Sexism and the Far-Right Vote: The Individual Dynamics of Gender Backlash Eva Anduiza and Guillem Rico Universitat Autònoma de Barcelona Abstract: This article contends that sexism plays a fundamental role in the electoral rise of the far right, both as a predis- position and as a changing attitude. Using panel data from Spain, we show that modern sexism is indeed among the most important attitudinal predictors of voting for the far-right party Vox. The results also show that internal individual changes in levels of modern sexism impact far-right voting. Backlash attitudinal change, defined as increases in sexism occurring in a context of feminist momentum, contributed significantly to the recent emergence of the radical right. Our findings indi- cate that sexism is not a crystalized attitude but rather susceptible to showing short-term changes with important political consequences. This highlights the importance of understudied context-dependent individual dynamics of gender backlash in far-right voting. Verification Materials: The data and materials necessary to verify the computational reproducibility of the results, procedures, and analyses in this article are available on the American Journal of Political Science Dataverse within the Harvard Dataverse Network, at: https://doi.org/10.7910/DVN/A11CD5 T he relationship between gender and the far right identify what is really at stake in contemporary debates has been the object of significant attention in about gender equality as well as proper measures of such terms of discursive, movement, and policy as- concepts. The concept and original operationalization of pects (e.g., Akkerman 2015; Köttig, Bitzan, and Pető modern sexism (Swim et al. 1995) serves this purpose. 2017; Kováts, Poim, and Pető 2015; Mudde 2019). Com- Second, we argue that sexism is a neglected yet poten- paratively, it is surprising how little consideration has tially key explanatory factor of far-right voting, both as been given to attitudes toward gender equality as an a preexisting predisposition and as an attitude that can explanation as why people vote for far-right parties. change within a relatively short time span. Third, we ar- Beyond some recent attempts to understand Donald gue that not all attitudinal changes are equally conse- Trump’s victory, we know relatively little about how gen- quential for vote choice. We distinguish between backlash der attitudes affect the electoral surge and growth of the and normalization attitudinal change and show that the far right. The virtual absence of any mention of atti- former can be more consequential than the latter regard- tudes toward gender equality in the most important lit- ing the surge of the far right, which suggests that there is erature reviews of the state of far-right voting is proof indeed an individual dynamic to gender backlash. of this blind spot (Arzheimer 2018; Golder 2016; Stock- Empirically, we use panel data from Spain collected emer, Lentz, and Mayer 2018). In this article, we analyze before, during, and after the massive feminist protests the relationship between gender attitudes and far-right of 2018 and 2019, and the ensuing surge of the far- voting, and make three claims. right party Vox. This context allows us to assess the ef- First, we argue that to properly assess the role of fect of individual changes in levels of sexism, distinguish- gender attitudes in far-right voting, we need concepts to ing between increases occurring in a context of feminist Eva Anduiza, Department of Political Science and Public Law, Universitat Autònoma de Barcelona, Edifici B, Campus de Bellaterra, 08193 Cerdanyola del Vallès, Barcelona, Spain (eva.anduiza@uab.cat). Guillem Rico, Department of Political Science and Public Law, Universitat Autònoma de Barcelona, Edifici B, Campus de Bellaterra, 08193 Cerdanyola del Vallès, Barcelona, Spain (guillem.rico@uab.cat). We thank Marta Fraile for helpful feedback on earlier versions of this article as well as the editors and reviewers at the AJPS; participants in EuroPOW and seminars held at the universities of Barcelona (IPERG), Gothenburg, Liverpool, Lisbon (ICS), Mannheim (MZES), Stanford (CASBS), Vienna (Institut für Staatswissenschaft), and Wissenschaftszentrum Berlin für Sozialforschung; participants in EPSA, German Political Psychology Network, and ISPP meetings; and participants in workshops at Lausanne (FORS), Florence (Scuola Normale Superiore), and Pittsburgh (European Studies Center). Support for our study was provided by the Ministry of Science and Innovation (grants CSO2014-52950-R and CSO2017-83086-R) and by the ICREA foundation. American Journal of Political Science, Vol. 0, No. 0, November 2022, Pp. 1–16 © 2022 The Authors. American Journal of Political Science published by Wiley Periodicals LLC on behalf of Midwest Political Science Association. DOI: 10.1111/ajps.12759 This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and repro- duction in any medium, provided the original work is properly cited. 1 15405907, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/ajps.12759 by Consorci De Serveis Universitaris De Catalunya, Wiley Online Library on [01/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License 2 EVA ANDUIZA and GUILLEM RICO momentum (backlash attitudinal change) and in a far- the many analyses that show that gender was and is a right surge context (normalization attitudinal change). powerful element in the definition of far-right discourse, Our analyses show that sexist attitudes—measured be- anti-feminist mobilization, and policy agendas, gender fore either feminist mobilization or the emergence of the issues have been largely ignored; attention has, at most, far right—were key to explaining the votes for Vox when reached the analysis of the gender gap in far-right vote it entered parliament. In addition, our data show that (Harteveld et al. 2015). In trying to understand why there have been short-term changes in sexist attitudes, people vote for far-right parties, attention has been and these impacted vote choice. In particular, increases devoted to ideological and attitudinal explanations, such in sexism that occurred at a time of heightened feminist as left–right orientation, attitudes toward redistribution, mobilization and declining overall levels of sexism had a attitudes toward law and order, anti-EU attitudes (in significant impact on the vote later. This evidence leads Europe), and in particular attitudes toward migration us to conclude that the attitudinal backlash reaction to (e.g., Arzheimer 2018; Edo et al. 2019; Georgiadou, Rori, feminist mobilization was a key element in the electoral and Roumanias 2018; Zhirkov 2014) but not to gender. rise of the far right in Spain. Except for a handful of exceptions we shall discuss below These findings inform our understanding of why (Green and Shorrocks 2021; Lodders and Weldon 2019; people choose to vote for a far-right party. Our results Spierings and Zaslove 2015), attention and findings have show that sexism matters more than has been acknowl- been focused elsewhere. If one were to identify the main edged in the comparative literature of far-right support. issue that accounts for the rise of the far right according Thus, the disregard of sexist attitudes—so far scarcely to previous work, it would be a reaction against immi- and poorly measured—should be corrected. gration or globalization rather than a reaction against Furthermore, our longitudinal analysis contributes greater gender equality. to unpacking and clarifying the micro dynamics of gen- Events, such as the 2016 election resulting in the der backlash. We show that sexism is an attitude with proclamation of a blatantly sexist U.S. president or the a hybrid nature wherein both crystalized predisposi- wave of feminist mobilizations that followed around the tions and changing evaluations must be considered (Al- globe, have put gender attitudes more center stage in barracin and Shavitt 2018). Our analysis discloses the electoral politics, and there are a significant handful of individual-level dynamics of the systemic argument of works investigating how different forms of sexist atti- gender backlash that asserts that the rise of the far right is tudes impacted vote choice in the 2016 U.S. election a reaction against cultural change and feminist mobiliza- (Bock, Byrd-Craven, and Burkley 2017; Hanley 2021; tion: increases in sexist attitudes in a context of feminist Schaffner, MacWilliams, and Nteta 2018; Valentino, mobilization and overall decreasing sexism explain rising Wayne, and Oceno 2018). Important as this case may preferences for the far right. be, it could also be exceptional, and not representative of how sexism matters for the far-right vote. We need to extend this incipient attention to other cases, unpack the reasons why sexism should matter, assess whether sexism Cultural Backlash, Gender, and the is a stable crystalized predisposition or can be subject to Rise of the Far Right short-term change, and gauge its electoral consequences. One of the central interpretations for the rise of the far right is as a conservative reaction to progressive value change. A “tectonic” long-term generational shift toward Sexism in