Running Head: Essentialist Explanations for Gender Differences The Effects of System Justifying Motives on Endorsement of Essentialist Explanations for Gender Differences

People have a fundamental motive to view their social system as just, fair, and good and will engage in a number of strategies to rationalize the status quo (Jost & Banaji, 1994). We propose that one way in which individuals may ̳justify the system‘ is through endorsement of essentialist explanations, which attribute group differences to deep, essential causes. We suggest that system justifying motives lead to greater endorsement of essentialist explanations because those explanations portray group differences as immutable. Study 1 employs an established system threat manipulation and finds that activating system justifying motives increases both male and female participants‘ endorsement of essentialist explanations for gender differences, and that this effect is mediated by beliefs in immutability. Study 2 uses a goal contagion manipulation and finds that both male and female participants primed with a system-justifying goal are significantly more likely to agree with essentialist explanations for gender differences. Study 3 demonstrates that providing an opportunity to explicitly reject a system threat (an alternative means of satisfying the goal to defend the system) attenuates system threat effects on endorsement of essentialist explanations, further suggesting that this process is motivated. Finally, Studies 4a and 4b dissociate the type of cause (biological vs. social) from whether group differences are portrayed as mutable vs. immutable and find that system threat increases endorsement of immutable explanations, independent of the type of cause. Essentialist Explanations for Gender Differences 3 ―Biology has always been used as a curse against women. From Darwin to Desmond Morris, from Freud to Robin Fox, from animal behaviorists who consider themselves open minded but „realistic‟ to the sober professors of ethology, the message has rarely changed: men are biologically suited for their life of power, pleasure, privilege, and women must accept subordination, sacrifice, and submission‖ (Weisstein, 1982, p. 41). Scholars from a variety of fields have amassed examples of how, throughout history, essentialist explanations for gender differences have been used to justify women‘s subordination and thereby perpetuate the status quo (e.g., Bem, 1993; Fausto-Sterling, 1985; Russett, 1989). For example, when American women first began to attend colleges and universities in the 19th century, prominent education leaders of the day argued that higher education would harm women‘s reproductive capacity because it would divert women‘s energy from the development of their reproductive organs to the development of their brains (Clarke, 1873). Likewise, many other scholars since then, such as Becker (1973) and Tiger (1970), argued that the sex-based division of labor was biologically ordained. In short, women, being ―naturally‖ more nurturing than men, were unsuited for roles outside the home. The present research seeks to understand whether and how desires to view one‘s social system as just, fair, and good (i.e., system justification motives; Jost & Banaji, 1994) are causally related to endorsement of such essentialist explanations for group differences (e.g., Haslam & Whelan, 2008; Haslam, Rothschild & Ernst, 2000; Keller, 2005). Specifically, we investigate how a motivation to justify the social system may lead people to endorse explanations that portray gender differences as stemming from deep, essential causes. In so doing, we identify a novel consequence of system justification, and more broadly, the factors that lead people toward particular explanatory frameworks, such as essentialism. Essentialist Explanations for Gender Differences 4 Psychological essentialism Psychological essentialism is the tendency to assume that category members possess a deep, underlying ―essence‖ that is responsible for phenomenal properties and category membership (e.g., Atran, 1987; Gelman, 2003; Hamilton, Sherman, & Rodgers, 2004; Hirschfeld & Gelman, 1994; Keil, 1989; Medin & Ortony, 1989; Newman & Keil, 2008; Plaks, Levy, Dweck, & Stroessner, 2004; Yzerbyt, Corneille, & Estrada, 2001). For example, people may assume that there is an essence of tigers that is common to all tigers, and causes tigers to have their unique tiger properties. The concept of essentialism has a long intellectual history (see Sober, 1994) and has been defined in diverse, albeit overlapping, ways (see Haslam, Rothschild & Ernst, 2004 for a review). For example, essentialism as it applies to natural kinds and certain socially-relevant categories (such as race and gender) is a multi-faceted phenomenon, including such elements as having high inductive potential, sharp category boundaries, and being homogeneous, identity-determining, historically invariant, and immutable (Gelman, 2003; Haslam et al., 2004). In some cases, psychological essentialism implies biological causation (Bem, 1993; Haslam & Whelan, 2008). However, it need not, as people also seem to assume the existence of essences (e.g., the essence of being a scientist) even in cases where a biological explanation is not possible (Knobe, Prasada & Newman, 2013). Evidence for psychological essentialism has been found in a diverse range of populations throughout the world (Gil-White, 2001; Norenzayan & Heine, 2005; Sousa, Atran, & Medin, 2002; Waxman, Medin, & Ross, 2007) and in children as young as three or four (Gelman, 2003), leading some researchers to propose that psychological essentialism is a human universal that is rooted in the basic cognitive architecture of the mind (Dar-Nimrod & Heine, 2011; Bloom 2004; 2010; Gelman, 2003; Keil, 1989). Importantly, however, people tend to think of essences in Essentialist Explanations for Gender Differences 5 terms of an ―essence placeholder‖ (Medin & Ortony, 1989) rather than an actual physical entity, which means that folk conceptions of essence tend to be relatively ―fuzzy.