<<

Radke Helena (Orcid ID: 0000-0003-2811-0976)

Running Head: SOCIAL DOMINANCE ORIENTATION AND

Negotiating the hierarchy: Social dominance orientation among women is associated with the endorsement of benevolent sexism

Helena R. M. Radke1, Matthew J. Hornsey2, Chris G. Sibley3, and Fiona Kate Barlow2

1 Institute for , University of Osnabrück

2 School of Psychology, University of Queensland

3 School of Psychology, University of Auckland

Word Count: 6916

KEYWORDS: , Political Psychology, /Stereotyping, Sexism, Social Dominance

Orientation, Status

Authors’ Note: Please direct all correspondence to Helena R. M. Radke, Institute for Psychology,

University of Osnabrück, Osnabrück, Germany, 49074, E-Mail: [email protected]

This is the author manuscript accepted for publication and has undergone full peer review but has not been through the copyediting, typesetting, pagination and proofreading process, which may lead to differences between this version and the Version of Record. Please cite this article as doi: 10.1111/ajpy.12176

This article is protected by copyright. All rights reserved. Figure_1 (1).tiff

This article is protected by copyright. All rights reserved. Radke Helena (Orcid ID: 0000-0003-2811-0976)

Running Head: SOCIAL DOMINANCE ORIENTATION AND SEXISM

Negotiating the hierarchy: Social dominance orientation among women is associated with the endorsement of benevolent sexism

Word Count: 6916

This article is protected by copyright. All rights reserved. SOCIAL DOMINANCE ORIENTATION AND SEXISM 2

Abstract Objective: We examine how women high in social dominance orientation reconcile supporting a social system that seemingly disadvantages them. We propose that women high in social dominance orientation are more likely to adopt a benevolently sexist worldview.

Method: This paper contains data from two survey studies. Study 1 used data from the New Zealand Attitudes and Values Survey (N = 10,485) which is a study of social attitudes, personality, and health outcomes among New Zealanders. Measures of Social Dominance Orientation, as well as Hostile and Benevolent Sexism were included in Study 1. Study 2 consisted of data from a smaller sample of American females (N = 269). In addition to the variables described above, we also measured the extent to which participants perceived the gender hierarchy to be legitimate and their perceived personal need for protection from men. Control variables included hostile sexism and demographic variables (participants’ age, race, and socio-economic status).

Results: Results revealed that social dominance orientation was a stronger predictor of benevolent sexism among women compared to men. Serial mediation analysis was then conducted in Study 2. Study 2 confirmed the hypothesised mechanism: Higher social dominance orientation was associated with perceptions that women’s low status is legitimate which, in turn, was associated with higher perceived personal need for protection from men. This was then associated with the endorsement of a benevolently sexist worldview.

Conclusion: Our findings suggest that social dominance orientation might help to explain why some women endorse benevolent sexism.

KEYWORDS: Gender, Political Psychology, Prejudice/Stereotyping, Sexism, Social

Dominance Orientation, Status

This article is protected by copyright. All rights reserved. SOCIAL DOMINANCE ORIENTATION AND SEXISM 3

Key Points What is already known about this topic; • Social dominance orientation (SDO) is a “general attitudinal orientation toward , reflecting whether one generally prefers such relations to be equal, versus hierarchical” (Pratto, Sidanius, Stallworth, & Malle, 1994, p. 742). • SDO is often studied among advantaged group members but not disadvantaged group members • Some studies have found a link between SDO and sexism among men and women

What this topic adds; • We contribute to the limited literature investigating SDO among disadvantaged group members • We find that SDO is a stronger predictor of benevolent sexism among women compared to men • We identify a novel process through which SDO is associated with benevolent sexism for women

This article is protected by copyright. All rights reserved. SOCIAL DOMINANCE ORIENTATION AND SEXISM 4

Negotiating the hierarchy: Social dominance orientation among women is associated with the endorsement of benevolent sexism

According to Glick and Fiske’s (1996) ambivalent sexism theory, sexism manifests in two complementary ways. Hostile sexism is an overt prejudice usually directed toward women who embody non-traditional gender roles. In contrast, those high in benevolent sexism hold seemingly “positive” views of women, but they do so as part of a broader worldview that obligates men to protect “traditional” women. Both hostile and benevolent sexism have been found to negatively impact women; for example, they are associated with in the workplace (Good & Rudman, 2010), and the promotion of sexual violence towards women (Abrams, Viki, Masser, & Bohner, 2003). It is not just men, however, who display sexism towards women. In their influential paper investigating cross-cultural differences in hostile and benevolent sexism, Glick and colleagues (2000) reported a perplexing finding – that some women (especially those from gender unequal nations) were just as likely to endorse benevolent sexism as men. Since the publication of these results, researchers have been trying to understand why some women internalise this form of prejudice towards their own group. In this paper we explore the possibility that women who adopt a benevolently sexist worldview do so because they orient towards the world in a socially dominant way. The Relationship between Social Dominance Orientation and Sexism Social dominance orientation (SDO) is a “general attitudinal orientation toward intergroup relations, reflecting whether one generally prefers such relations to be equal, versus hierarchical” (Pratto, Sidanius, Stallworth, & Malle, 1994, p. 742). As one might expect, those group members who are advantaged by the existing social system are more likely to endorse a socially dominant orientation compared to those who are disadvantaged by the social hierarchy. Men, for example, have consistently higher levels of SDO relative to women (Pratto, Sidanius, & Levin, 2006). The link between hostile sexism and SDO is theoretically intuitive and well-supported by the psychological literature. People high in SDO support hierarchies, and so one would expect them to support the (patriarchal) gender hierarchy. Given that hostile sexism is directed towards women who challenge the gender hierarchy (e.g., feminists and career women) it is not surprising that SDO is positively correlated with hostile sexism among both men and women (Christopher & Mull, 2006; Christopher & Wojda, 2008; Fraser, Osborne, &