a Far-Right Context more inclusive, libertarian, cosmopolitan, universalistic values is naturally followed by a reaction in the oppo- Attitudes toward gender equality include a large num- site direction that far-right parties embody (Ignazi 2003; ber of different concepts (Burns and Gallagher 2010). Kitschelt 1995; Norris and Inglehart 2019). Even if atti- Our focus is on sexism as a belief or attitude that re- tudes toward gender equality are not entirely consistent flects negative evaluations of individuals based on their with postmaterialist values (Hayes, McAllister, and Stud- sex or contributes to the maintenance of sex-based in- lar 2000), it is difficult to understand cultural liberalism equality in society (Becker and Sibley 2016; Swim and without including a defense of gender equality. Hyers 2009). Although sexism can technically be directed It is, however, striking that in the comparative litera- toward both sexes, it is mostly directed toward women ture that aims to explain the vote for the far right, gender because of existing sexual hierarchies. We select sexism is mostly absent from the empirical analysis. In spite of over other gender-related attitudes because the concept 15405907, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/ajps.12759 by Consorci De Serveis Universitaris De Catalunya, Wiley Online Library on [01/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License THE FAR-RIGHT VOTE AND GENDER BACKLASH 3 has adapted very well to social changes regarding gender Christian societies" and voicing unfair requests. Femi- roles and stereotypes, and connects remarkably with the nism or so-called “gender ideology” are dangerous exis- debates at stake in the conflict between feminism and the tential threats to family values, children’s well-being, the far right. stability of Western societies, and the integrity of their In terms of values, advanced democracies have em- cultural basis. Gender works as a “symbolic glue” that braced the primacy of the principle of equality between allows the bringing together of different questions re- the sexes. Many countries have long past the time when garding the organization of society, such as care work, male supremacy could be proclaimed. This does not nec- reproductive rights, gender mainstreaming, and educa- essarily mean that every individual endorses such princi- tion (Kováts, Poim, and Pető 2015; Pető 2015). Building ples in their hearts, let alone has the same consideration on gender as a key political conflict is a second impor- of what gender equality should entail in practice. Neither tant characteristic of far-right discourse, which opposes does the endorsement of this principle entail, of course, both feminist mobilization and policies aimed at correct- real equality or even equity between men and women. ing gender-based inequalities and discrimination. However, socially the principle is well rooted as a value Moving into the realm of individual attitudes, the (Rhodes et al. 2020; Scarborough, Sin, and Risman 2019). concept of modern sexism precisely captures these el- Sexism today takes more subtle forms, which are ements. It was initially elaborated on by Swim and her present in the discourses of far-right parties. This is well colleagues inspired by the attempt to distinguish old- reflected in the evolution of the concept and its measures. fashioned or overt and subtle or covert forms of racism In blatant, hostile, or benevolent sexism, the core element (Swim et al. 1995; Swim and Cohen 1997). The concept is beliefs in stereotypes regarding women (and some- has traveled across time and space well and is strongly times also men) (Glick and Fiske 1997; Swim et al. 1995). connected to the contemporary far-right discourse de- While these forms of sexism are still found in our soci- scribed above. Modern sexism involves three dimensions. eties and in the discourses of political leaders, they leave The first one is the denial of women’s discrimination. Let out two questions that are of paramount importance for us remember that sexism consists not only of attitudes the analysis of gender in the political realm today: percep- and beliefs that harm women but also of resistance to tions of discrimination against women and the consider- end sex-based inequality. The most obvious form of ation of gender not as a question of stereotypes or values such resistance is to deny existing inequalities and dis- but as a question that entails an assessment of a political crimination. The second component is the rejection and situation where conflict is involved. In other words, de- delegitimization of any complaint involving women’s bates today are not articulated around whether women discrimination. Protests, mobilization, or outrage are should work outside the home, for example, but rather unjustified since there is no discrimination to correct. around whether they are still facing discrimination when Finally, the third component of modern sexism is the they do so, and what should be done about it. rejection of any measures taken to correct inequality Far-right parties use the value of women’s rights and discrimination, which are considered unfair favors and gender equality to legitimize their attacks on Is- or even as discriminatory against men: institutional lam and Islamic migrants (Akkerman 2005; Akkerman campaigns, legislation on affirmative action, or even and Hagelund 2007; Lépinard 2020; Möser 2020). They mere attention to such matters, are contested. often assume that migrants’ countries of origin have Modern sexism is more socially acceptable than pronounced gender inequality and discrimination chal- hostile forms of sexism (Becker and Sibley 2016), and lenges, in contrast to their own countries, which they therefore less prone to produce an underestimation of view as having achieved real equality between men and sexism due to social desirability. However, its main added women (Schwab et al. 2019). Even when making familist value in our view comes from the fact that it is able to discourses wherein women are seen as belonging primar- capture the key elements of the far right’s discourses on ily to the private sphere, they do so wrapped in the terms gender equality and transform these into an individual of equal rights (Grzebalska and Pető 2018). Denial of dis- attitude. While sexism could be interpreted as a stable crimination is, therefore, a key element in far-right dis- predisposition that should be measured as different and course. as distant from politics as possible (Schaffner 2021), we In addition, gender appears in the debate not only as argue that it is precisely the connection of the individual a question about what men and women should be like or attitude with the discourses of far-right parties that how they should behave but rather as a key conflict where makes it an interesting predictor of vote choice for these some (“gender ideology,” “feminists,” “women’s organi- parties as well as an attitude prone to short-term change. zations”) are threatening the core principles of "Western If measures of individual attitudes were disconnected 15405907, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/ajps.12759 by Consorci De Serveis Universitaris De Catalunya, Wiley Online Library on [01/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License 4 EVA ANDUIZA and GUILLEM RICO from ongoing debates, we would be less well equipped Oceno 2018). The results are less clear for benevolent sex- to grasp the extent to which sexism is still present in our ism (Cassese and Holman 2019; Ratliff et al. 2019). There society, its changes, and its political consequences. is evidence that other attitudes, such as gendered nation- alism, also affected the 2016 vote (Deckman and Cassese 2019). Perceptions of discrimination against women af- fected voting among men as much as partisanship, ide- Sexism as a Predisposition ology, or economic evaluations (Simas and Bumgard- ner 2017). The relationship between modern sexism and We expect far-right parties to be significantly more at- perceptions of how Trump and Clinton fit their gen- tractive to people who hold sexist attitudes, believe sex- der stereotypes has been studied (Godbole, Malvar, and based discrimination is irrelevant, and oppose any at- Valian 2019). Very recently, Hanley (2021) compared the tention given to the issue aiming to alter the status quo. effect of some hostile and modern sexism measures on Hence, we expect sexism to matter in its own right, inde- presidential voting, concluding that while the former pendent of other attitudinal explanations of vote choice have been more frequently used in the literature, only the that have received extensive attention in the far-right latter show an effect on vote choice in 2016, which prob- vote literature. Since sexism can also increase as a con- ably points to an underestimation of the effects of sexism sequence of a person becoming a far-right supporter, we on vote choice in the U.S. also need to consider the possibility that the relationship Overall, we have some cross-sectional evidence runs in both directions. pointing toward an association between sexism and vote Evidence of how sexism is related to the far-right choice in the