‖ For example, young children readily appeal to essentialist explanations for differences between natural kinds (Newman & Keil, 2008) as well as socially-relevant differences such gender (Gelman, 2003) well before they possess firm commitments about what those essences might be. Moreover, a great deal of evidence suggests that lay understanding of the influence of biological factors on human behavior is incomplete (e.g., Alper & Beckwith, 1993; Hinshaw & Stier, 2008; Shostak, Freese, Link & Phelan, 2009). For example, one study found that over half the population did not know that genes are located in cells (Lanie, et al., 2004). Compounding this problem is the fact that people receive most of their information about science from the media (Conrad, 1997) and those descriptions tend to be grossly oversimplified (Conrad, 1997; Bubela & Caulfield, 2004), deterministic (Alper & Beckwith, 1993), and subject to political biases (Brescoll & LaFrance, 2004). Taken together, past research suggests that we are predisposed toward essentialist explanations (Gelman, 2003). However, those explanations may function more as conceptual placeholders and may lead to causal inferences that are incomplete or overly simplified. As a result, while the basic structure of essentialist explanations may be an inevitable byproduct of the mind, the ―deployment‖ of essentialist explanations may be highly susceptible to the influence of motivated reasoning based on differences in political ideology (Brescoll & LaFrance, 2004) or, as we argue here, system justification (Jost & Hamilton, 2005; Yzerbyt, Rocher & Schadron, 1997). Essentialist Explanations for Gender Differences 6 Essentialism as a system justifying ideology The existing literature suggests that essentialist explanations for group differences may typically imply two distinct, but related ideas (Brescoll & LaFrance, 2004): first, that the causes of gender differences are biological, and second, that they are fixed and not capable of being changed. (Bem, 1993; Bleier, 1984; Fausto-Sterling, 1985; Russett, 1989; Sayers, 1982). For example, if an individual believes that men have a higher aptitude in math than women because of innate, biological factors, then this implies not only that the difference is genetic, but also that the lack of women pursuing math-related careers is an inevitable ―fact‖ that is unlikely to change over time (Bem, 1993). Therefore, in the present studies, we conceive of biological attributions and immutability as two distinct but related features of a more general tendency to attribute group differences to deep, essential causes (i.e., essentialism). In fact, we propose that it is this latter ―immutable‖ aspect of essentialism (and not biology per se) that makes essentialist explanations for group differences an especially potent means of satisfying system justifying motives. System justification theory proposes that people have a fundamental motive to view their social system as just, fair, and good and will engage in a number of strategies to rationalize the status quo (Jost & Banaji, 1994; Jost, Banaji, & Nosek, 2004; see also Sidanius & Pratto, 1999). Rationalization of the status quo is aided by psychological processes of motivated reasoning (Kunda, 1990). Indeed, people are extraordinarily adept at adopting ideas, evidence, and standards that allow them to draw the conclusions they would like to reach (Dunning & Cohen, 1992; Dunning, Leuenberger, & Sherman, 1995; Kunda, 1987; Norton, Vandello, & Darley, 2004; Simon, 2004; Simon, Krawczyk, Bleicher, & Holyoak, 2008; Uhlmann & Cohen, 2005; Uhlmann, Pizarro, Tannenbaum, & Ditto, 2009). Essentialist Explanations for Gender Differences 7 One way of justifying the system and existing status hierarchies, such as those between men and women, may be to explain group differences as immutable. For example, if one perceives the existing social structure as an inevitable ―fact‖ that is unlikely to change over time, then it limits the criticisms that can be made of the status quo, and lends support to the current system. In other words, essentialist explanations imply that existing social structures (such as differences between social groups) are inherent, not likely to change, and therefore ―right,‖ which can be used as a means of reaffirming the status quo. Therefore, we propose that motivations to justify the system will lead people to seek out essentialist explanations for gender differences because such explanations portray the system as stable and incapable of being changed. Indeed, the existing literature provides some support for the idea that essentialism is a system justifying ideology as it has been shown to correlate with a number of beliefs typically associated with system justification (e.g., sexism, racism, social dominance, and authoritarianism; Jayaratne et al., 2006; Keller, 2005; Sidanius & Pratto, 1999). For example, Keller (2005) found that belief in genetic determinism is correlated with conservatism, nationalism, negative racial stereotyping, modern sexism, and measures of prejudice. Analogously, using a ―Biological Basis of Essentialism Scale,‖ Bastian and Haslam (2006) found a relationship between endorsing items such as ―the kind of person someone is can be largely attributed to their genetic inheritance‖ and the extent to which participants agreed with stereotypes about a range of social groups. Experimental studies further show that exposing people to essentialist explanations for group differences causes them to more strongly endorse other ideologies that uphold the social system. For example, Brescoll and LaFrance (2004) assigned participants to read a fictional Essentialist Explanations for Gender Differences 8 newspaper article about gender differences in plant identification that was said to be caused either by biological or socio-cultural factors. Participants exposed to the biological explanation were more likely to endorse a range of gender stereotypes compared to participants exposed to the socio-cultural explanation. Keller (2005) found that having students read an essay promoting genetic explanations caused them to report stronger in-group biases compared to reading a neutral essay. And, Williams and Eberhardt (2008) found that participants who read that there was a biological basis for race were less interested in becoming friends with a student who was a different race from their own. In sum, the existing literature suggests that essentialism