This article is protected by copyright. All rights reserved. SOCIAL DOMINANCE ORIENTATION AND SEXISM 5

Sibley, 2015; Kelly, Dubbs, & Barlow, 2013; Kteily, Ho, & Sidanius, 2012; Sibley & Overall, 2011). Some studies also suggest that SDO and benevolent sexism are positively related for both women and men (Christopher & Mull, 2006; Christopher & Wojda, 2008; Fraser et al., 2015; Kteily et al., 2012), although this relationship is smaller and less reliable than the link between SDO and hostile sexism. For example, in an article examining prejudice towards women in the workplace, Christopher and Wojda (2008) found that social dominance orientation (measured using the Pratto et al., [1994] 14-item scale) and benevolent sexism (measured using the Glick & Fiske [1996] 11-item scale) were positively and significantly correlated (.27) in a sample of 349 U.S residents. In another paper, Fraser et al. (2015) investigated whether benevolent sexism moderated the relationship between social dominance orientation and support for gender-based affirmative action in a nationally representative sample of over 5,000 New Zealanders. They measured SDO using the six items from the original SDO scale (Pratto et al., 1994) and five items from the benevolent sexism scale (Glick & Fiske, 1996), and found that SDO was positively and significantly associated with benevolent sexism for both men (.11) and women (.20). These results, however, are counter to other research which has found an inconsistent relationship between SDO and benevolent sexism for men and women. In a paper investigating gender differences in intimate partner preferences, a positive and significant correlation between SDO (measured using 4 items from the Sidanius & Pratto [1999] scale) and benevolent sexism (measured using 11 items from the Glick & Fiske [1996]) was found for women (.21) but not for men (.10) in a sample of 492 New Zealand undergraduate students (Sibley & Overall, 2011). Mosso and colleagues (2013) found no relationship between SDO (measured using the SDO6 scale; Pratto et al., [2006]) and benevolent sexism (measured using the 11-item Glick & Fiske [1996] scale) among women (-.03) and men (-.12) in a sample of 544 Italian students (with ages ranging from 14 to 20 years). They did, however, find a positive and significant relationship between SDO and benevolent sexism among women (.27) and men (.18) in a sample of 297 American students (with ages again ranging from 14 to 20 years). This article was investigating the role of legitimizing ideologies (i.e., SDO, ) when predicting ambivalent sexism across cultures. These conflicting results suggest the need for further research in this area. We speculate that one reason for these conflicting results is that some of the previously described studies do not control for relevant extraneous variables. Specifically, hostile and benevolent

This article is protected by copyright. All rights reserved. SOCIAL DOMINANCE ORIENTATION AND SEXISM 6

sexism are correlated with one another, and it is recommended by Glick and Fiske (1996) to control for hostile sexism when examining the effects of benevolent sexism and vice versa. It is possible, especially when examining effects in small samples, that it is necessary to control for hostile sexism to reveal the unique association between benevolent sexism and SDO. Consequently, we controlled for hostile sexism in all analyses in this paper, and further, test our predictions in a large and more powerful sample in Study 1. Above we cite research showing mixed support for our core proposition that women who prefer hierarchically based social structures (i.e., are high in SDO) might also be more likely to report high levels of benevolent sexism. It is important, however, to consider not only past research, but the theoretical basis for the claims we make. Below we draw on social dominance theory, as well as related theories (belief in a just world, system justification theory, cognitive dissonance theory) to make the case that SDO should theoretically be related to women’s levels of benevolent sexism. People are motivated to believe that the world in which they live is just and fair (i.e., just world beliefs; Furnham, 2003; Lerner, 1980). Building on this, system justification theory (Jost, Banaji, & Nosek, 2004) proposes that individuals have a psychological need to maintain stability in their lives, which prompts them to justify the existing social order (cf. Brandt, 2013). These researchers argue that lower-status group members are more likely to justify the social system because these individuals experience an inconsistency between their need to justify the existing social order and the awareness that they are disadvantaged by the social system because of their group membership (Jost, Pelham, Sheldon, & Sullivan, 2003). Previous research supports this argument; Jost and Kay (2005) found that exposure to benevolent sexism among women was associated with increased support for the status quo, and other research has found that these system justifying beliefs map closely onto SDO (Jost & Thompson, 2000). Drawing from this line of research, we start to see some theoretical reasons for why SDO might be linked to women’s benevolent sexism. In particular, women who support the current hierarchy might also be motivated to see it as just and fair. Such a perception, we suggest, might easily flow on to adopting a worldview that obligates help and protection from those higher up in the hierarchy, such as men. Cognitive dissonance theory, which is closely linked to system justification theory, is another approach that can help explain our proposed relationship between women’s endorsement of SDO and benevolent sexism. This theory proposes that when people experience inconsistent beliefs or behaviours they either justify their cognitions or change