U.S. but not so much for Europe. Measures vote is at best unbalanced. In Europe, very few studies are particularly poor outside of the U.S. context. The address the effects of attitudes toward gender equality on concept of modern sexism has not been fully measured vote choice (Hayes 1997). One of the first studies to con- to assess its relationship with vote choice in Europe (and sider empirical measures of attitudes toward economic only in Hanley’s study for the U.S.). Our first objective is gender roles as a predictor of far-right vote choice found to evaluate the extent to which modern sexism—the con- no significant effects (Spierings and Zaslove 2015). In cept that best reflects far-right discourses on gender— their analysis of four European cases, Lodders and Wel- works as a predisposition that affects the probability to don (2019) found some effect, but it was still far smaller vote for a far-right party. than other explanations such as left–right self-location or attitudes toward migration. These works are based on ev- idence prior to the massive women’s mobilization of re- cent years and use the European Social Survey’s limited Changes in Sexism: Individual measures of sexism. More recently, Green and Shorrocks Dynamics of Gender Backlash (2021) find evidence of a significant relationship between a measure of male resentment based on perceptions of The individual dynamics of the relationship between sex- discrimination and the Brexit vote, and Geus, Ralph- ism and the far-right vote involve at least three differ- Morrow, and Shorrocks (2022) find an association be- ent facets. First, there is the question of the extent and tween hostile sexism and voting Conservative. patterns of attitude change: are levels of sexism stable Sexism and prejudice against women have been or do they change over time? Second, we must con- comparatively more analyzed in the U.S., where re- sider how the relationship between attitudes and vote search has looked at how sexism and gender attitudes choice changes over time: is the link between sexism affect perceptions of female candidates and other polit- and far-right vote choice stable, or can sexism be acti- ical attitudes with mixed findings (Cassese, Barnes, and vated/deactivated? Finally, we must consider the conse- Branton 2015; Ditonto 2019; Dwyer et al. 2009; Mc- quences of attitudinal change for the vote: are increases Thomas and Tesler 2016; Mo 2015; Simas and Bumgard- in sexism linked with increases in the probability to vote ner 2017). The Trump victory clearly spurred attention for the far right? to this question. A significant number of studies now Not much evidence tracking longitudinal changes in show that hostile sexism is positively related to Trump sexism is available. Archer and Kam (2020) use panel data support (Bock, Byrd-Craven, and Burkley 2017; Cassese to explore levels of sexism before and after the #Metoo and Barnes 2019; Cassese and Holman 2019; Frasure- movement emerged. They found these attitudes were on Yokley 2018; Knuckey 2019; Ratliff et al. 2019; Schaffner, average highly stable. This stability may be specific to MacWilliams, and Nteta 2018; Valentino, Wayne, and the U.S., where the feminist mobilization was partly a 15405907, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/ajps.12759 by Consorci De Serveis Universitaris De Catalunya, Wiley Online Library on [01/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License THE FAR-RIGHT VOTE AND GENDER BACKLASH 5 reaction to Trump’s victory. Limitations in the measure- evidence of backlash dynamics not only in terms of ment instruments may also be at work, as Archer and macro processes but also in individuals’ attitudes. Do Kam’s scale uses three items, only one of which is taken some people actually move from more progressive to from Swim et al.’s original battery of modern sexism indi- less progressive positions when a feminist tide emerges? cators. In any case, aggregate stability may be hiding con- Do these changes matter for vote choice? Increases in sequential internal changes within individuals that need sexism in a context of heightened feminist mobiliza- to be explored. tion can be considered as observed backlash attitudinal Regarding the stability of the relationship between change. sexism and vote choice, we know from previous work Normalization has also been widely used in the con- that sexism can be activated as a political issue. Activa- text of the far right to refer to a process by which taboos tion is the strengthening of the association between the becomes destigmatized and acceptable—when “shame- attitude and vote choice, which may be due to different less discourse” becomes visible and present in institu- underlying processes (such as heightened accessibility, tions (Selvanathan and Leidner 2021; Valentim 2021; salience, or issue voting) with similar observable impli- Wodak, Culpeper, and Semino 2021). Increases in sex- cations (Hopkins 2021). Cassese and Barns (2019) argue ism might occur in a favorable context of heightened far- that hostile sexism was activated in the U.S. presidential right visibility. We refer to such increases as normaliza- election of 2016: it mattered much more than in 2012 tion attitudinal change. and so did perceptions of discrimination. Valentino et al. (2018) argue along the same lines: sexism played a minor role before 2016. In 2016, its effect became substantial, similar to authoritarianism (but still less important than Case, Data, and Measurement racial resentment or party identification). Hanley (2021) Spain: Feminist Mobilization and Far-Right also finds evidence of the activation of sexism, which in Visibility 2016 became a more relevant predictor of the vote, par- ticularly due to Trump’s ability to repel nonsexist voters. We analyze the relationship between sexism and the far- While this is informative of how important sexism was right vote and its individual dynamics using evidence as an explanation for vote choice in these cases (with from the Spanish case for three reasons. First, we have some potential risks of endogeneity because of the use data measuring modern sexism via the full original bat- of cross-sectional data), this is only part of the story; we tery of items in a sample of citizens taken during four also need to assess when attitudinal change matters for time points between 2017 and 2020. We ensure that our vote choice. This question has received less attention, as key concept of interest—modern sexism—is, therefore, it requires longitudinal data, which to our knowledge accurately measured over time. These longitudinal data remain largely unavailable so far. allow us to estimate the effect of prior values of sexism as Within-individual attitude change is not simply a well as the effect of changes in sexism over time. source of variation that allows us to control for time- Second, Spain has witnessed a succession of events, invariant unobservables when explaining vote choice. including two significantly different moments, in which The direction and the timing of attitudinal changes may attitudinal change may have happened. In 2018, the also be relevant. The concepts of backlash and normal- country witnessed a moment of heightened feminist ization can be applied to the characterization of changes mobilization without any visible or significant far-right in sexism and their consequences. Both “backlash” and actor. This was followed in 2019 by the surge of a far-right “normalization” have been widely used in the analysis of party (Vox), which garnered intense media attention. far-right movements and parties, though more as a nar- This sequence of events allows us to trace changes in rative framework than as theoretical models with observ- levels of sexism under different circumstances. While we able implications. make no causal claim as to how these different contexts Backlash has been defined as a reaction to a shift in affect attitudes and behavior, the two different scenarios power (Mansbridge and Shames 2008), or more specif- allow us to explore the dynamics of attitudinal change ically to our concerns, a reaction to a threat of forth- in a unique way by assessing which type of attitudinal coming shifts in gendered power relations (Sanbonmatsu change is relevant for party preferences. It could be 2008). The use of the concept, even if, extended in the that increases in sexism that happened before the far- analysis of the far right, has spurred some debate as right surge (that is between 2017 and 2018) could have to whether it is adequate as a generalized explanation possibly induced the surge and enhanced support for (Kováts 2018; Paternotte 2020). One of the pending is- the far right, suggesting a backlash dynamic. It could sues in this discussion is the extent to which we find also be that increases in sexism that happened once 15405907, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/ajps.12759 by Consorci De Serveis Universitaris De Catalunya, Wiley Online Library on [01/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License 6 EVA ANDUIZA and GUILLEM RICO Vox were in the spotlight (between 2018 and 2019) are manded the removal of the law against gender violence electorally consequential, suggesting