may serve as a system justifying ideology. However, past research has not demonstrated a causal effect of system justifying motives on essentialist explanations for group differences, nor has the theoretical mechanism linking these two phenomena been explored. The present research attempts to fill this gap by demonstrating that experimentally activating motivations to justify the system can cause individuals to endorse essentialist explanations for gender differences. In the present studies, we predominantly operationalize essentialist explanations as ̳biological essentialism‘ (Keller, 2005) because this is the most common form of essentialist explanations for gender differences (Bem, 1993; Brescoll, 2005). However, we propose that perceived immutability provides the critical link between system justification and endorsement of essentialist explanations. Therefore, in later studies (Studies 4a and 4b) we show that activating system justifying motives lead individuals to endorse biological explanations for gender differences if and only if those influences are portrayed as immutable. A further point concerns the extent to which the effect of system justification on essentialist explanations reflects motivated reasoning. Although there is of course difficulty Essentialist Explanations for Gender Differences 9 inherent in distinguishing motivated reasoning from a ―non-motivational, pure cognitive‖ effect (Aarts, Gollwitzer, & Hassin, 2004; Greenwald & Ronis; 1978; Tetlock, 1984), here we attempt to distinguish these concepts as follows: A non-motivational account may attempt to explain an apparent effect of system justifying motives on essentialist explanations via the activation of semantic networks or some associative link. By contrast, a motivational explanation instead conceives of essentialist explanations for gender differences as satisfying some goal or psychic need (e.g., the desire to reaffirm or uphold the system). For example, experiencing a threat to the existing status quo (Jost, Kivetz, Rubini, Guermandi & Masso, 2005; Kay, Jost & Young, 2005) should lead individuals to seek out ways of defending the status quo and in a sense, ―repair‖ the effects of that threat. This distinction is useful as it makes different predictions regarding how activating system justification should influence essentialism, as well as certain moderating conditions. Only a motivated account uniquely predicts that the activation of system justifying motives via the adoption of goals should lead to increases in essentialist explanations (a possibility we test in Study 2). Further, a motivational account predicts that providing alternative means to satisfy the goal to support the system (such as the opportunity to explicitly reject the source of the system threat) should diminish the effects of the threat on endorsement of essentialist explanations (examined in Study 3). A final prediction is that the effects of system justifying motives should generalize across both male (advantaged) and female (disadvantaged) individuals (Jost & Banaji, 1994; Kay et al., 2009; Jost, Pelham, Sheldon, & Sullivan, 2003). SJT proposes that even people who do not stand to benefit from endorsing the status quo (e.g., women endorsing beliefs that reify gender inequality or poor people supporting policies that do not favor income redistribution) often will Essentialist Explanations for Gender Differences 10 do so because everyone has the same basic motive to uphold the existing social system. Consistent with this idea, women are just as likely as men to exhibit backlash against women who promote their abilities (Rudman, 1998; Rudman & Glick, 1999), women who speak ―too much‖ in organizational settings (Brescoll, 2011), and women who express anger in the workplace (Brescoll & Uhlmann, 2008). Therefore, we examine the link between system justification and essentialist explanations for gender differences in both men and women, with the expectation that desires to uphold the system should have equivalent effects, regardless of participant gender. Overview Five empirical studies examined the causal link between system justifying motives and endorsement of essentialist explanations for gender differences. Study 1 demonstrates that system threat leads both men and women to endorse essentialist explanations for gender differences. Additionally, Study 1 offers evidence for our proposed mechanism by demonstrating that perceived immutability mediates the relationship between system justification and endorsement of essentialist explanations. Study 2 provides evidence of a motivated process by using a goal contagion prime to show that activating a pro-system goal can likewise cause participants to endorse essentialist explanations for gender differences. Study 3 demonstrates that providing an opportunity to explicitly reject the system threat, and thus defend the system by an alternative means (Cutright, Wu, Banfield, Kay and Fitzsimons, 2011) attenuates the system threat effect on endorsement of essentialist explanations, lending further support to a motivational account. Finally, Studies 4a and 4b provide a further test of the proposed mechanism by separating the type of cause for gender differences (biological vs. social) from whether the influence is described as mutable or immutable. Results from these experiments Essentialist Explanations for Gender Differences 11 indicate that system threats lead to an increase in endorsement of immutable explanations, independent of the type of cause, thereby providing additional empirical support for the notion that perceived immutability is the key mechanism linking system justification to essentialist explanations. Study 1 Study 1 used an established manipulation to activate system-justifying motives (Jost et al., 2005; Kay et al., 2005). In this manipulation, participants read an ostensive news story reporting either that the U.S. had reached a social, economic and political nadir (system threat) or that the country was doing well in these same areas (system affirmation). Consistent with the predictions outlined above, we expected that reading information that threatened the social status quo, as compared to affirming it, would lead both male and female participants to endorse essentialist