presume that, for sex to be satisfying, men must be in charge or dominant and women more sub- missive, they should score low on the feminist attitude and support for civil rights measures. If so, specific concerns about men’s sexual needs may also play a role in preserving sexism. Finally, Study 3 allowed us to replicate Study 2’s finding that romantic conflict negatively predicted feminist orien- tations. To enhance the generalizability of our results, we used different measures to assess (a) attitudes toward fem- inism and (b) romantic conflict (i.e., perceptions that fem- inism and romance are incompatible). We expected sexual and romantic conflict to be negative predictors of feminist attitudes and support for women’s civil rights. Method Participants Heterosexual volunteers (N = 173; 106 women, 67 men) participated in exchange for partial credit toward their In- troductory Psychology research participation requirement. Of these, 90 (52%) were European American, 54 (31%) were Asian American, 21 (12%) were African American, 15 (9%) were Latino/a, and the remainder reported another ethnic identity. Of the original sample (N = 180), 7 were eliminated because they indicated they were lesbian, gay, or bisexual (4 women, 3 men). Materials Feminist orientations. Participants completed the Feminist and Women’s Movement Scale (Fassinger, 1994), responding on a scale ranging from 1 (strongly disagree) to 7 (strongly agree). Sample items included “Feminist princi- ples should be adopted everywhere,” “Feminists are a men- ace to society,” and “The Women’s Movement is too radical and extreme in its views.” After appropriate recoding, items were averaged to form the feminist attitude index (α = .82), on which high scores reflected pro-feminist attitudes. Par- ticipants also completed Study 2’s civil rights index (α = .84). Sexual and romantic conflict. As a measure of sexual conflict, participants responded to four items on a scale ranging from 1 (strongly disagree) to 10 (strongly agree). The items were: “Men perform better sexually when they are in charge”; “Romance depends, in part, on men be- ing allowed to be in charge”; “If men and women behaved the same, it would take the mystery out of love”; and “If Feminism and Romance 133 Table 5 Correlations and Means as a Function of Participant Gender (Study 3) Women Men Feminist Romantic Sexual Feminist Romantic Sexual attitude conflict conflict attitude conflict conflict Romantic conflict −.47∗∗ −.64∗∗ Sexual conflict −.40∗∗ .08 −.27∗ .14 Civil rights .39∗∗ −.27∗∗ −.35∗∗ .31∗ −.40∗∗ −.28∗∗ Mean 4.94a 3.69a 3.31a 4.38b 3.91a 3.98b SD 0.74 1.38 1.21 0.83 1.41 1.18 Note. Correlations are based on 106 women and 67 men. Means not sharing a subscript differ between female and male participants at the p < .01 level or greater. ∗p < .05. ∗∗p < .01. women want to be loved, they should not compete against men.” Items were averaged to form the sexual conflict index (α = .72).4 For the measure of romantic conflict, partici- pants rated their agreement with a single item, “Feminism has negatively influenced relationships between men and women,” on a scale ranging from 1 (strongly disagree) to 7 (strongly agree). Procedure Measures were administered by a computer program to par- ticipants in individual booths in the order described above. Within each measure, items were randomly presented. Par- ticipants were then thanked and debriefed. Results and Discussion Preliminary Analyses Table 5 shows descriptive statistics for the Study 3 vari- ables separately by gender. Women scored higher on the feminist attitude index than did men; the effect size was moderate (d = .67). In addition, men showed higher sexual conflict scores than women (d = −.69). That is, men were more likely than women to view female assertiveness and independence as predictors of sexual conflict. In contrast to Study 2, there was no significant gender difference on the romantic conflict scale (d = −.16). That is, women were no less likely than men to agree that feminism has negatively influenced male–female relations. Finally, as in Study 2, and not shown in Table 5, women scored higher than men on the civil rights index (M = 9.27 vs. 8.41, SD = 1.00 and 1.41, respectively, d = .69). Predicting Feminist Orientations From Romantic Relationship Conflicts In line with Study 2, we expected romantic conflict scores to negatively covary with feminist attitudes and civil rights support. To extend Study 2’s findings, we expected a simi- lar pattern for the sexual conflict index. Table 5 shows the key relationships separately by gender. As predicted, the romantic conflict and sexual conflict indexes were nega- tively associated with the feminist attitude and civil rights measures for both genders. The pattern is consistent with the hypothesis that low enthusiasm for feminists may stem, in part, from a