a normalization (Bernardez-Rodal, Rey, and Franco 2020; Ferreira 2019). dynamic. The year 2019 is, therefore, the year of the emergence of Third, existing accounts of the birth and growth of the far right in Spain. this new far-right party thus far do not consider gender- In line with other studies of the far-right vote, gen- based attitudes and instead point to the relevance of the der attitudes have been barely considered in explana- territorial conflict. According to some interpretations, tions of the electoral rise of Vox. However, some qualita- Vox’s electoral success is the result of a reaction against tive research has highlighted the relevance of gender and the Catalan secessionist bid (Turnbull-Dugarte 2019). anti-feminism in Vox’s discourses, which include all the Migration, hardly as salient an issue in Spain as in other components of modern sexism. Vox denies the existence European countries, has been a secondary issue (Mendes of discrimination against women (“Spain is a country and Dennison 2021). Likewise, gender does not play a where men and women are equal before the law”), and significant role in accounts of why Spain has not had if any discrimination exists it is carried out by Muslim a far-right party until recently (Alonso and Kaltwasser migrant workers (Cabezas 2022). Vox attacks feminists 2012) nor on the rise of Vox (Rama et al. 2021). Anti- calling them “feminazis” and considers them a threat to feminism is considered a characterizing feature of Vox, the family as an institution, responsible for “totalitarian but it is not its raison d’être (Ferreira 2019). In the same indoctrination,” and hence should not receive any public way, individual-level explanations of vote choice for Vox funding (Cabezas 2022; Fernández-Suárez 2021). Finally, have considered and emphasized explanatory factors re- Vox fervently rejects existing legislation against gender lated to left–right orientation, territorial conflict, and im- violence. In their view, the 2004 act—passed with unani- migration but not gender (Rama et al. 2021; Turnbull- mous approval in parliament—“liquidates constitutional Dugarte 2019). This should make Spain a relatively hard guarantees and removes basic civil rights for half of the case in which to identify important effects of sexism on population” by discriminating against men (Fernández- vote choice. Suárez 2021). These are all the core elements of modern Spain followed the global wave of feminist mobi- sexism. lization against sexual harassment and violence, with marches of historical proportions in 2018 and 2019, spurred by a case of gang rape that generated an out- Panel Data and Measures pouring of outrage. Protests took place throughout the country, mobilizing hundreds of thousands of people, We rely on data from the Spanish Political Attitudes mostly women. The term “feminism” doubled its pres- dataset (Hernández Pérez et al. 2021), an online panel ence in the mainstream media, with less than 1,400 pub- study conducted yearly on a sample with quotas based lished news or articles on the topic in 2017 and almost on sex, age, education, region, and municipality size to 3,000 in 2018.1 Therefore, in 2018 there was a heightened ensure a balanced representation of the Spanish adult saliency of issues regarding gender equality and violence population between 18 and 56 years of age.2 The analysis against women, the presence of feminist marches of his- focuses on the four waves fielded between 2017 and torical proportions, and an amplified media visibility of 2020, for which the modern sexism battery is available. feminism. The supporting information (SI, p. 2) reports the main This feminist tide was followed by the increased vis- characteristics of the panel. This longitudinal data help ibility of Vox, the far-right party that splintered from shed light on the overall and individual dynamics of the conservative People’s Party (PP) in 2011. Vox gath- gender attitudes in a context where gender issues gained ered significant media attention in October 2018 (Olalla, significant attention, first in the context of feminist Chueca, and Padilla 2019). It entered the regional An- mobilization, and later, with Vox as a new key political dalusian Parliament in December 2018 and the national actor attacking feminist policies and protests. Most Parliament in April 2019 with 24 seats out of 350. When importantly, they allow us to test the role of sexism in elections were repeated in November, Vox acquired 52 2 Because internet use among the older portion of the country’s seats, securing visibility in the Spanish chamber and population was still comparatively low when the study was first the media. Vox painted a picture of feminists as “vio- launched in 2010, the original sample was restricted to Spanish lent” and “communists and radicals,” and repeatedly de- residents aged between 16 to 44 years. In terms of our findings, we would expect older citizens to be more (Norris and Inglehart 2019) or at least not less (Schäfer 2021) likely to exhibit cultural backlash, 1 Personal elaboration data obtained from ABC, El Mundo, El Pais, and therefore our estimation, if anything, would be a conservative La Vanguardia. one. 15405907, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/ajps.12759 by Consorci De Serveis Universitaris De Catalunya, Wiley Online Library on [01/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License THE FAR-RIGHT VOTE AND GENDER BACKLASH 7 FIGURE 1 Mean Levels of Sexism over Four Waves and Timeline of Political Events Notes: Sexism is measured on a 0–1 scale. Mean levels with 95% confidence intervals. the emergence of Vox while overcoming endogeneity four time points, recoded to range between zero and one, concerns arising from reverse causation. It is key to our with higher values denoting higher sexism (Figure A1 in purposes that the first wave of data was collected before SI p. 9 shows the distribution of modern sexism in each the first massive feminist mobilization and the second wave). wave before the rise of Vox. Our vote models account for the effect of other at- Because we are interested in capturing respondents’ titudinal factors found to predict support for the rad- support for Vox at the precise moment of each of their ical right in the comparative literature. Specifically, we interviews, we used intention to vote for Vox (rather include controls for general ideological identification than reported vote in actual past elections) as our depen- (11-point left–right scale), authoritarianism (4-item bat- dent variable. Respondents were asked which party they tery based on childrearing values, Feldman and Stenner would vote for if there were general elections tomorrow. 1997), nativism (assessing attitudes to the economic and Those who mentioned Vox were coded as one and the rest cultural consequences of migration), populist attitudes as zero.3 Given that support for Vox was negligible until (Akkerman, Mudde, and Zaslove 2014), and territorial late 2018, our analyses of intended vote focus on the 2019 preferences (higher values denoting greater support for and 2020 waves.4 decentralization). Since part of our arguments rely on assessing the The models also controlled for sex, age, education impact of changes in modern sexism, it is important to (middle school or less, high school/vocational training, use a valid and reliable measure of this construct. Un- college), household income (a scale with 12 intervals), like previous works that have struggled with limited and whether the respondent lives with a partner, and a 4- scarce instruments, we use the full battery proposed by point measure of interest in politics. All variables except Swim et al. (1995) with an additional item to better gauge age (in years) have been recoded to run from zero to one resentment about measures to correct gender inequality (see SI, pp. 7–8 for detailed information about all our and discrimination (see SI, p. 8 for measurement details). measures). The internal reliability of the composite score was satis- factory, with alphas ranging between 0.86 and 0.89 across waves. Figure 1 shows the timeline of the panel waves and political events as well as average levels of sexism for the Results 3 Trajectories in Modern Sexism Nonresponses and intended nonvoters were coded as zero. 4 The intended vote for Vox was below two percent before 2019 Before looking at the consequences of sexism and sex- (only 15 and 26 respondents in 2017 and 2018, respectively). ism changes for far-right support, we first explore how 15405907, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/ajps.12759 by Consorci De Serveis Universitaris De Catalunya, Wiley Online Library on [01/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License 8 EVA ANDUIZA and GUILLEM RICO FIGURE 2 Distribution of Changes in Modern hort, education, household income, living with a part- Sexism ner, interest in politics, ideology, partisanship (based on reported vote in the 2016 general elections), and engage- ment in the 2018 Women’s Day protests. Note that be- cause this analysis was restricted to respondents who par- ticipated in all four waves (N = 807) and covariates are treated as time-invariant (fixed at their 2017 