explanations for gender differences. In the present study, we operationalized essentialist explanations as ̳biological essentialism‘ (Keller, 2005). This study also included a neutral control condition and an existential threat condition (McGregor, Zanna, Holmes, & Spencer, 2001). The control condition tested the direction of the effect—i.e., whether system threat increases endorsement of essentialist explanations, or whether system affirmation causes participants to be less likely to endorse essentialist explanations. The existential threat manipulation allowed us to test whether it is experiencing a threat to the system specifically, and not just experiencing any type of threat that causes people to endorse biological essentialist explanations for gender differences. The second goal of Study 1 was to test the proposed mechanism of immutability. Our theory proposes that system justifying motives increases endorsement of essentialist explanations because they imply that the status quo is immutable. Therefore, in addition to the main Essentialist Explanations for Gender Differences 12 dependent measure assessing endorsement of biological essentialism for gender differences, we also included a series of items which specifically asked about immutability, with the hypothesis that perceived immutability would mediate the effect of system threat on endorsement of biological essentialism. Method Participants and Design One hundred thirty five (53 males, 80 females; 2 did not indicate their gender) adults (Mage = 36.6) were randomly assigned to the system threat condition, the system affirmation condition, the existential threat condition, or a control condition. Participants were recruited from a national online database maintained by an East Coast university. Materials and Procedure Experimental conditions. Participants in the system threat and affirmation conditions first read one of two ostensive news stories, which were presented as part of a separate study on human memory (Jost et al., 2005; Kay et al., 2005). Prior work (Kay et al., 2005) established that this manipulation threatens/affirms the perceived stability of the social system without affecting indices of threat to the self. The system threat manipulation read as follows: ―These days, many people in the United States feel disappointed with the nation‟s condition. Many citizens feel that the country has reach a low point in terms of social, economic and political factors. It seems that many countries are enjoying better social, economic and political conditions than the U.S. More and more Americans express a willingness to leave the United States and emigrate to other nations.‖ The system affirmation manipulation read as follows: ―These days, despite the difficulties the nation is facing, many people in the United States feel safer and more secure relative to the past. Many citizens feel that the country is relatively stable in terms of social, economic and political factors. It seems that compared with many countries in the world the social, economic and political conditions Essentialist Explanations for Gender Differences 13 in the U.S. are relatively good. Very few Americans express a willingness to leave the United States and emigrate to other nations.‖ Participants then completed a manipulation check masked as a ―memory quiz‖ asking them ―How did the article report that the United States is doing relative to other countries?‖ and, ―According to the article, how do most Americans feel about the condition of the United States?‖ Four participants (2.8%) did not pass the manipulation check and thus were not included in the following analyses. However, including (or excluding) their data did not alter the results. In the existential threat condition, we utilized a manipulation from McGregor et al. (2001) where participants were first asked to ―briefly describe the events, people, and location associated with an important, vivid memory from your childhood or adolescence.‖ After this, participants were then asked to: ―(w)rite down how you imagine the scene of this above memory might be changed if you revisited it in the year 2035 (be as specific as possible). How does it make you feel to imagine this?‖ All participants wrote between 2-5 sentences for each question and an examination of their responses found that nearly all participants mentioned that they were unsure if they would view things the same in the future, suggesting that the subjects correctly engaged with this manipulation of uncertainty. In the control condition, participants simply completed the dependent measures. Dependent Measures Biological Essentialism. All participants completed seven items adapted from Keller‘s (2005) biological essentialism measure reworded to refer to gender differences (1 = completely disagree, 11 = completely agree). The items included, ―I think that differences between men and women in personality are largely determined by genetic factors,‖ ―I believe that men pursue math and science careers more than women because of the innate difference between the Essentialist Explanations for Gender Differences 14 genders,‖ ―I think that differences between men and women in behavior are largely determined by the biological differences between the genders,‖ ―Men commit the majority of violent crimes in this country because they have a greater predisposition towards violence than women,‖ ―Part of the reason why women are more emotional than men is because of the way they‟re hardwired,‖ ―Women get in fewer physical fights than men because men have less of an inborn tendency for aggression,‖ and ―I think that the reason why there are more male math professors than female math professors is due to some biological differences between the sexes.‖ We intentionally included items that would assess not only biological attributions for gender differences in general, but also for specific behaviors (e.g., violent behavior, math performance) that people commonly believe that men and women differ on. Basic Beliefs About Gender Differences. To assess more clear-cut beliefs about gender differences (which should not vary as a function of system threat because they do not have implications for justifying the gender hierarchy), we utilized a measure by McCauley and Thangavelu (1991). Participants separately estimated the height for ―an adult American female‖ and an ―adult American male.