This article is protected by copyright. All rights reserved. SOCIAL DOMINANCE ORIENTATION AND SEXISM 7

their behaviours (Festinger, 1957). We argue that women who are relatively high in SDO experience an internal inconsistency between their endorsement of the social hierarchy and the knowledge that they have a lower-status position in society. This discrepancy may motivate women to orient towards worldviews that are (1) consistent with the existing gender hierarchy, and (2) simultaneously place special importance on women’s position in society (i.e., benevolent sexism). More specifically, the argument that women’s SDO might contribute to their benevolent sexism is consistent with preliminary research into women’s ambivalent sexism. Specifically, some researchers argue that women are motivated to adopt benevolently sexist views to protect themselves in hostilely sexist environments (i.e., the protection racket effect; Fischer, 2006; Glick et al., 2000; Sibley & Overall, 2011). For example, Fischer (2006) made the case that women might see (and espouse) benevolent sexism in an attempt to deflect male negativity. As highlighted above, while benevolent sexism does prescribe restrictive and limited roles to women, it also obligates men to behave in a protective and “chivalrous” manner towards women. Consistent with this argument, Fisher (2006) found that women who were told that men held negative attitudes towards women subsequently reported higher benevolent sexism than women who had been told that men held positive attitudes, or those in a control condition. Drawing on this work, we propose that women might respond to a general social system that disadvantages them by increasing the extent to which they support an ideology that obligates male support of women. Specifically, we suggest that those women who are relatively high in SDO will see the social system as legitimate and, through this, perceive that they require male protection. We expect that this perceived personal need for male protection will predict the adoption of a benevolently sexist worldview. While we have argued that women high in SDO will recognise (and support) a hierarchy in which men are afforded higher status than women, we acknowledge that an intuitive alternate hypothesis can be made. Namely, it is possible that women high in SDO do not acknowledge or accept that women have a low-status position in the social hierarchy. They may instead argue that women are privileged and high-status. We can test our main hypothesis against this competing possibility by examining the association between women’s SDO and the extent to which they see the existing gender hierarchy (that favours men over women) as legitimate in Study 2. This argument, however, is contrary to previous research which has found that socially dominant disadvantaged group members support a social

This article is protected by copyright. All rights reserved. SOCIAL DOMINANCE ORIENTATION AND SEXISM 8

hierarchy which maintains their lower-status (Pratto et al., 2006; Sidanius, Levin, Federico, & Pratto, 2001). The Current Studies In Study 1 we tested the basic relationship between SDO and benevolent sexism among men and women using a large, representative sample of New Zealanders. We predicted that SDO among both men and women would be associated with increased levels of benevolent sexism, and further, that this relationship would hold while controlling for hostile sexism. We also tested whether any relationship between SDO and benevolent sexism differed for men and women. In Study 2 we sought to identify the process through which SDO is associated with benevolent sexism among women. This paper has three cases for theoretical impact. First, it seeks to further our understanding of women’s endorsement of benevolent sexism. Second, it contributes to the limited literature examining SDO among lower-status group members; SDO research has largely focused on advantaged group members and so we know relatively little about the psychology of disadvantaged group members who are high in SDO. Finally, Study 2 offers a novel explanation for the relationship between SDO and benevolent sexism among women. Study 1 As discussed above, there is some research demonstrating a positive correlation between SDO and benevolent sexism among men, and a sub-set of this literature has also established the relationship among women. There are, however, some limitations to this empirical base, mainly that (1) the effect is unreliable, and (2) the previous literature typically does not control for relevant extraneous variables such as hostile sexism and demographic variables. Thus, before testing our full mediational model in Study 2, we conducted a basic test of the association between SDO and BS for women in Study 1 in a large, national probability sample. We also controlled for relevant extraneous factors (i.e., hostile sexism, and demographic variables). Method Participants We used data from the Time 4 (2012) New Zealand Attitudes and Values Survey (NZAVS; https://www.psych.auckland.ac.nz/en/about/our-research/research-groups/new- zealand-attitudes-and-values-study.html). The NZAVS is a longitudinal national probability study of social psychological and health outcomes among New Zealanders. We utilised this data as it is a large study in which measures of SDO, Hostile, and Benevolent Sexism are pre-

This article is protected by copyright. All rights reserved. SOCIAL DOMINANCE ORIENTATION AND SEXISM 9

existing. It was thus the ideal dataset with which to rigorously and reliably test our proposed link between SDO and benevolent sexism. A large group of New Zealanders (12,182) filled out some version of the NZAVS. Of these, 10,485 filled out a version of the survey that contained measures of benevolent and hostile sexism. This is the sample that we use for our survey. The participants were 3,922 males and 6,560 females (three participants did not report their gender) with a mean age of 49.65 years (SD = 15.19). The sample was composed mostly of European New Zealanders (76%), followed by Asians (4%), M� oris (3%), Pacific peoples (3%), a combination of these groups (12%), and other racial/ethnic groups (2%). We also measured socio-economic status using the 2013 New Zealand Deprivation Index which gave participants a score from 1 (most affluent) to 10 (most impoverished) depending on the area where they lived. Participants had an average level of deprivation on this scale (M = 4.94; SD = 2.82). Measures Unless otherwise indicated, all measures were recorded on a 1 (strongly disagree) to 7 (strongly agree) scale. Social Dominance Orientation (SDO) was measured using six items from the SDO6 Scale (Pratto et al., 2006; e.g., “It is OK if some groups have more of a chance in life than others”; ± = .73). Higher scores indicated higher levels of social dominance orientation. Hostile Sexism was measured using five items from Glick and Fiske’s (1996) Hostile Sexism Scale (e.g., “Women seek to gain power by getting control over men”; ± = .82). Higher scores indicated higher levels of hostile sexism. Benevolent Sexism was measured using five items from Glick and Fiske’s (1996) Benevolent Sexism Scale (e.g., “Women, compared to men, tend to have greater moral sensibility,”; ± = .74). Higher scores indicated higher levels of benevolent sexism. The Hostile and Benevolent Sexism Scales are subscales of the Ambivalent Sexism Inventory. Demographic variables. In addition to the above variables we included a measure of participant age (measured in years), gender (coded as male = 1 and female = -1), race (coded as European New Zealander = 1 and Non-European New Zealander = -1), and socio- economic status (see below). Participants who identified as a combination of European New Zealander and other ethnic groups were coded as European New Zealander. Socio-economic status was measured using the 2013 New Zealand Deprivation Index (Atkinson, Salmond, & Crampton, 2014; Sibley, 2015). This index combines nine items measuring deprivation in the national census (i.e., adults who received a means-tested benefit,