values), the trajectories trace the evolution of exactly the same groups of individuals over the whole period. Figure 3 presents the predicted trajectories by the most relevant covariates (sex, ideology, partisanship, and engagement in Women’s Day protests), based on the esti- mates of the multilevel growth curve model (Table A5 in SI, pp. 10–11 includes the full details). Although the pat- terns of change look remarkably similar after 2018, with all groups following roughly the same upward trend, no- ticeable variations occurred between 2017 and 2018. The gaps in levels of sexism increased by sex, left-right identi- fication, party support and engagement in Women’s Day protests (trajectories of other variables are shown in Fig- ure A2 in SI, p. 12). Thus, some people were resistant to the overall trend of declining sexism. The estimates sug- Notes: Sexism was measured on a 0–1 scale. The proportion of gest that sexist attitudes realigned considerably between respondents who increased their levels of sexism was 37.8% in 2018 (16.2% by more than 0.1 points on the sexism scale), 54.1% 2017 and 2018, following the feminist protests and just (26.4%) in 2019, and 53.8% (24.5%) in 2020. before the emergence of Vox. Differences remained sta- ble afterwards, with all groups gradually becoming more sexism changed during the analyzed period. Figure 1 de- sexist at a similar pace. In other words, the time when picts average levels of sexism by wave. gender issues were increasingly polarizing voters was just Sexist attitudes declined slightly in 2018, from 0.36 before Vox appeared, not afterwards. to 0.34 on a 0–1 scale, likely reflecting the rise of the #MeToo movement and the massive turnout at the 2018 Women’s Day marches around the country. However, Modern Sexism and the Rise of Vox sexism increased to 0.37 in June 2019, just after Vox We next consider how sexist attitudes, along with other entered the national parliament for the first time, and factors typically connected to the far-right vote, are as- then increased again to 0.41 in June 2020, once the party sociated with support for Vox. Table 1 presents the es- had increased its representation in the early elections of timates of two cross-sectional logit models of intended November 2019. Although these changes are relatively vote for the 2019 and 2020 waves, respectively: moderate in size, paired t-tests indicate that the differ- ences between adjacent waves are all highly statistically voxit = sexismit + other attitudesit + controlsit (1) significant (p < 0.01).5 Note, however, that as shown in Figure 2, changes in attitudes happen in all directions where individuals are indexed as i and time (wave) as t; for all waves, revealing that individual trajectories were other attitudes refer to ideology, authoritarianism, na- much more diverse than aggregate figures might suggest. tivism, territorial preferences, and populism; and the To gain further insight into the variation in individ- controls are sex, age, education, income, living with a ual trajectories of sexism, we employed multilevel growth partner, and interest in politics. curve modeling, which treats observations for each wave The models reveal that the effects of the attitudinal (level 1) as nested within individual respondents (level variables all go in the expected direction: voting for Vox 2). Specifically, we estimated a random slope model with is positively influenced by right-wing orientation, sex- time as a categorical predictor, and covariates for sex, co- ism, nativism, and populist attitudes (if only significantly in 2020) and negatively by attitudes in favor of decen- 5 The results are nearly identical when the sample is restricted to tralization. The impact of modern sexism is surpassed the respondents who completed all four waves. only by that of ideology. Holding all other factors at their 15405907, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/ajps.12759 by Consorci De Serveis Universitaris De Catalunya, Wiley Online Library on [01/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License THE FAR-RIGHT VOTE AND GENDER BACKLASH 9 FIGURE 3 Predicted Levels of Modern Sexism by Sex, Ideology, Partisanship, and Women’s Day Engagement Notes: Predicted levels of modern sexism with 95% confidence intervals, based on the estimates of Table A5 (SI, pp. 10–11). The dependent variable is the modern sexism scale running from 0–1. Inde- pendent variables are measured in the first wave (2017) unless otherwise indicated. Ideology. 0–2 = Far left, 3–4 = Left, 5 = Center, 6–7 = Right, 8–10 = Far right. Partisanship: based on reported vote in the 2016 general election; “Others” includes voters of other parties and nonrespondents. Women’s Day protest engagement (measured in 2018): 0–1 composite score where 1 indicates participation in the four activities asked, including striking, demonstrating, mobilizing others to participate, and talk- ing about the protests. Results for age, education, income, living with a partner (2018), and interest in politics are shown in Figure A2 (SI, p. 12). observed values, individuals at the ninety-fifth percentile intention. Following Lenz’s (2012) specification strategy, of sexism are 8.6 (2019) and 9.9 (2020) percentage points we first examine how prior attitudes influence changes in more likely to express support for Vox than those at the vote intention, regressing vote intention at time t on atti- fifth percentile. tudes and vote intention measured at time t−1, plus con- Although the previous results suggest the relevance trols. This model tests whether the shifts in Vox support of modern sexism in explaining the vote for Vox, the registered in 2019 and 2020 were associated with previ- fact that the predictors were measured at the same time ous levels of sexism and other attitudes, allowing us to (i.e., wave) as the outcome does not allow us to draw identify any differences between the breakthrough and firm conclusions about the causal precedence of sexism consolidation phases. and the remaining attitudinal factors to vote intention. It may be the case that voters adjust their views to bring voxit = voxit −1 + sexismit −1 + other attitudesit −1 them in line with their preferred party’s positions (Lenz +controlsit (2) 2012). We, thus, cannot rule out that Vox voters’ opinions on women’s discrimination are actually a consequence, As shown in Table 2, the results of this test indicate rather than a cause, of their partisan preferences. that sexism measured in the previous wave significantly One way to circumvent this difficulty is to assess the predicts change in intended vote for Vox both in 2019 effect of previously measured attitudes on changes in vote and (albeit marginally) in 2020. That is, respondents who 15405907, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/ajps.12759 by Consorci De Serveis Universitaris De Catalunya, Wiley Online Library on [01/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License 10 EVA ANDUIZA and GUILLEM RICO TABLE 1 Predictors of Intention to Vote for Vox fifth percentile. The lagged values of ideological orienta- in 2019 and 2020 tion and territorial preferences also have an effect in both years, while the effect of populism is not statistically sig- 2019 2020 nificant in either year and that of nativism is only signif- Female 0.118 −0.145 icant in 2020. The heightened impact of attitudes toward (0.277) (0.220) immigrants is indeed the most remarkable difference be- tween the two waves. This suggests that immigration was Age 0.004 −0.009 not a key factor in the emergence of Vox and that it only (0.016) (0.010) became relevant after the party attained representation— hence once its stances on this issue gained visibility. By High school / Vocational −0.716 0.198 contrast, previous sexist attitudes were paramount for (0.442) (0.286) the party’s breakthrough and remained relevant after- College −0.075 0.080 wards, although slightly deactivated. (0.300) (0.256) We can take further advantage of our panel study to examine how changes in sexist attitudes affect subsequent Income −0.529 −0.061 changes in vote intention. To this end, we regress vote in- (0.575) (0.446) tention at time t on changes in attitudes and vote inten- Lives with partner 0.192 0.018 tion between times t−2 and t−1 as well as their values at (0.301) (0.231) t−2, plus controls. This model tests how prior shifts in individuals’ views lead them to change their vote inten- Interest in politics 0.634 0.915∗ tion, bringing their partisan preferences in line with their (0.478) (0.364) updated issue opinions, as follows: Authoritarianism −0.499 0.136 voxit = voxit −2 + sexism it −2 + other attitudesit −2 (0.540) (0.408) + voxit −1 +  sexismit −1 ∗∗ ∗∗ Ideological identification 5.497 4.965 + other attitudesit −1 + controlsit (3) (0.729) (0.587) The results, reported as Models 1 and 2 in Table 3, Nativism 2.646∗∗ 2.280∗∗ indicate that earlier changes in sexism significantly pre- (0.655) (0.564) dict changes in the probability of intending to vote for Vox in 2019 but not in 2020.6 People who became more Territorial preference −1.314∗ −1.905∗∗ sexist between 2017 and 2018 were more likely to switch (0.528) (0.398) their support to Vox in 2019. As an illustration, the es- Populism 0.894 1.418∗ timates imply that for someone whose level of sexism (0.741) (0.625) decreased by 0.26 points (fifth percentile) in 2018, the probability of supporting the radical right in 2019 was Sexism 4.159∗∗ 2.983∗∗ 3.8%, while the probability was 8.2% for someone whose (0.749) (0.583) sexism increased by 0.20 points (ninety-fifth percentile). However, those who increased their sexism between 2018 Constant −9.712∗∗ −8.419∗∗ and 2019 were not significantly more likely to switch to (1.189) (0.829) Vox in 2020. Observations 1651 1972 The preceding models assume that the impact of shifts in sexist attitudes are symmetrical, that is, that Notes: The dependent variable is intended vote for Vox. Logistic regression coefficients, standard errors in parentheses. decreases in sexism have the same effect as increases † p < 0.1, ∗ p < 0.05, ∗∗ p < 0.01. in sexism but in the opposite direction. To relax this assumption, Models 3 and 4 in Table 3 use two separate held sexist attitudes in 2018 or 2019 were more likely measures of changes in sexism: one for increases, which to switch their vote to the radical right one year later. 