‖ They also estimated the average salary for an ―adult American male per year,‖ and did the same for American females. Next, participants completed a task where they estimated the percentage of women in certain occupations in the United States (McCauley & Thangavelu, 1991). They estimated the percentages of females who are police officers, school teachers, lawyers, engineers, bank tellers, physicians, secretaries, scientists, bus drivers, and social workers. These occupations were selected not only because they are familiar to most participants, but because some are more common among one gender or the other and some are gender-balanced (McCauley & Thangavelu, 1991). Essentialist Explanations for Gender Differences 15 Immutability. To assess beliefs about the immutability of gender differences, we adapted items from Levy, Stroessner and Dweck‘s (1998) measure assessing implicit person theory beliefs applied to groups. We added two reverse-scored items, and re-wrote the items so that they would assess immutability/mutability and not biological explanations for gender differences. Participants completed the following items on a 6-point scale (1 = strongly disagree, 6 = strongly agree): ―The differences in behavior between men and women can't be changed very much,‖ ―Individuals can do things differently, but the important differences between men and women can't really be changed,‖ ―Even the most basic differences between men and women can be easily changed,‖ (r) ―The basic differences between how men and women act are fixed, and nothing can really change that,‖ and ―Men and women can easily become more like one another‖ (r). 1 Mood, Arousal and Uncertainty. To assess mood, arousal and uncertainty, we asked participants to rate how they were feeling in the moment, using a 9-point scale (1 = not at all, 9 = extremely). Following McGregor and colleagues (2001), we included three items to measure felt uncertainty (―bothered,‖ ―uncomfortable,‖ ―uneasy‖). Elliot and Devine (1994) found that these items were associated with cognitive inconsistency in dissonance experiments. To assess positive and negative mood, we used six items from Diener and Emmons‘s scale (1984): joyful, happy, pleasant, bored, angry, and unpleasant. Finally, to assess psychological arousal, we included three items: awake, aroused, and sleepy (r) (Diener & Emmons, 1984). 1 We performed a factor analysis on the items in the immutability and biological essentialism scales. As expected, inspection of the scree plot and Varimax factor analysis revealed two distinct factors: immutability and biological essentialism. All items had factor loadings over .60 and, taken together, these two factors accounted for 67% of the total variance. Essentialist Explanations for Gender Differences 16 Demographic information. In all of the present studies, participants reported their degree of political conservatism (1=very liberal, 7=very conservative), as well as their gender, age, race/ethnicity, and country of origin. Funneled debriefing. Finally, in this and subsequent studies in this paper, participants were administered a funneled debriefing probing for their awareness of the influence of the experimental manipulations (Bargh & Chartrand, 2000). Participants in the system threat and system affirmation conditions were asked: ―Did the news story you read influence your social attitudes in any way?‖ (1=Definitely Not, 9=Definitely Yes). If the answered affirmatively, participants were asked to describe the influence of the news story. Participants in the existential threat condition were also administered a funneled debriefing but with slightly different wording: ―Did the paragraph you wrote about a childhood or adolescent memory influence any of your other answers in this survey?‖ (1=Definitely Not, 9=Definitely Yes). Participants were then asked to explain how and why writing about this experience may have influenced their responses on the rest of the survey. Nearly all participants (93%) did not believe the manipulation had affected their subsequent responses and among the minority that thought it may have influenced them, none of these participants were able to accurately describe how it may have done so. Similar funneled debriefings in the remaining studies likewise revealed no significant evidence of awareness of the influence of the system justification manipulations among our research participants. Results Biological Essentialism. A 2 (participant gender) x 4 (system threat vs. affirmation vs. existential threat vs. control) ANOVA indicated a significant main effect of condition, F (3, 125) = 3.37, p < .05, no main effect of participant gender, F (1, 125) < 1, ns, and no interaction Essentialist Explanations for Gender Differences 17 between condition and participant gender, F (1, 125) < 1, ns (Figure 1; see Table 1 for correlations between all dependent measures). A planned contrast analysis revealed that participants in the system threat condition (M = 7.26, SD = 2.17) were significantly more likely to endorse biological essentialist beliefs about gender than participants in the affirmation condition (M = 5.97, SD = 2.13), the existential threat condition (M = 5.88, SD = 1.61), and the control condition (M = 5.93, SD = 2.00), t(131) = 3.42, p < .01. In addition, there were no significant differences between the affirmation, control, and existential threat conditions on the essentialism scale, all Fs < 1, ns. But, each of the three conditions were significantly different from the system threat condition (system threat vs. affirmation, t(73) = 2.61, p < .05; system threat vs. existential threat, t(59) = 2.69, p < .01; and system threat vs. control, t(71) = 2.74, p < .001). Thus, it appears as though experiencing a threat to the social order, and not an affirmation of the social order or an existential threat, specifically caused participants to increase their endorsement of biological essentialism. Additional analyses examining participant gender separately revealed that men and women had nearly identical patterns of responding on the dependent measure. A planned contrast analysis revealed that male participants were more