perceived conflict between feminism and harmonious romantic relationships. Finally, romantic and sexual conflict scores were negligibly related for women and men, rs = .08 and .14, ns, respectively. This suggests that they are distinguishable as predictors of feminist ori- entations. In sum, Study 3’s findings supported the hypothesis that beliefs about the incompatibility of romance and feminism decrease support for feminism, and we extended our anal- ysis to include specific concerns about sexual harmony. Be- cause sexual conflict and romantic conflict scores were un- related, individuals can have independent concerns about the impact of feminism on the health of their intimate rela- tionships; however, each can negatively predict support for feminism and women’s civil rights. GENERAL DISCUSSION Gender represents a unique social category for intergroup relations research because heterosexual men and women rely on each other to fulfill basic needs, including sexual gratification and relational well being. In three investiga- tions, we found support for our hypothesis that intimate re- lationship concerns can undermine feminism’s appeal and, therefore, perhaps undermine collective action on women’s behalf. Results consistently showed a negative association between feminist orientations and perceived conflict be- tween feminism and romantic relationship concerns (at- tractiveness, sex appeal, romance, and sexual harmony) for women and men alike. Are Beauty and Romance Incompatible With Feminism? Study 1 revealed that the unattractive feminist stereotype is alive and well, just as it was 30 years ago (Goldberg et al., 1975). Both genders rated plain women as more likely to be feminists compared to pretty women. However, the stereo- 134 RUDMAN AND FAIRCHILD type was fully explained by beliefs that unattractive women are likely to be lesbians. Moreover, the negative link be- tween perceived beauty and lesbianism was fully accounted for by the belief that plain women are low on sex appeal (i.e., unpopular dating-wise with men). As a result of these unfavorable beliefs, young adults may view feminism as un- romantic and a hindrance to their own relationships. Past research has consistently uncovered a bias against feminists. The present investigations were concerned with the possibility that people may avoid identifying with and liking feminism (i.e., treat it as the F word) to the extent that it is perceived as a roadblock to emotional or sexual hap- piness. Support for this hypothesis was shown in Studies 2 and 3. Beliefs that feminism was problematic for hetero- sexual relationships negatively predicted feminist identity and attitudes and support for women’s civil rights, even af- ter accounting for the lesbian feminist stereotype (Study 2) and sexual conflict (Study 3). Although men tended to show higher romantic and sexual conflict scores than women, the predictive utility of these scores vis-à-vis feminist orien- tations was remarkably similar for both genders in both studies. Limitations and Future Directions Although we can claim with confidence that a perceived conflict between feminism and romance is linked to low support for feminists, our research cannot speak to the causal direction. Future research is needed to determine whether people justify negative reactions to feminists on the basis of relationship concerns, as opposed to these concerns leading to less support for feminism. Moreover, our research cannot determine the accuracy of beliefs that feminism is incompatible with romance and sexual intimacy. Future investigations should examine whether feminist women (or men with feminist partners) are likely to have troubled relationships, compared with people in tra- ditional relationships. Finally, future investigations should employ measures of feminist identity development (Bargad & Hyde, 1991; Downing & Roush, 1985), as well as feminist activism (Stake, Roades, Rose, Ellis, & West, 1994), which could provide more nuanced results. In concert, the present findings suggest that a mar- riage between romantic relationship and sexism research is overdue when considering the factors that underlie gen- der hegemony. Although the fact that gender represents a special case for intergroup relations has been appreciated (Fiske & Stevens, 1993; Rudman & Goodwin, 2004), the in- fluence of heterosexuality on sexism has been largely over- looked (cf. Glick & Fiske, 1996; Holland & Eisenhart, 1990; Rudman & Goodwin, 2004; Rudman & Heppen, 2003). As a result, there are several roads that might be taken by researchers. First, future investigations of the negative link between beauty and perceived feminism are needed to illu- minate its underpinnings. For example, heterosexuals may presume that pretty women can obtain social status and prestige indirectly, through men, and therefore need not commit