6 According to the estimates, the predicted probability of Given the low number of respondents expressing a vote intention switching to Vox is 5.7 (2019) and 3.4 (2020) percent- for Vox before 2019, the inclusion of prior change of this variable as a predictor causes a problem of separation. To deal with this age points higher for an individual at the ninety-fifth issue, the models in Table 3 use penalized maximum likelihood percentile of the sexism scale than for someone at the estimation (Firth 1993; Kosmidis et al. 2021). 15405907, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/ajps.12759 by Consorci De Serveis Universitaris De Catalunya, Wiley Online Library on [01/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License THE FAR-RIGHT VOTE AND GENDER BACKLASH 11 TABLE 2 Effect of Prior Attitudes on Intended takes the value of change if the shift is positive and zero Vote for Vox otherwise; and one for decreases, which takes the value of change if the shift is negative and zero otherwise. The 2019 2020 results indicate that the 2017–2018 increases in modern Female 0.095 −0.114 sexism had an impact on the probability to switch to (0.283) (0.305) Vox in 2019, whereas decreases in sexism did not have a statistically significant effect. This suggests that the Age −0.009 −0.026 impact of attitudinal change was mainly driven by peo- (0.015) (0.017) ple who countered the overall downward trend in sexist attitudes observed in the context of the massive feminist High school / Vocational −0.905† −0.387 mobilization of 2018. Specifically, the estimates of Model (0.496) (0.448) 3 predict that those whose sexist attitudes increased by 0.20 points had a likelihood of supporting Vox of 9.3%, College −0.039 −0.135 compared to 4.8% among those who maintained their (0.306) (0.347) level of sexism. On the other hand, Model 4 confirms Income 0.153 0.104 that changes occurring between 2018 and 2019 did not (0.564) (0.642) affect switching to Vox in 2020. Overall, the evidence in Table 3 suggests that, in line Lives with partner −0.086 0.052 with the backlash interpretation, increases in sexism oc- (0.289) (0.314) curring in a context of 2018’s massive feminist mobiliza- tions paved the way for the rise of Vox. Conversely, in- Interest in politics 0.946† 1.249∗ creases in sexism occurring at a time when the radical (0.497) (0.496) right obtained institutional representation had no clear impact on the party’s subsequent electoral gains. It is Prior Values (t−1) worth noting that, apart from ideological identification, Vox intention 3.071∗∗ 3.367∗∗ just as only changes in sexism predicted switching to Vox (0.699) (0.355) in 2019, only changes in nativist attitudes did so in 2020. This is consistent with the argument that the electoral Authoritarianism 0.702 0.368 success of the radical right might have contributed to the (0.538) (0.582) enhanced visibility of xenophobic rhetoric and to the ac- tivation of the immigration issue. Ideological identification 3.559∗∗ 2.208∗∗ To further probe the causal role of modern sexism (0.753) (0.817) in the emergence of Vox, we conducted a placebo test by replicating the models in Tables 2 and 3, using the in- Nativism 0.734 3.881∗∗ tention to vote for the PP, the mainstream conservative (0.626) (0.724) party, as the outcome variable. In contrast to what we ob- served for the far right, switching to the mainstream right Territorial preference −1.433∗∗ −1.097∗ was unaffected neither by prior levels of sexism nor by (0.496) (0.538) prior changes in sexism (see Tables A6 and A7 in SI, pp. Populism 1.515† 1.070 13–14). This strengthens the interpretation that modern (0.784) (0.798) sexism matters when parties display a discourse reflecting modern sexism, as far-right parties often do. The test also Sexism 2.293∗∗ 1.377† shows that modern sexism was not a relevant predictor of (0.751) (0.823) vote choice in 2018 (i.e., before the rise of Vox), neither in terms of predispositions nor of changing attitudes. Constant −7.132∗∗ −7.191∗∗ (1.126) (1.302) Observations 1503 1360 Discussion Notes: The dependent variable is intended vote for Vox at time t. Logistic regression coefficients, standard errors in parentheses. The findings presented here show that modern sexism— † p < 0.1, ∗ p < 0.05, ∗∗ p < 0.01. which captures denial of discrimination against women, 15405907, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/ajps.12759 by Consorci De Serveis Universitaris De Catalunya, Wiley Online Library on [01/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License 12 EVA ANDUIZA and GUILLEM RICO TABLE 3 Effect of Prior Change in Attitudes and Vote Intention on Intended Vote for Vox 2019 2020 2019 2020 Female 0.041 −0.144 0.059 −0.144 (0.355) (0.313) (0.353) (0.312) Age −0.025 −0.032† −0.023 −0.032† (0.020) (0.018) (0.020) (0.018) High school / Vocational −0.548 −0.572 −0.597 −0.569 (0.597) (0.463) (0.603) (0.462) College 0.456 −0.249 0.469 −0.248 (0.398) (0.352) (0.395) (0.351) Income −1.069 0.271 −1.123 0.272 (0.705) (0.648) (0.707) (0.647) Lives with partner 0.046 −0.012 0.084 −0.011 (0.370) (0.316) (0.374) (0.315) Interest in politics 1.910∗∗ 1.220∗ 1.863∗∗ 1.218∗ (0.645) (0.509) (0.640) (0.509) Prior Values (t−2) Vox intention 2.975† 4.695∗∗ 3.181† 4.659∗∗ (1.577) (1.529) (1.632) (1.515) Authoritarianism 1.170 0.956 1.205 0.957 (0.753) (0.697) (0.759) (0.696) Ideological identification 5.326∗∗ 1.691† 5.177∗∗ 1.691† (1.100) (1.003) (1.102) (1.004) Nativism 0.062 4.083∗∗ 0.018 4.072∗∗ (0.963) (0.841) (0.958) (0.839) Territorial preference −1.266† −1.363∗ −1.212† −1.363∗ (0.709) (0.626) (0.705) (0.625) Populism 1.499 0.259 1.267 0.260 (1.119) (0.931) (1.118) (0.929) Sexism 2.908∗ 1.065 2.990∗ 1.064 (1.184) (0.941) (1.194) (0.939) Prior Change (t−2 minus t−1) Vox intention 4.749∗∗ 3.219∗∗ 4.309∗ 3.209∗∗ (1.840) (0.377) (1.719) (0.379) Authoritarianism 0.422 0.098 0.490 0.099 (0.768) (0.666) (0.766) (0.665) Ideological identification 2.679∗ 1.904∗ 2.639∗ 1.903∗ (1.208) (0.934) (1.205) (0.932) (Continued) 15405907, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/ajps.12759 by Consorci De Serveis Universitaris De Catalunya, Wiley Online Library on [01/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License THE FAR-RIGHT VOTE AND GENDER BACKLASH 13 TABLE 3 (Continued) 2019 2020 2019 2020 ∗∗ Nativism 0.861 3.448 0.829 3.437∗∗ (1.002) (0.820) (0.990) (0.819) Territorial preference −0.603 −0.918 −0.516 −0.913 (0.701) (0.621) (0.694) (0.622) Populism 1.033 1.426 0.954 1.425 (1.159) (0.905) (1.145) (0.904) Sexism 2.610∗ 0.801 (1.162) (1.034) Increase in sexism 5.037∗∗ 0.827 (1.788) (1.425) Decrease in sexism −0.190 0.657 (1.776) (2.079) Constant −7.764∗∗ −6.186∗∗ −7.925∗∗ −6.183∗∗ (1.561) (1.407) (1.574) (1.410) Observations 1092 1264 1092 1264 Notes: The dependent variable is intended vote for Vox at time t. Logistic regression coefficients estimated using penalized maximum likelihood, standard errors in parentheses. † p < 0.1, ∗ p < 0.05, ∗∗ p < 0.01. antagonism toward women’s demands, and resentment of predispositions; it must also include the analysis of at- against antidiscrimination policies—can be an impor- titudinal changes and their consequences for vote choice. tant predictor of far-right vote choice. Spanish panel data We have observed modest but significant short-term show that, as a preexisting attitude, modern sexism mat- changes in levels of sexism that contradict the notion of tered at least as much as other attitudes that are typically an unwavering predisposition. Overall, levels of sexism included in explanatory models of far-right vote choice. decreased after the wave of feminist demonstrations that We should no longer assume gender attitudes are ignor- took place in 2018 and increased after Vox entered par- able, negligible, or even secondary in explanations of far- liament in 2019. This is at odds with the stability that right support. Sexism deserves specific conceptual and sexism has shown in the U.S. context (Archer and Kam empirical attention for understanding the electoral rise 2020). These differences between the