likely to endorse biological essentialism in the system threat condition (M = 7.69, SD = 2.05) than the affirmation condition (M = 5.90, SD = 2.01), the existential threat condition (M = 6.24, SD = 1.39), and the control condition (M = 5.67, SD = 2.09), t (49) = 2.09, p < .05. Female participants were also more likely to endorse biological essentialism in the system threat condition (M = 7.18, SD = 2.22) compared to the system affirmation condition (M = 5.82, SD = 2.27), the existential threat condition (M = 5.56, SD = 1.77), and the control condition (M = 5.94, SD = 1.94), t (76) = 2.91, p < .01. Thus, regardless of participants‘ own gender, being exposed to information that Essentialist Explanations for Gender Differences 18 threatened the established social order led to increased endorsement of biological essentialism compared to reading information affirming the social order, being exposed to an existential threat to oneself, or reading nothing at all. Basic Beliefs About Gender Differences. A 2 (participant gender) x 4 (system threat vs. affirmation vs. existential threat vs. control) MANOVA on the male and female height and salary estimations, indicated no main effect for condition, F (12, 323) = 1.51, ns, Wilk's λ = 0.86, partial ε2 = .047, a significant main effect of participant gender, F (4, 122) = 3.21, p < .05, Wilk's λ = 0.91, partial ε2 = .095, and no interaction between condition and participant gender, F (12, 323) = .44, ns, Wilk's λ = 0.96, partial ε2 = .014. Regardless of condition, female participants estimated the average American male to be taller than male participants‘ estimates (M = 69.95 inches, SD = 1.60; M = 68.93 inches, SD = 2.42, respectively), F (1, 125) = 5.59, p < .05. Female participants also estimated that the average American male earned a higher yearly salary than male participants did (M = $50,217, SD = $16,221; M = $44,327, SD = $17,296, respectively) F (1, 125) = 5.12, p < .05, again regardless of what condition the female participants were assigned to. For the estimations of the percentage of women in certain occupations in the United States, a 2 (participant gender) x 4 (system threat vs. affirmation vs. existential threat vs. control) MANOVA revealed no main effect for condition, F (30, 350) = 1.06, ns, Wilk's λ = 0.78, partial ε2 = .08, no main effect of participant gender, F (10, 119) = 1.39, ns, Wilk's λ = 0.895, partial ε2 = .105, and no interaction between condition and participant gender, F (30, 350) = .75, ns, Wilk's λ = 0.834, partial ε2 = .105. Thus, as expected, condition type (system threat vs. affirmation vs. existential threat vs. control) did not appear to influence either set of measures assessing basic beliefs about gender. Essentialist Explanations for Gender Differences 19 Mood, Arousal and Uncertainty. We computed separate indices of arousal (α = .80), uncertainty (α = .95), positive affect (α = .81), and negative affect (α = .76). Examining the effects of condition on psychological arousal, a 2 (participant gender) x 4 (system threat vs. affirmation vs. existential threat vs. control) ANOVA found no main effect for condition, F (3, 125) = 1.28, ns, no main effect for gender, F (1, 125) = 1.50, ns, and no interaction between condition and participant gender, F (3, 125) = .47, ns. Similarly, a 2 (participant gender) x 4 (system threat vs. affirmation vs. existential threat vs. control) ANOVA revealed no main effect for condition, F (3, 125) = 1.50, ns, no main effect for gender, F (1, 125) = .67, ns, and no interaction between condition and participant gender, F (3, 125) = .78, ns, on positive mood. For negative mood, a 2 (participant gender) x 4 (system threat vs. affirmation vs. existential threat vs. control) ANOVA revealed no main effect for condition, F (3, 125) = .56, ns, no main effect for gender, F (1, 125) = .08, ns, and no interaction between condition and participant gender, F (3, 125) = .29, ns. These results suggest that the system threat, system affirmation, and existential threat manipulations did not significantly impact participants‘ positive mood, negative mood, or subjective sense of psychological arousal. The items assessing felt uncertainty were averaged to form a scale (α = .95). A 2 (participant gender) x 4 (system threat vs. affirmation vs. existential threat vs. control) ANOVA revealed the expected main effect for condition, F (3, 125) = 2.47, p = .06, no main effect for participant gender, F (1, 125) = 1.59, ns, and no interaction between condition and participant gender, F (3, 125) = .86, ns. As expected, participants reported feeling more uncertainty in the threat conditions (existential threat, M = 3.40, SD = 1.95; system threat, M = 3.31, SD = 2.26) compared to the affirmation (M = 2.57, SD = 1.80) and control conditions (M = 2.65, SD = 1.87), t (130) = 2.14, p < .05. Essentialist Explanations for Gender Differences 20 Immutability. The items assessing participants‘ beliefs about the immutability formed a reliable scale (Chronbach‘s α = .87). A (participant gender) x 4 (system threat vs. affirmation vs. existential threat vs. control) ANOVA revealed the hypothesized main effect for condition, F (3, 125) = 3.63, p < .05, no main effect of participant gender, F (1, 125) = .09, ns, and no interaction between condition and participant gender, F (3, 125) = .83, ns. A planned contrast analysis revealed that participants in the system threat condition (M = 3.39, SD = .90) were significantly more likely to believe that gender differences are immutable than participants in the affirmation condition (M = 3.36, SD = 1.04), the existential threat condition (M = 3.37, SD = .78), and the control condition (M = 3.28, SD = .86), t(131) = 3.07, p < .01. Furthermore, there were no significant differences between the affirmation, control, and existential threat conditions on the immutability scale, all F‘s < 1, ns. However, each of the three conditions were significantly different from the system threat condition (system threat vs. affirmation, t(73) = 2.34, p < .05; system threat vs. existential threat, t(59) = 2.27, p < .05; and system threat vs. control, t(71) = 2.93, p < .001). Thus, it appears that when participants experience a threat to their system, they are more likely to endorse the belief that gender