themselves to obtaining gender parity to the same degree that plain women must to be financially secure. Sec- ond, it might be fruitful to explore the relationship between romantic ideology and gender competition. For example, women may be reluctant to change the gender dynamic from benevolence (men protecting women) to competing with men for economic resources for fear that it will impede their love life, including their ability to marry. Men, too, might resist a change in this dynamic, given their vested in- terest in male hegemony. In short, romance may be viewed as a means of achieving power more so for women than for men, and this fact may make feminism threatening to women and men for different, as well as similar, reasons. Fi- nally, individual differences in the perception that feminism conflicts with romance should be investigated. For example, young women, who may be particularly interested in having romantic relationships, may be more afraid than their older counterparts that men will not want to date them if they are feminists. In addition, women in secure marriages may be less concerned with unromantic perceptions of feminists. The successful protection and enlargement of women’s rights requires women’s enthusiastic defense, which, we suspect, entails recognizing that feminism is not anti-male. Instead, feminism is a humanist movement, dedicated to gender equality, rather than benefiting women at the ex- pense of men. Therefore, understanding the specific barri- ers to embracing feminism for both genders is critical for guaranteeing its health. For this reason, we believe that a consideration of the implications of heterosexual relation- ships for feminism is long overdue, but will ultimately bear fruit in the struggle to overcome resistance to gender equity. Initial submission: April 5, 2005 Initial acceptance: February 24, 2006 Final acceptance: April 14, 2006 NOTES 1. Cohen’s d is an effect size measure. By convention, small, mod- erate, and large effect sizes correspond to .20, .50, and .80, respectively (Cohen, 1988). 2. It is interesting to note that, in their research, Unger et al. (1982) found a general tendency for political deviants to be rated as unattractive, not just feminists. 3. We used photos from 1974 to replicate Goldberg’s paradigm and because contemporary high school senior photos resem- ble model photo shoots, complete with costumes and various settings. In particular, the overt emphasis on sex appeal (e.g., provocative clothing and poses) prohibited using contempo- rary photos. We used our own judgment to determine pretty and plain targets; a manipulation check from our participants confirmed our decisions (see the Results section). 4. We originally factor analyzed a 7-item scale and chose the four items in the sexual conflict index because they formed the first factor (eigenvalue = 3.97, variance accounted for = 41%, all factor loadings > .70), whereas the remaining three Feminism and Romance 135 items formed the second factor (eigenvalue = 1.22, variance accounted for = 12%, all factor loadings > .56). The items in the second factor appeared to reflect traditional gender roles (“Women should relax and let men be the primary breadwin- ner in the family,” “Women and men should not compete for the same jobs,” and “Women should not try to succeed in male occupations”). REFERENCES Bargad, A., & Hyde, J. S. (1991). Women’s studies: A study of femi- nist identity development in women. Psychology of Women Quarterly, 15, 181–201. Baron, R. M., & Kenny, D. A. (1986). The moderator-mediator variable distinction in social psychological research: Con- ceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology, 51, 1173–1182. Baumeister, R. F., Catanese, K. R., & Vohs, K. D. (2001). Is there a gender difference in strength of sex drive? Theoretical views, conceptual distinctions, and a review of relevant ev- idence. Personality and Social Psychology Review, 5, 242– 273. Bell, D., & Klein, R. (1996). Radically speaking: Feminism re- claimed. Victoria, Australia: Spinifex Press. Browning, J. R., Hatfield, E., Kessler, D., & Levine, T. (2000). Sexual motives, gender, and sexual behavior. Archives of Sexual Behavior, 29(2), 135–153. Buschman, J. K., & Lenart, S. (1996). “I am not a feminist, but. . .”: College women, feminism, and negative experiences. Polit- ical Psychology, 17, 59–75. Cejka, M. A., & Eagly, A. H. (1999). Gender-stereotypic images of occupations correspond to the sex segregation of em- ployment. Personality and Social Psychology Bulletin, 25, 413–423. Chodorow, N. (1979). Feminism and difference: Gender, rela- tion, and difference in psychoanalytic perspective. Socialist Review, 46, 42–64. Cohen, J. (1988). Statistical power for the behavioral sciences. Hillsdale, NJ: Erlbaum. Dion, K. K., Berscheid, E., & Walster, E. (1972). “What is beautiful is