U.S. and Spain may of the far right. Our argument is not that modern sexism be due to factors such as the timing or the intensity of is the key predictor in all contexts but rather that differ- feminist mobilization, or to the use of different measures. ent types of sexism may matter in different contexts and But our evidence suggests that—even if only to a moder- that modern sexism should be carefully considered and ate extent—modern sexism can eventually be sensitive to measured in connection with far-right parties. context and show short-term changes, which should be Consistently with what previous work in the U.S. has the object of further scrutiny in future research. established, our results confirm that sexism can be acti- These attitudinal changes in turn may have relevant vated as well as deactivated. Sexism as a predisposition electoral consequences. We use internal individual varia- was particularly important when Vox appeared in 2019 tion in sexism not only to assess its effect on the far-right following a period of heightened saliency of feminist de- vote but also with a more contextualized perspective to mands and less so once the “feminist threat” was less assess what kind of changes in sexism affect vote choice. pressing, likely because of the COVID-19 pandemic and The two moments in which attitudinal change happens the very presence of the far-right itself as an anti-feminist in our study (before and after a feminist tide, and be- actor. But the analysis of the dynamic relationship be- fore and after a far-right party became a significant polit- tween sexism and vote choice goes beyond the activation ical actor), allow us to distinguish between backlash and 15405907, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/ajps.12759 by Consorci De Serveis Universitaris De Catalunya, Wiley Online Library on [01/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License 14 EVA ANDUIZA and GUILLEM RICO normalization attitudinal change. Backlash involves in- Akkerman, Tjitske, and Anniken Hagelund. 2007. “’Women creases in sexism that happen at a time of feminist mo- and Children First!’ Anti-Immigration Parties and Gender in Norway and the Netherlands.” Patterns of Prejudice 41(2): mentum, while normalization, in turn, involves increases 197–214. in sexism that happen at a time of increased far-right vis- Albarracin, Dolores, and Sharon Shavitt. 2018. “Attitudes and ibility. Our findings show that these backlash changes in Attitude Change.” Annual Review of Psychology 69(1): 299– sexism had important electoral consequences, while nor- 327. malization changes did not. Alonso, Sonia, and Cristóbal Rovira Kaltwasser. 2012. “The The overall decline in sexism after the feminist Spanish ’Exceptionalism’ In Western Europe: No Country movement of 2018 was rather heterogeneous, con- For The Populist Radical Right?” South European Society and Politics (February 2015): 1–50. ditioned by respondents’ characteristics, such as sex, Archer, Allison, and Cindy D., Kam. 2020. “Modern Sexism in partisanship, ideology, or engagement with the feminist Modern Times Public Opinion in the #Metoo Era.” Public demonstrations. This seems to indicate some realign- Opinion Quarterly 84(4): 813–37. ment or polarization around the issue. Those who were Arzheimer, Kai. 2018. “Explaining Electoral Support for the less likely to absorb the overflowing feminist messages Radical Right." In The Oxford Handbook of the Radical Right, of the moment or even perceived a threatening element Oxford: Oxford University Press. in the wave of feminist mobilization presumably reacted Becker, Julia, and Chris Sibley. 2016. “Sexism." In Handbook of with increasing levels of modern sexism. It was this atti- Stereotyping Prejudice and Discrimination, New York, Lon- don: Psychology Press. tudinal backlash in sexism that initially attracted voters Bernardez-Rodal, Asuncion, Paula Requeijo Rey, and Yanna who were previously comfortable voting for a moderate G. Franco. 2020. “Radical Right Parties and Anti-Feminist party to this new far-right party with an anti-feminist Speech on Instagram: Vox and the 2019 Spanish General discourse. Later on, sexism continued to change, but Election.” Party Politics 28(2): 272–83. these changes were not dependent on political predispo- Bock, Jarrod, Jennifer Byrd-Craven, and Melissa Burkley. 2017. sitions and did not evolve with a polarization pattern. “The Role of Sexism in Voting in the 2016 Presidential Election.” Personality and Individual Differences 119: 189– The increases in sexism that we witness once the far right 93. was on the scene seem to reflect some sort of generalized Burns, Nancy E., and Katherine Gallagher. 2010. “Public Opin- period effect, which although very much concerning in ion on Gender Issues: The Politics of Equity and Roles.” An- itself, did not have direct electoral consequences for the nual Review of Political Science 13: 425–43. far right. Cabezas, Marta. 2022. “Silencing Feminism? Gender and the Our analysis shows the complex micro processes Rise of the Nationalist Far Right in Spain.” Signs: Journal of through which feminist mobilization played an impor- Women in Culture and Society 47(2): 319–45. tant role in the rise of Vox and its breakthrough into par- Cassese, Erin, and Tiffany Barnes. 2019. “Reconciling Sexism and Women’s Support for Republican Candidates: A Look liament, which is consistent with a backlash hypothesis at Gender, Class, and Whiteness in the 2012 and 2016 Pres- that, although present in the literature, has been more idential Races.” Political Behavior 41(3): 677–700. often suggested than empirically tested. Further research Cassese, Erin, Tiffany Barnes, and Regina Branton. 2015. should elaborate on the individual consequences of fem- “Racializing Gender: Public Opinion at the Intersection.” inist mobilization, which seems to have the paradoxical Politics and Gender 11(1): 1–26. effect of initially reducing overall levels of sexism but also Cassese, Erin, and Mirya Holman. 2019. “Playing the Woman increasing the degree of polarization around the issue, Card: Ambivalent Sexism in the 2016 U.S. Presidential Race.” Political Psychology 40(1): 55–74. which in turn benefits the far right. Deckman, Melissa, and Erin Cassese. 2019. “Gendered Nation- alism and the 2016 US Presidential Election: How Party, Class, and Beliefs about Masculinity Shaped Voting Behav- References ior.” Politics and Gender 17(2): 277–300. Ditonto, Tessa. 2019. “Direct and Indirect Effects of Preju- dice: Sexism, Information, and Voting Behavior in Politi- Akkerman, Agnes, Cas Mudde, and Andrej Zaslove. 2014. cal Campaigns.” Politics, Groups, and Identities 7(3): 590– “How Populist Are the People? Measuring Populist Atti- 609. tudes in Voters.” Comparative Political Studies 49(9): 1324– 53. Dwyer, Caitlin E., Daniel Stevens, John L. Sullivan, and Barbara Allen. 2009. “Racism, Sexism, and Candidate Evaluations in Akkerman, Tjitske. 2005. “Anti-Immigration Parties and the the 2008 U.S. Presidential Election.” Analyses of Social Issues Defence of Liberal Values: The Exceptional Case of the List and Public Policy 9(1): 223–40. Pim Fortuyn.” Journal of Political Ideologies 10(3): 337–54. Edo, Anthony, Yvonne Giesing, Jonathan Öztunc, and Panu Akkerman, Tjitske. 2015. “Gender and the Radical Right in Poutvaara. 2019. “Immigration and Electoral Support for Western Europe: A Comparative Analysis of Policy Agen- the Far-Left and the Far-Right.” European Economic Review das.” Patterns of Prejudice 49(1–2): 37–60. 115: 99–143. 15405907, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/ajps.12759 by Consorci De Serveis Universitaris De Catalunya, Wiley Online Library on [01/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License THE FAR-RIGHT VOTE AND GENDER BACKLASH 15 Feldman, Stanley, and Karen Stenner. 1997. “Perceived Threat Knuckey, Jonathan. 2019. “’I Just Don’t Think She Has a and Authoritarianism.” Political Psychology 18(4): 741–70. Presidential Look’: Sexism and Vote Choice in the 2016 Fernández-Suárez, Belén. 2021. “Género e inmigración en Election∗.” Social Science Quarterly 100(1): 342–58. VOX :” Migraciones.” Publicación del Instituto Universitario Kosmidis, Ioannis, Euloge Clovis Kenne Pagui, Kjell Konis, and de Estudios sobre Migraciones (51): 241–68. Nicola Sartori. 2021. "Brglm2: Bias Reduction in General- Ferreira, Carles. 2019. “Vox Como Representante de La ized Linear Models.” https://CRAN.R-project.org/package= Derecha Radical En España : Un Estudio Sobre Su Ide- brglm2 (May 16, 2022). ología.” Revista Espanola de Ciencia Politica 51: 73–98. Köttig, Michaela, Renate Bitzan, and Andrea Pető, eds. 2017. Firth, David. 1993. “Bias Reduction of Maximum Likelihood Gender and Far Right Politics in Europe. Cham: Springer In- Estimates.” Biometrika 80(1): 27–38. ternational Publishing. Frasure-Yokley, Lorrie. 2018. “Choosing the Velvet Glove: Kováts, Eszter. 2017. “The Emergence of Powerful Anti-Gender Women Voters, Ambivalent Sexism, and Vote Choice in Movements in Europe and the Crisis of Liberal Democracy." 2016.” Journal of Race, Ethnicity, and Politics 3(1): 3–25. In Gender and Far Right Politics in Europe, Cham: Springer International Publishing, 175–89. http://link.springer.com/ Georgiadou, Vasiliki, Lamprini Rori, and Costas Roumanias. 