This article is protected by copyright. All rights reserved. SOCIAL DOMINANCE ORIENTATION AND SEXISM 10

low household income) for distinct geographical areas in New Zealand called “meshblocks”. These meshblocks are defined by Statistics New Zealand (2013), and contain a median of approximately 81 residents (Sibley, 2015). The index assigns a rank from 1 (most affluent) to 10 (most impoverished) to each meshblock. Therefore, residents in the 10% of meshblocks that are most deprived receive a score of 10, and residents in the 10% of meshblocks that are most affluent receive a score of 1. We control for socio-economic status in this study because lower socio-economic status is associated with the endorsement of both SDO (Pratto et al., 2006) and benevolent sexism (Glick et al., 2000). All scales within the NZAVS were necessarily shortened due to time constraints. Items for the scales were selected because they were balanced and broadly representative of the scales. Recent research has found that the 6-item short SDO scale used correlates very strongly with the full SDO6 scale (see Pehrson, Carvacho, & Sibley, 2017, for more information). Previous research has also published results using the shortened hostile and benevolent sexism scales administered in this study (for an example, see Fraser et al., 2015). Results and Discussion Descriptive statistics and correlations among the variables are summarised in Table 1. In Study 1, age, socio-economic status, SDO, benevolent sexism, and hostile sexism were skewed and/or kurtotic. In Study 2, SDO, personal protection, and age were skewed and/or kurtotic. However, we do not believe that this is a problem because regression conducted with a large sample size is robust when the assumption of normality is violated (Howell, 2009). Moreover, log transformation of the variables described above did not change the results in studies 1 and 2 so we reported the untransformed data. Further, less than .6% of the data was missing on any of the core independent and dependent variables (i.e., SDO, participant gender, benevolent sexism, hostile sexism). Sixteen participants (.2% of the sample) did not report their age, 248 participants (2.4% of the sample) did not provide their residential address (which was required to calculate the measure of socio-economic status), 53 participants (.5% of the sample) did not complete the SDO measure, 2 participants (less than .01% of the sample) did not complete the benevolent sexism measure, 1 participant (less than .01% of the sample) did not complete the hostile sexism measure, and three participants (less than .01% of the sample) did not report their gender Consequently, with the power afforded by our large sample size, balanced with our desire to retain the fidelity of the observed pattern of intercorrelations, we used listwise deletion to deal with the small amount of missing data (see Howell, 2009). After list-wise deletion our sample

This article is protected by copyright. All rights reserved. SOCIAL DOMINANCE ORIENTATION AND SEXISM 11

consisted of 10,426 participants when examining the relationship between SDO and benevolent sexism (controlling for hostile sexism) among men and women. We conducted a hierarchical multiple regression to test the relationship between SDO and benevolent sexism for both men and women when controlling for hostile sexism. Hostile sexism was included in Block 1, participant gender and level of SDO was included in Block 2, and the two-way interaction between participant gender and SDO was included in Block 3. As can be seen in Table 2, a significant amount of the variance in benevolent sexism 2 was accounted for in Block 1 (R adj = .18, F (1, 10,424) = 2,311.89, p < .001). Hostile sexism (² = .43, p < .001; 95% CI = .39, .42) was positively associated with benevolent sexism. SDO and participant gender also accounted for a significant amount of the variance in benevolent 2 sexism when entered in Block 2 (R adj = .19, Fch (2, 10,422) = 32.19, p < .001). Specifically, SDO (² = .08, p < .001; 95% CI = .07, .12) was positively correlated with benevolent sexism. In Block 3 the two-way interaction between SDO and participant gender accounted for a 2 significant amount of the variance in benevolent sexism (R adj = .19, Fch (1, 10,421) = 46.12, p < .001). The relationship between SDO and benevolent sexism was qualified by participant gender (² = -.06, p < .001; 95% CI = -.10, -.06). Simple slopes analysis revealed that SDO was a stronger predictor of benevolent sexism among women (² = .09, p < .001; 95% CI = .09, .16) than men (² = .04, p = .011; 95% CI = .01, .08). We also ran the model controlling for participant age, race, and socio-economic status in Block 1, hostile sexism in Block 2, SDO and participant gender in Block 3, and the interaction between SDO and participant gender in Block 4. The interaction remained significant. For the sake of thoroughness, we tested whether participant gender moderated the relationship between SDO and hostile sexism when controlling for benevolent sexism (see Table 3). Benevolent sexism (² = .43, p < .001; 95% CI = .43, .47) was positively associated with hostile sexism in Block 1. In Block 2, SDO (² = .25, p < .001; 95% CI = .30, .35) was positively associated with hostile sexism. Men were also more likely to endorse hostile sexism than women (² = .17, p < .001; 95% CI = .19, .23). However, there was no interaction between SDO and participant gender on hostile sexism in Block 3 (² = -.00, p = .616; 95% CI = -.03, .02). We also ran the model controlling for participant age, race, and socio-economic status in Block 1, hostile sexism in Block 2, SDO and participant gender in Block 2, and the interaction between SDO and participant gender in Block 4. The interaction remained non- significant. We acknowledge, however, that our findings are relatively small given the large