differences are immutable than participants exposed to an existential threat, an affirmation of the system or no manipulation at all. Examining male and female participants separately revealed that they had similar patterns of responding on the immutability measure. A planned contrast analysis revealed that male participants were more likely to believe that gender differences were immutable in the system threat condition (M = 4.27, SD = .63) than the affirmation condition (M = 3.20, SD = 1.01), existential threat condition (M = 3.25, SD = .73), or control condition (M = 3.16, SD = .80), t (49) = 2.77, p < .01. Female participants were also more likely to believe that gender differences were immutable in the system threat condition (M = 3.81, SD = .94) compared to the system Essentialist Explanations for Gender Differences 21 affirmation condition (M = 3.52, SD = 1.06), existential threat condition (M = 3.47, SD = .83), or control condition (M = 3.30, SD = .89), t (76) = 1.77, p = .08. Thus, it appears that both male and female participants were similarly impacted by exposure to a system threat in they were both more likely to believe that gender differences were immutable compared to male and female participants exposed to an existential threat, an affirmation of the system, or not exposed to anything at all. Mediation analysis. A series of regression analyses tested our prediction that beliefs about immutability would help to explain why threats to the system cause people to endorse biological essentialist explanations for gender differences. Using the neutral control condition as the comparison (and combining male and female subjects), exposure to a system threat was significantly related to immutability, ß (48) = .30, p < .05, and endorsement of biological essentialism, ß (48) = .27, p = .05. Furthermore, beliefs about immutability were related to biological essentialism, ß (48) = .68, p < .001. When we used exposure to the system threat and beliefs about immutability as independent variables to simultaneously predict endorsement of biological essentialism, only the coefficient for beliefs in immutability remained significant, ß (47) = .66, p < .001. The coefficient for system threat fell from ß = .27 to ß = .07 and was no longer significant, Sobel z = 2.08, p < .05. Thus, it appears that beliefs about immutability mediated the relationship between system threat and endorsement of biological essentialism for gender differences. Discussion The results of this study were consistent with the prediction that that a motive to justify the system can lead people to endorse essentialist explanations for gender differences, regardless of the participants‘ gender. Prior research has established correlational relationships between Essentialist Explanations for Gender Differences 22 these constructs (Keller, 2005; Sidanius & Pratto, 1999), and has further shown that exposure to biological attributions leads to increased endorsement of gender stereotypes (Brescoll & LaFrance, 2004; Keller, 2005; Williams & Eberhardt, 2008). However, to our knowledge, our study is the first to demonstrate that experimentally activating system-justifying motives can actually cause both men and women to endorse essentialist explanations for gender differences. Specifically, we found that when people were exposed to a system threat, they were more likely to endorse biological essentialism compared to those exposed to a system affirmation, an existential threat, or no message of any kind (a neutral control condition). This demonstrates that it is system threat that causes increased endorsement of essentialist explanations for gender differences, rather than system affirmation that causes decreased endorsement of essentialist ideologies. Furthermore, Study 1 also shows that it is not just any type of threat that causes an increased endorsement of essentialism as exposing participants to a commonly used manipulation of existential threat (McGregor, et al., 2001) did not increase participants‘ agreement with essentialist explanations for gender differences. Study 1 also found that exposure to a system threat manipulation did not cause an increase in negative mood or psychological arousal. Although Kay et al. (2005) have demonstrated that the system threat manipulation does not induce a threat to the self, it is nevertheless possible that if a threat manipulation causes a negative mood state or elevated level of psychological arousal, it could influence responses on the essentialism scale. However, as expected, we did not find this to be the case, supporting the idea that negative mood or psychological arousal do not explain why experiencing a threat to the system leads to increased agreement with essentialist explanations for gender differences. Essentialist Explanations for Gender Differences 23 Study 1 also examined a number of what we call ―basic‖ beliefs about gender whose endorsement (or lack of endorsement) would not have direct implications for justifying gender inequality (e.g., estimating the average height of men and women). As expected, we did not find that exposure to a threat to the system impacted these types of beliefs about gender whereas we did find that exposure to a system threat increased participants‘ endorsement of essentialist explanations for gender differences. Together, this suggests that experiencing a threat to the system activates a motive to restore it and thereby ̳set things right‘ by endorsing ideologies which reinforce the gender hierarchy, but not beliefs that are unrelated to the legitimacy of the hierarchy (such as the height differential between men and women). Finally, Study 1 offered evidence that immutability explains the relationship between system threat and endorsement of biological essentialism for gender differences. Specifically, we adapted items from Levy and colleagues‘ (1998) measure assessing implicit person theory to specifically assess participants‘ beliefs about the (im)mutability of gender differences and found that these beliefs increased when participants were exposed to a system threat, and more importantly, that they fully mediated the relationship between experiencing a threat to the system and endorsement of biological essentialism for gender differences. In Studies 