10.1007/978-3-319-43533-6_12 (May 31, 2019). 2018. “Mapping the European Far Right in the 21st Cen- tury: A Meso-Level Analysis.” Electoral Studies 54: 103–15. Kováts, Eszter. 2018. “Questioning Consensuses: Right-Wing Populism, Anti-Populism, and the Threat of ’Gender Ide- de Geus, Roosmarijn, Elizabeth Ralph-Morrow, and Rosalind ology."” Sociological Research Online 23(2): 528–38. Shorrocks. 2022. “Understanding Ambivalent Sexism and Its Relationship with Electoral Choice in Britain.” British Kováts, Eszter, Maari Poim, and Andrea Pető. 2015. Gender as Journal of Political Science 52(4): 1564–83. Symbolic Glue: The Position and Role of Conservative and Far-Right Parties in the Anti-Gender Mobilizations in Europe. Glick, Peter, and Susan T. Fiske. 1997. “Hostile and Benevolent Brussels: Foundation for European Progressive Studies. Sexism.” Psychology of Women Quarterly 21(1): 119–35. Lenz, Gabriel S. 2012. Follow the Leader? How Voters Respond to Godbole, Maya A., Noelle A. Malvar, and Virginia V. Valian. Politicians’ Policies and Performance. Chicago: The Univer- 2019. “Gender, Modern Sexism, and the 2016 Election.” Pol- sity of Chicago Press. itics, Groups, and Identities 7(3): 700–12. Lépinard, Éléonore. 2020. Feminist Trouble. Oxford University Golder, Matt. 2016. “Far Right Parties in Europe.” Annual Re- Press. view of Political Science 19(1): 477–97. Lodders, Vanna, and Steven Weldon. 2019. “Why Do Women Green, Jane, and Rosalind Shorrocks. 2021. “The Gender Back- Vote Radical Right? Benevolent Sexism, Representation lash in the Vote for Brexit.” Political Behavior https://doi. and Inclusion in Four Countries.” Representation 55(4): org/10.1007/s11109-021-09704-y (August 3, 2021). 457–74. Grzebalska, Weronika, and Andrea Pető. 2018. “The Gendered Mansbridge, Jane, and Shauna L. Shames. 2008. “Toward a Modus Operandi of the Illiberal Transformation in Hun- Theory of Backlash: Dynamic Resistance and the Central gary and Poland.” Women’s Studies International Forum 68: Role of Power.” Politics and Gender 4(4): 623–34. 164–72. McThomas, Mary, and Michael Tesler. 2016. “The Growing In- Hanley, Eric. 2021. “Sexism as a Political Force: The Impact fluence of Gender Attitudes on Public Support for Hillary of Gender-Based Attitudes on the Presidential Elections of Clinton, 2008–2012.” Politics & Gender 12(01): 28–49. 2012 and 2016.” Social Science Quarterly 102(4): 1408–27. Mendes, Mariana S., and James Dennison. 2021. “Explaining Harteveld, Eelco, Tim Immerzeel, Hilde Coffé, and Tanja van the Emergence of the Radical Right in Spain and Portugal: der Lippe. 2015. “The Gender Gap in Populist Radical-Right Salience, Stigma and Supply.” West European Politics 44(4): Voting: Examining the Demand Side in Western and Eastern 752–75. Europe.” Patterns of Prejudice 49(1–2): 103–34. Mo, Cecilia Hyunjung. 2015. “The Consequences of Explicit Hayes, Bernadette C. 1997. “Gender, Feminism and Electoral and Implicit Gender Attitudes and Candidate Quality in the Behaviour in Britain.” Electoral Studies 16(2): 203–16. Calculations of Voters.” Political Behavior 37(2): 357–95. Hayes, Bernadette C, Ian McAllister, and Donley T. Studlar. Möser, Cornelia. 2020. “Sexual Politics as a Tool to "un- 2000. “Gender, Postmaterialism and Feminism in Com- Demonize Right-Wing Discourses in France." In Right- parative Perspective.” International Political Science Review Wing Populism and Gender: European Perspectives and Be- 21(4): 425–39. yond, Bielefield: transcript Verlag. Hernández Pérez, Enrique et al. 2021. "POLAT Project. Spanish Mudde, Cas. 2019. The Far Right Today. Cambridge: Polity Political Attitudes Panel Dataset (Waves 1–6).” https://ddd. Press. uab.cat/record/243399 (September 13, 2021. Norris, Pippa, and Ronald Inglehart. 2019. Cultural Backlash Hopkins, Daniel J. 2021. “The Activation of Prejudice and Pres- and the Rise of Populism: Trump, Brexit, and Authoritarian idential Voting: Panel Evidence from the 2016 U.S. Elec- Populism. Cambridge University Press. tion.” Political Behavior 43(2): 663–86. Olalla, Sergio, Enrique Chueca, and Javier Padilla. 2019. Ignazi, Piero. 2003. Extreme Right Parties in Western Europe. “Spain Is No Longer Exceptional: Mainstream Media Oxford ; New York: Oxford University Press. and the Far-Right Party Vox.” Euro Crisis in the Press Kitschelt, Herbert. 1995. The Radical Right in Western Europe: https://blogs.lse.ac.uk/eurocrisispress/2019/01/10/spain-is- A Comparative Analysis. Ann Arbor: University of Michigan no-longer-exceptional-mainstream-media-and-the-far- Press. right-party-vox/(July 26, 2021). 15405907, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/ajps.12759 by Consorci De Serveis Universitaris De Catalunya, Wiley Online Library on [01/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License 16 EVA ANDUIZA and GUILLEM RICO Paternotte, David. 2020. “Backlash: A Misleading Narra- Stockemer, Daniel, Tobias Lentz, and Danielle Mayer. tive.” Engenderings,LSE https://blogs.lse.ac.uk/gender/2020/ 2018. “Individual Predictors of the Radical Right-Wing 03/30/backlash-a-misleading-narrative/(February 1, 2022). - ProQuest.” Government and Opposition 53(3): 569– Pető, Andrea. 2015. “’Anti-Gender’ Mobilisational Discourse 93. of Conservative and Far Right Parties as a Challenge for Swim, Janet K., Kathryn J. Aikin, Wayne S. Hall, and Barbara Progressive Politics"." In Gender as Symbolic Glue, Friedrich A. Hunter. 1995. “Sexism and Racism: Old-Fashioned and Ebert Siftung. Modern Prejudices.” Journal of Personality and Social Psy- Rama, José, Lisa Zanotti, Stuart J. Turnbull-dugarte, and An- chology 68(2): 199–214. drés Santana. 2021. VOX: The Rise of the Spanish Populist Swim, Janet K., and Laurie L. Cohen. 1997. “Overt, Covert, and Radical Right. Routledge. Subtle Sexism.” Psychology of Women Quarterly 21(1): 103– Ratliff, Kate A., Liz Redford, John Conway, and Colin Tucker 18. Smith. 2019. “Engendering Support: Hostile Sexism Pre- Swim, Janet K., and Lauri L. Hyers. 2009. “Sexism.” In Hand- dicts Voting for Donald Trump over Hillary Clinton in the book of Prejudice, Stereotyping, and Discrimination, New 2016 U.S. Presidential Election.” Group Processes and Inter- York, NY, US: Psychology Press, 407–30. group Relations 22(4): 578–93. Turnbull-Dugarte, Stuart J. 2019. “Explaining the End of Span- Rhodes, Jesse H., Elizabeth A. Sharrow, Jill S. Greenlee, and ish Exceptionalism and Electoral Support for Vox.” Research Tatishe M. Nteta. 2020. “Just Locker Room Talk? Explicit and Politics 6(2): 1–8. Sexism and the Impact of the Access Hollywood Tape on Valentim, Vicente. 2021. “Parliamentary Representation and Electoral Support for Donald Trump in 2016.” Political the Normalization of Radical Right Support.” Comparative Communication 37(6): 741–67. Political Studies 54(14): 2475–511. Sanbonmatsu, Kira. 2008. “Gender Backlash in American Poli- Valentino, Nicholas A, Carly Wayne, and Marzia Oceno. 2018. tics?” Politics and Gender 4(4): 634–42. “Mobilizing Sexism: The Interaction of Emotion and Gen- Scarborough, William J., Ray Sin, and Barbara Risman. 2019. der Attitudes in the 2016 US Presidential Election.” Public “Attitudes and the Stalled Gender Revolution: Egalitarian- Opinion Quarterly 82(suppl_1): 213–35. ism, Traditionalism, and Ambivalence from 1977 through Wodak, Ruth, Jonathan Culpeper, and Elena Semino. 2021. 2016.” Gender & Society 33(2): 173–200. “Shameless Normalisation of Impoliteness: Berlusconi’s and Schäfer, Armin. 2021. “Cultural Backlash? How (Not) to Ex- Trump’s Press Conferences.” Discourse & Society 32(3): plain the Rise of Authoritarian Populism.” British Journal of 369–93. Political Science: 1–17. Zhirkov, Kirill. 2014. “Nativist but Not Alienated: A Compar- Schaffner, Brian F. 2021. “Optimizing the Measurement of Sex- ative Perspective on the Radical Right Vote in Western Eu- ism in Political Surveys.” Political Analysis: 1–17. rope.” Party Politics 20(2): 286–96. Schaffner, Brian F., Matthew Macwilliams, and Tatishe Nteta. 2018. “Understanding White Polarization in the 2016 Vote for President: The Sobering Role of Racism and Sexism.” Political Science Quarterly 133(1): 9–34. Supporting Information Schwab, Klaus et al. 2019. Global Gender Gap Report 2020 In- sight Report. Geneva: World Economic Forum. Selvanathan, Hema Preya, and Bernhard Leidner. 2021. "Nor- Additional supporting information may be found online malization of the Alt-Right: How Perceived Prevalence and in the Supporting Information section at the end of the Acceptability of the Alt-Right Is Linked to Public Attitudes." article. Group Processes & Intergroup Relations Online first: 1–22. Simas, Elizabeth N., and Marcia Bumgardner. 2017. “Modern Appendix A. The Spanish Political Attitudes Dataset Sexism and the 2012 U.S. Presidential Election: Reassessing Appendix B. Panel Attrition the Casualties of the War on Women.” Politics and Gender Appendix C. Measurement 13(3): 359–78. Appendix D. Distribution of Sexist Attitudes Spierings, Niels, and Andrej Zaslove. 2015. “Gendering the Appendix E. Multilevel Growth Curve Models Vote for Populist Radical-Right Parties.” Patterns of Preju- dice 49(1–2): 135–62. Appendix F. Placebo Tests