This article is protected by copyright. All rights reserved. SOCIAL DOMINANCE ORIENTATION AND SEXISM 12

sample size we used. Nonetheless, they support our core proposition – that women’s SDO predicts the extent to which they endorse benevolent sexism. Study 2 In Study 2 we aimed to replicate our effect, and determine mechanisms through which SDO is associated with benevolent sexism among women. We proposed that social dominance orientation among women would be associated with benevolent sexism because women high in SDO (a) believe the existing gender hierarchy is legitimate, which in turn predicts (b) a perceived personal need for protection from those at the top of the hierarchy (i.e. men). We tested this possibility in Study 2. Method Participants

A total of 269 female participants (Mage = 33.65; SD = 11.47) living in the U.S. were recruited through an online scientific survey platform (Amazon’s Mechanical Turk or MTurk; www.mturk.com). Participants were paid 90 U.S. cents for completing the study. MTurk has been found to be a useable online platform for recruiting participants (Buhrmester, Kwang, & Gosling, 2011). The sample was composed mostly of Whites (81%), followed by Asians (7%), Blacks (7%), Hispanics/Latinas (4.5%), and other racial/ethnic groups (.5%). We also asked participants to indicate their socio-economic status from 1 (extremely poor) to 7 (extremely wealthy; M = 3.61; SD = 1.20). The majority of participants reported that they had an average (33%; score of 4 on the above scale) or just below average (25%; score of 3 on the above scale) socio-economic status. Measures Unless otherwise indicated, all measures were recorded on a 1 (strongly disagree) to 7 (strongly agree) scale. Race was coded as 1 = White, and -1 = Non-White for this study. Social Dominance Orientation was measured using the complete 16-item SDO Scale (Pratto et al., 2006; ± = .95), Hostile Sexism was measured using the complete 11-item scale (± = .93), and Benevolent Sexism was measured using the complete 11-item scale (± = .91). Legitimacy of the Gender Hierarchy was measured using Glick and Whitehead’s (2010) 6-item scale (e.g., “For the most part, there are legitimate and fair reasons for current differences in men’s and women’s positions in society”, ± = .82). Higher scores indicated greater belief in the legitimacy of the gender hierarchy. Personal Need for Protection was measured using two items (e.g., “I need to be protected by the men in my life” and “I need to be looked after by the men in my life”; r(267)

This article is protected by copyright. All rights reserved. SOCIAL DOMINANCE ORIENTATION AND SEXISM 13

= .80, p < .001). Higher scores indicated increased perceived need for personal protection from men. Results and Discussion Descriptive statistics and correlations among the variables are summarised in Table 4. Serial mediation analysis was conducted using bootstrapping procedures (SPSS PROCESS macro; Hayes, 2013) to examine the indirect effect of SDO on benevolent sexism via legitimacy of the gender hierarchy and then personal need for protection from men. We used 5000 bootstrap samples to estimate -corrected standard errors and 95% percentile confidence intervals for the indirect effect of SDO on the dependent variable via the two serial mediators, while controlling for hostile sexism. As can be seen in Figure 1, legitimacy of the gender hierarchy and then the personal need for protection mediated the association between SDO and benevolent sexism (B = .01, BSE = .01; 95% CI = .00, .03; this pathway was still significant when not controlling for hostile sexism). Our model also held when controlling for participant age, race, socio-economic status and level of hostile sexism. We also tested an alternative pathway, where we flipped the order of the two mediators (personal need for protection and then legitimacy of the gender hierarchy). This model was non-significant (B = .00, BSE = .00; 95% CI = -.00, .02). Moreover, there was no relationship between SDO and hostile sexism through legitimacy of the gender hierarchy and personal need for protection (controlling for benevolent sexism; B = .00, BSE = .00; 95% CI = -.00, .02). Study 2 found that socially dominant women perceived their lower-status position in society to be legitimate and, through this, reported a greater personal need for protection from men. In turn, this need for protection predicted increased benevolent sexism. These findings support the proposed pathway through which SDO is associated with benevolent sexism, and shows that this relationship is specific to this form of prejudice. General Discussion In this paper we investigated how women who are socially dominant might negotiate their lower-status position within the social hierarchy. In doing so, we identified how women who prefer hierarchical social structures (i.e., are high in SDO) reconcile supporting a social system that disadvantages them. In Study 1, we clarified the relationship between SDO and benevolent sexism among men and women using a large, national probability sample. We found that SDO was a stronger predictor of benevolent sexism among women compared to men. Study 2 identified underlying mechanisms for the relationship between SDO and benevolent sexism among women: Socially dominant women perceived women’s lower-