4a and 4b, we further examine the role of perceived mutability in our effects by experimentally manipulating whether biological and socio-cultural explanations for gender differences are presented as mutable or immutable. Study 2 Study 2 sought to conceptually replicate the results of our first study using a goal contagion paradigm (Aarts et al., 2004) to implicitly prime a motive to defend the existing social order. Previous work on goal contagion indicates that exposure to a target person who pursues a Essentialist Explanations for Gender Differences 24 goal causes the perceiver to implicitly adopt that same goal, even if the target and participant are from different social groups (Aarts, Gollwitzer, & Hassin, 2004; Wheeler & Petty, 2001). For example, reading about a student who wanted to earn extra money over the summer led participants to rush through their current task in order to get to a later task for which they could earn additional pay (Aarts et al., 2004). This study used a goal contagion manipulation in particular because it allowed us to test whether the effect of system justification on essentialism is a motivated process, as SJT predicts (Jost & Banaji, 1994). The goal contagion manipulation is qualitatively different from the system threat manipulation used in Study 1 in that, in the goal contagion manipulation, participants themselves are not experiencing a direct threat to the system, rather they are reading about someone else‘s goal to support the system. Therefore, we predict that adoption of pro-system motives via goal contagion should cause increased endorsement of essentialist explanations for gender differences, and that endorsement should be greater than in either a control condition, or a condition in which anti-system goals are activated. As in Study 1, we operationalized essentialist explanations as ̳biological essentialism.‘ Method Participants and Design Fifty-seven undergraduates (41 males and 16 females) were recruited from an Introductory Psychology subject pool at an East Coast University and were randomly assigned to either the pro-system goal condition, the anti-system goal condition, or the neutral prime condition. Essentialist Explanations for Gender Differences 25 Materials and Procedure Goal contagion manipulation. Participants were randomly assigned to carefully read one of three paragraphs ostensibly as part of a ―memory study.‖ Participants then completed the dependent measures and then were asked to write down the gist of the paragraph they read at the beginning of the study, which served as a manipulation check. All participants successfully summarized the goal contagion paragraph. Participants read a first-person account from the perspective of either a soldier who described fighting for a government he believed in, a rebel soldier fighting against a government he didn‘t believe in, or (in the control condition) a backpacker hiking through the woods. The anti-system goal condition read: ―We jumped into chest-high water. We held our rifles over our heads and waded to shore. We got on the beach and across the sand and then we got pinned down by machine-gun fire. The government‟s soldiers were lobbing grenades down on us. I thought to myself, “I‟m fighting against a government I don‟t believe in, against a society that doesn‟t care about me, and against leaders who I don‟t trust.‖ The pro-system goal condition was the same as the anti-system goal condition except that the word ―government‖ was replaced with ―enemy‖ and the final sentence read: ―I thought to myself, “I‟m fighting for a cause I believe in, for a society that cares about me, and for leaders who I trust.‖ The neutral condition read: ―We jumped into chest-high water. We held our backpacks over our heads and waded to shore. We got on the beach and across the sand and then made our way around some bushes and trees. Pine cones were falling from the tree branches above us. I thought to myself, „I‟m hiking through the brush, in a forest, and up a hill.‖ Biological Essentialism. In order to expand our assessment of endorsement of essentialist explanations for gender differences, we had participants read a passage about a hypothetical debate between two professors on why there are fewer female than male faculty members in math, science, and engineering. The professor who took the biological position was said to Essentialist Explanations for Gender Differences 26 believe that the under-representation in math, science and engineering faculty was due to ―biological gender differences in temperaments and abilities‖ in that ―women evolved to nurture children and families, and they prefer jobs working with people rather than things.‖ The professor advocating the socio-cultural position argued that the under-representation of women in these careers was ―caused by environmental factors like gender role socialization and job discrimination‖ in that ―parents think that their sons are more talented at math and science than their daughters, even when their daughters‘ grades and standardized test scores are just as high‖ (Appendix A). The pairing of the professors‘ respective names (Patton and Yanover) with the biological and socio-cultural positions was counterbalanced between-subjects, as was the order in which the two positions were presented. There were no significant effects of counterbalancing order, which is not discussed further. After reading each position, participants completed five items assessing their agreement with the positions on 8-point Likert-type scales. The items included: “If you had to choose, would you say you agree more with Patton, or with Yanover?”, “To what extent do you agree with Patton‟s position?”, “To what extent did you find Patton‟s arguments compelling?”, “To what extent do you agree with Yanover‟s position?”, and “To what extent did you find Yanover‟s arguments compelling?‖ The items assessing agreement with the professor who advocated the socio-cultural position were reverse-scored so that higher numbers on the scale indicated more agreement with the biological explanation for gender differences in math, science, and engineering. Finally, participants completed the same biological essentialism scale from Study 1. Essentialist Explanations for Gender Differences 27