good.” Journal of Personality and Social Psychology, 24, 285–290. Downing, N., & Roush, K. L. (1985). From passive acceptance to active commitment: A model of feminist identity de- velopment for women. Counseling Psychologist, 13, 695– 709. Eagly, A. H. (1987). Sex differences in social behavior: A social role interpretation. Hillsdale, NJ: Erlbaum. Fassinger, R. E. (1994). Development and testing of the Attitudes Toward Feminism and the Women’s Movement (FWM) Scale. Psychology of Women Quarterly, 18, 389–402. Fiske, S. T., Cuddy, A. J., Glick, P., & Xu, J. (2002). A model of (often mixed) stereotype content: Competence and warmth respectively follow from perceived status and competition. Journal of Personality and Social Psychology, 82, 878– 902. Fiske, S. T., & Stevens, L. E. (1993). What’s so special about sex? Gender stereotyping and discrimination. In S. Oskamp & M. Costanzo (Eds.), Gender issues in contemporary society (pp. 173–196). Newbury Park, CA: Sage. Gilligan, C. (1982). In a different voice: Psychological theory and women’s development. Cambridge, MA: Harvard University Press. Glick, P., & Fiske, S. T. (1996). The Ambivalent Sexism Inventory: Differentiating hostile and benevolent sexism. Journal of Personality and Social Psychology, 70, 491–512. Glick, P., & Fiske, S. T. (1999). The Ambivalence toward Men Inventory: Differentiating hostile and benevolent beliefs about men. Psychology of Women Quarterly, 23, 519– 536. Goldberg, P. A., Gottesdiener, M., & Abramson, P. R. (1975). Another put-down of women? Perceived attractiveness as a function of support for the feminist movement. Journal of Personality and Social Psychology, 32, 113–115. Greenwald, A. G., Banaji, M. R., Rudman, L. A., Farnham, S. D., Nosek, B. A., & Mellott, D. S. (2002). A unified theory of im- plicit attitudes, stereotypes, self-esteem, and self-concept. Psychological Review, 109, 3–25. Haddock, G., & Zanna, M. P. (1994). Preferring “housewives” to “feminists”: Categorization and the favorability of attitudes toward women. Psychology of Women Quarterly, 18, 25– 52. Holland, D. C. (1992). How cultural systems become desire: A case study of American romance. In R. G. D’Andrade & C. Strauss (Eds.), Human motives and cultural models (pp. 61–89). New York: Cambridge University Press. Holland, D. C., & Eisenhart, M. A. (1990). Educated in romance: Women, achievement, and college culture. Chicago: Univer- sity of Chicago Press. Impett, E. A., & Peplau, L. A. (2003). Sexual compliance: Gender, motivational, and relationship perspectives. Journal of Sex Research, 40, 87–100. Jackman, M. R. (1994). The velvet glove: Paternalism and conflict in gender, class, and race relations. Berkeley, CA: University of California Press. Misciagno, P. S. (1997). Rethinking feminist identification: The case for de facto feminism. Westport, CT: Praeger. Pratto, F., Stallworth, L. M., Sidanius, J., & Siers, B. (1997). The gender gap in occupational role attainment: A social domi- nance approach. Journal of Personality and Social Psychol- ogy, 72, 37–53. Renzetti, C. M. (1987). New wave or second stage? Attitudes of college women toward feminism. Sex Roles, 16, 265– 277. Rubin, L. R., Nemeroff, C. J., & Russo, N. F. (2004). Explor- ing feminist women’s body consciousness. Psychology of Women Quarterly, 28, 27–37. Rudman, L. A., & Goodwin, S. A. (2004). Gender differences in automatic ingroup bias: Why do women like women more than men like men? Journal of Personality and Social Psy- chology, 87, 494–509. Rudman, L. A., & Heppen, J. (2003). Implicit romantic fantasies and women’s interest in personal power: A glass slipper ef- fect? Personality and Social Psychology Bulletin, 29, 1357– 1370. Sanchez, D., Crocker, J., & Boike, K. R. (2005). Doing gender in the bedroom: Investing in gender norms and the sexual experience. Personality and Social Psychology Bulletin, 31, 1445–1455. Sobel, M. E. (1982). Asymptotic confidence intervals for indirect effects in structural equations models. In S. Leinhart (Ed.), 136 RUDMAN AND FAIRCHILD Sociological methodology 1982 (pp. 290–312). San Fran- cisco: Jossey-Bass. Stake, J. E., Roades, L., Rose, S., Ellis, L., & West, C. (1994). The women’s studies experience: Impetus for fem- inist activism. Psychology of Women Quarterly, 18, 17– 24. 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 Psy- chology, 21, 61–68. Unger, R. K., Hilderbrand, M., & Madar, T. (1982). Physical at- tractiveness and assumptions about social deviance: Some sex-by-sex comparisons. Personality and Social Psychology Bulletin, 8, 293–301. Williams, R., & Wittig, M. A. (1997). “I’m not a feminist, but. . .”: Factors contributing to the discrepancy between pro-feminist orientation and feminist social identity. Sex Roles, 37, 885–904. Wolff, N. (1991). The beauty myth: How images of beauty are used against women. New York: William Morrow.