This article is protected by copyright. All rights reserved. SOCIAL DOMINANCE ORIENTATION AND SEXISM 14

status position in society to be legitimate and, through this, perceived a greater need for protection from men. This personal need in turn predicted support for an ideology that reveres women (albeit restrictively) and demands protection of women by men (i.e., benevolent sexism). The first component of the model (SDO  legitimacy) is in line with previous research which has found that disadvantaged group members high in SDO endorse the existing hierarchy (rather than their own ingroup; Pratto et al., 2006; Sidanius et al., 2001). The remainder of the model is broadly consistent with the “protection racket” interpretation of benevolent sexism; that women endorse benevolent sexism because they have limited opportunity for direct access to status and resources in hostilely sexist (and presumably high- SDO) environments (Fischer, 2006; Glick et al., 2000; Sibley & Overall, 2011). Our findings also align with the broader psychological literature investigating why disadvantaged group members support the status quo. For example, through the lens of system justification theory it can be argued that women might internalise their lower status position because they have a need to justify the status quo even if it disadvantages them (Jost et al., 2003; Jost et al., 2004). Our results are also in line with cognitive dissonance theory (Festinger, 1957). Specifically, it is possible that socially dominant women experience a dissonance between their belief that the social hierarchy is natural and good, and the knowledge that they are disadvantaged by the hierarchy. It is also possible that they resolve this dissonance by (1) legitimising the hierarchy, and (2) taking on an ideology that affords them special benefits from men. Socially dominant women could also endorse benevolent sexism because they hold tightly to the belief that the world is just and fair (Furnham, 2003; Lerner, 1980). We are limited in our inability to infer causality from our results because of the cross- sectional nature of our data, the correlational study design we used, and the mediation analysis we conducted. Future research should consider manipulating SDO or examining the relationship between SDO and benevolent sexism longitudinally to establish causality. For example, previous research has manipulated levels of SDO by making salient participants’ membership in low and high status groups (Huang & Liu, 2005). We should note, however, that our presumed causal model is supported by theory, in which SDO is routinely theorised to be an antecedent rather than a consequence of prejudice (e.g., Pratto et al., 1994). Furthermore, the model does not hold when the need for personal protection precede beliefs about the legitimacy of the gender hierarchy. Another alternative explanation for our data is that extraneous variables help account for the relationship between SDO and sexism. Yet this

This article is protected by copyright. All rights reserved. SOCIAL DOMINANCE ORIENTATION AND SEXISM 15

seems unlikely given that the model was tested with and without various extraneous variables (participant age, race, socio-economic status) controlled for in the analyses. Nevertheless, there is potential for future research to examine other variables that provide additional insights into the relationship between SDO and sexism among women. For example, women might be more likely to legitimise their lower-status position if they believe that people (and groups) get what they deserve (i.e., just world beliefs). This direction for future research is particularly relevant given the diverse and varied experiences women have in the social world. Moreover, it may be helpful to identify variables that motivates women to adopt a social dominance orientation in the first place (such as high levels of societal gender inequality or low socio-economic status; Fischer, 2006; Glick et al., 2000). As a final note, the effects we detected are generally small, which hints that other factors might be associated with the extent to which women come to endorse, or reject, benevolent sexism. Future work should therefore investigate SDO in conjunction with other key predictors of women’s benevolent sexism. In sum, we have made a motivated argument for the link between SDO and benevolent sexism among women, finding that women who support the existing hierarchy see their own low-status place in it as legitimate and, through this, perceive that they are in personal need of protection from those at the top of the hierarchy (i.e., men). On a broader note, our findings suggest that socially dominant ideologies might be one reason why some women endorse benevolent sexism as much as men (Glick et al., 2000). Moreover, recent theorizing has argued that women’s endorsement of benevolent sexism presents a barrier to feminist collective action (Radke, Hornsey, & Barlow, 2016). Future research should consider ways in which women’s levels of SDO could be reduced because this attitude may not only reinforce a benevolently sexist worldview that maintains women’s lower-status but may actively prevent women from participating in collective action for greater .

This article is protected by copyright. All rights reserved. SOCIAL DOMINANCE ORIENTATION AND SEXISM 16

References Abrams, D., Viki, G. T., Masser, B., & Bohner, G. (2003). Perceptions of stranger and acquaintance rape: The role of benevolent and hostile sexism in victim blame and rape proclivity. Journal of Personality and , 84, 111-125. doi.org/10.1037//0022-3514.84.1.111 Atkinson, J., Salmond, C., & Crampton, P. (2014). NZDep2013 Index of Deprivation. Department of Public Health, University of Otago, Wellington. Retrieved from http://www.otago.ac.nz/wellington/otago069936.pdf Brandt, M. J. (2013). Do the disadvantaged legitimize the social system? A large-scale test of the status-legitimacy hypothesis. Journal of Personality and Social Psychology, 104, 765-785. doi.org/10.1037/a0031751 Buhrmester, M., Kwang, T., & Gosling, S. D. (2011). Amazon's Mechanical Turk a new source of inexpensive, yet high-quality, data? Perspectives on Psychological Science,

6, 3-5. doi.org/10.1177/1745691610393980 Christopher, A. N., & Mull, M. S. (2006). Conservative ideology and ambivalent sexism. Psychology of Women Quarterly, 30, 223-230. doi.org/10.1111/j.1471- 6402.2006.00284.x Christopher, A. N., & Wojda, M. R. (2008). Social dominance orientation, right-wing authoritarianism, sexism, and prejudice toward women in the workforce. Psychology of Women Quarterly, 32, 65-73. doi.org/10.1111/j.1471-6402.2007.00407.x Festinger, L. (1957). A theory of cognitive dissonance. Stanford, CA: Stanford University Press. Fischer, A. R. (2006). Women's benevolent sexism as reaction to hostility. Psychology of Women Quarterly, 30, 410-416. doi.org/10.1111/j.1471-6402.2006.00316.x Fraser, G., Osborne, D., & Sibley, C. G. (2015). “We want you in the workplace, but only in a skirt!” Social dominance orientation, gender-based affirmative action and the moderating role of benevolent sexism. Sex Roles, 73, 231-244. doi.org/10.1007/s11199-015-0515-8 Furnham, A. (2003). Belief in a just world: Research progress over the past decade. Personality and Individual Differences, 34, 795-817. doi.org/10.1016/s0191- 8869(02)00072-7

This article is protected by copyright. All rights reserved. SOCIAL DOMINANCE ORIENTATION AND SEXISM 17

Glick, P., & Fiske, S. T. (1996). The Ambivalent Sexism Inventory: Differentiating hostile and benevolent sexism. Journal of Personality and Social Psychology, 70, 491-512. doi.org/10.1037//0022-3514.70.3.491 Glick, P., Fiske, S. T., Mladinic, A., Saiz, J. L., Abrams, D., Masser, B., . . . López, W. L. (2000). Beyond prejudice as simple antipathy: Hostile and benevolent sexism across cultures. Journal of Personality and Social Psychology, 79, 763-775. doi.org/10.1017/cbo9781139022736.005 Glick, P., & Whitehead, J. (2010). Hostility toward men and the perceived stability of male dominance. Social Psychology, 41, 177-185. doi.org/10.1027/1864-9335/a000025 Good, J. J., & Rudman, L. A. (2010). When female applicants meet sexist interviewers: The costs of being a target of benevolent sexism. Sex Roles, 62, 481-493. doi.org/10.1007/s11199-009-9685-6 Hayes, A. F. (2013). Introduction to mediation, moderation, and conditional process analysis: A regression-based approach. New York, NY: The Guildford Press. doi.org/10.1111/jedm.12050 Howell, D. C. (2009). Statistical methods for psychology (7th ed.). Belmont, CA: Wadsworth Cengage Learning. Huang, L. L., & Liu, J. H. (2005). Personality and social structural implications of the situational priming of social dominance orientation. Personality and Individual Differences, 38, 267-276. doi.org/10.1016/j.paid.2004.04.006 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, 881-919. doi.org/10.1111/j.1467-9221.2004.00402.x Jost, J. T., & Kay, A. C. (2005). Exposure to benevolent sexism and complementary gender : Consequences for specific and diffuse forms of system justification. Journal of Personality and Social Psychology, 88, 498-509. doi.org/10.2139/ssrn.386981 Jost, J. T., Pelham, B. W., Sheldon, O., & Sullivan, B. (2003). Social inequality and the reduction of ideological dissonance on behalf of the system: Evidence of enhanced system justification among the disadvantaged. European Journal of Social Psychology, 33, 13-36. doi.org/10.1002/ejsp.127

This article is protected by copyright. All rights reserved. SOCIAL DOMINANCE ORIENTATION AND SEXISM 18

Jost, J. T., & Thompson, E. P. (2000). Group-based dominance and opposition to equality as independent predictors of self-esteem, ethnocentrism, and social policy attitudes among African Americans and European Americans. Journal of Experimental Social Psychology, 36, 209-232. doi:10.1006/jesp.1999.1403 Kelly, A. J., Dubbs, S. L., & Barlow, F. K. (2014). Social dominance orientation predicts heterosexual men’s adverse reactions to romantic rejection. Archives of Sexual Behavior, 1-17. doi.org/10.1007/s10508-014-0348-5 Kteily, N., Ho, A. K., & Sidanius, J. (2012). Hierarchy in the mind: The predictive power of social dominance orientation across social contexts and domains. Journal of Experimental Social Psychology, 48, 543-549. doi.org/10.1016/j.jesp.2011.11.007 Lerner, M. J. (1980). The belief in a just world: A fundamental delusion. New York, NY: Plenum. Mosso, C., Briante, G., Aiello, A., & Russo, S. (2013). The role of legitimizing ideologies as predictors of ambivalent sexism in young people: Evidence from Italy and the USA. Social Justice Research, 26, 1-17. doi.org/10.1007/s11211-012-0172-9 Pehrson, S., Carvacho, H., & Sibley, C. G. (2017). Group differences in the legitimization of inequality: Questioning the role of social dominance orientation. British Journal of Social Psychology, 56, 28-46. doi.org/10.1111/bjso.12167 Pratto, F., Sidanius, J., & Levin, S. (2006). Social dominance theory and the dynamics of intergroup relations: Taking stock and looking forward. European Review of Social Psychology, 17, 271-320. doi.org/10.1080/10463280601055772 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. doi.org/10.1037//0022- 3514.67.4.741 Radke, H. R. M., Hornsey, M. J., & Barlow, F. K. (2016). Barriers to women engaging in collective action to overcome sexism. American Psychologist, 71, 893-874. doi.org/10.1037/a0040345 Sibley, C. G. (2015). Use of the New Zealand deprivation index in the NZAVS. NZAVS Technical Documents, e25. Retrieved from https://cdn.auckland.ac.nz/assets/psych/about/our- research/nzavs/NZAVSTechnicalDocuments/NZAVS-Technical-Documents-e25a- Geographic-and-area-unit-data.pdf

This article is protected by copyright. All rights reserved. SOCIAL DOMINANCE ORIENTATION AND SEXISM 19

Sibley, C. G., & Overall, N. C. (2011). A dual process motivational model of ambivalent sexism and gender differences in romantic partner preferences. Psychology of Women Quarterly, 35, 303-317. doi.org/10.1177/0361684311401838 Sidanius, J., Levin, S., Federico, C., & Pratto, F. (2001). Social dominance approach. In J. T. Jost & B. Major (Eds.), The psychology of legitimacy: Emerging perspectives on ideology, justice, and intergroup relations (pp. 307-322). New York, NY: Cambridge University Press. doi.org/10.2307/3094855 Sidanius, J., & Pratto, F. (1999). Social dominance theory: A new synthesis. In J. Sidanius & F. Pratto (Eds.), Social dominance (pp. 31-57). New York, NY: Cambridge University Press. doi.org/10.1017/cbo9781139175043.002 Statistics New Zealand. (2013). Geographic definitions. Retrieved from http://www.stats.govt.nz/Census/2013-census/info-about-2013-censusdata/2013- census-definitions-forms/definitions/geographic.aspx

This article is protected by copyright. All rights reserved. SOCIAL DOMINANCE ORIENTATION AND SEXISM 20

Table 1. Means, Standard Deviations, and Zero-Order Intercorrelations for Study 1.

Mean (SD) 1 2 3 4 5 6 7

1. Gender -0.25 (0.97) -

2. Age 49.65 (15.19) .14*** -

3. Race 0.74 (0.68) .00 .15*** -

4. SES 4.94 (2.82) -.02* -.07*** -.21*** -

5. SDO 2.43 (0.94) .20*** .13*** -.05*** -.00 -

6. Benevolent Sexism 3.88 (1.18) .11*** .08*** -.20*** .15*** .22*** -

7. Hostile Sexism 3.07 (1.23) .25*** .01 -.16*** .11*** .36*** .43*** -

Note. *p < .05, **p < .01, ***p < .001. Gender coded as male = 1, female = -1. Race coded as European New Zealander = 1, Non- European New Zealander = -1. SES = socio-economic status. Higher levels of SES indicate increased deprivation as the 2013 New Zealand Deprivation Index was used in this study.

This article is protected by copyright. All rights reserved. SOCIAL DOMINANCE ORIENTATION AND SEXISM 21

Table 2. Hierarchical Multiple Regression Predicting Benevolent Sexism as a Function of Participant Gender, SDO, and the Interaction between Participant Gender and SDO (Controlling for Hostile Sexism; Study 1).

Block 1 Block 2 Block 3

2 2 2 ² B BSE 95% CI sr ² B BSE 95% CI sr ² B BSE 95% CI sr

Hostile .43*** .41 .01 .39, .42 .18 .40*** .38 .01 .36, .40 .13 .40*** .38 .01 .36, .40 .13 Sexism

Gender -.01 -.01 .01 -.03, .02 .00 .00 .00 .01 -.02, .02 .00

SDO .08*** .10 .01 .07, .12 .01 .07*** .08 .01 .06, .11 .00

Gender -.06*** -.08 .01 -.10, -.06 .00 x SDO

2 R ch .18 .01 .00

2 R adj .18 .19 .19

Fch 2,311.89*** 32.19*** 46.12***

2 ƒ .22 .23 .24

Note. ***p < .001. Gender coded as male = 1, female = -1. 95% CI = 95% confidence intervals.

This article is protected by copyright. All rights reserved. SOCIAL DOMINANCE ORIENTATION AND SEXISM 22

Table 3. Hierarchical Multiple Regression Predicting Hostile sexism as a Function of Participant Gender, SDO, and the Interaction between Participant Gender and SDO (Controlling for Benevolent Sexism; Study 1).

Block 1 Block 2 Block 3

2 2 2 ² B BSE 95% CI sr ² B BSE 95% CI sr ² B BSE 95% CI sr

Benevolent .43*** .45 .01 .43, .47 0.18 .35*** .37 .01 .35, .39 .12 .35*** .37 .01 .35, .39 .12 Sexism

Gender .17*** .21 .01 .19, .23 .03 .17*** .21 .01 .19, .23 .03

SDO .25*** .33 .01 .30, .35 .06 .25*** .32 .01 .30, .35 .06

Gender -.00 -.01 .01 -.03, .02 .00 x SDO

2 R ch .18 .10 .00

2 R adj .18 .28 .28

Fch 2,311.89*** 725.05*** .25

.22 .39 .39 ƒ2

Note. ***p < .001. Gender coded as male = 1, female = -1. 95% CI = 95% confidence intervals.

This article is protected by copyright. All rights reserved. SOCIAL DOMINANCE ORIENTATION AND SEXISM 23

Table 4. Means, Standard Deviations, and Zero-Order Intercorrelations for Study 2.

Mean (SD) 1 2 3 4 5 6 7 8

1. Age 33.65 (11.47) -

2. Race 0.63 (0.78) .13* -

3. SES 3.61 (1.20) -.03 .05 -

4. SDO 2.16 (1.02) .02 .07 .23*** -

5. Legitimacy of the 3.25 (1.20) .10 .09 .15* .54*** - gender hierarchy

6. Personal need for 2.93 (1.71) .10 -.07 .21*** .50*** .51*** - protection

7. Benevolent sexism 3.23 (1.24) .04 -.08 .19** .51*** .57*** .61*** -

8. Hostile sexism 2.97 (1.30) .06 .06 .18** .60*** .76*** .52*** .63*** -

Note. *p < .05, **p < .01, ***p < .001. Race coded as White = 1, Non-White = -1. SES = socio-economic status.

This article is protected by copyright. All rights reserved. SOCIAL DOMINANCE ORIENTATION AND SEXISM 24

List of Figures

Figure 1.

The proposed model linking SDO and Benevolent Sexism (controlling for hostile sexism; Study 2).

Note. ² displayed. Direct effect in brackets. *p < .05, **p < .01, ***p < .001

This article is protected by copyright. All rights reserved.