How polarization, populist attitudes, and cultural backlash affect citizens’ support for democracy: Evidence from Spain

Affective polarization, populism, and cultural battles as attitudinal orientations

How are affective polarization, populism, and cultural backlash battles attitudinally connected? A review of the literature on the three phenomena led us to identify several common roots among them at the attitudinal level. Starting with polarization, scholars warn that we are today witnessing an increased affective sorting of citizens around party lines, which is fuelling levels of polarization among voters of contending parties. As typically defined, affective polarization refers to the distance between the affection or positive feelings individuals have toward parties, leaders, and followers whose ideas or political positions they share, and the rejection of parties, leaders, and followers who advocate opposing ideas and opinions (Iyengar et al., 2019Iyengar, Shanto, YphtachLelkes, MatthewLevendusky, NeilMalhotra, and Sean J.Westwood. 2019. “The Origins and Consequences of Affective Polarization in the United States.” Annual Review of Political Science 22: 129–146. doi:10.1146/nnurev-polisci-051117-073034). Therefore, affective polarization has two defining components, namely “in-group” affection, i.e. the positive feelings toward their own party, its leaders, and followers, and “out-group” polarization, i.e. the negative feelings toward opposing parties, its leaders, and followers.

The literature on affective polarization has discussed two types of mechanisms that contribute to explaining increasing levels of animosity toward political opponents in consolidated democracies. The first mechanism explains affective polarization as the result of the convergence of partisan identities with many other social identities. As West and Iyengar (2022West, Emily A., and ShantoIyengar. 2022. “Partisanship as a social identity: Implications for Polarization.” Political Behavior 44(2): 807–838. doi: 10.1007/s11109-020-09637) argue, partisanship is no longer exclusively synonymous with political preferences, but has become an important social identity which, ultimately, contributes to the strengthening of affective polarization. While partisan and ideological identities have traditionally been aligned, other prominent social identities, such as race, religion, or territorial identity are also increasingly converging into partisanship (Abramowitz and McCoy, 2019Abramowitz, Alan, and JenniferMcCoy. 2019. “United States: Racial Resentment, Negative Partisanship, and Polarization in Trump’s America.” The Annals of the American Academy of Political and Social Science 681(1): 137–156. doi:10.1177/0002716218811309; Iyengar et al., 2019Iyengar, Shanto, YphtachLelkes, MatthewLevendusky, NeilMalhotra, and Sean J.Westwood. 2019. “The Origins and Consequences of Affective Polarization in the United States.” Annual Review of Political Science 22: 129–146. doi:10.1146/nnurev-polisci-051117-073034). For instance, in the US, Mason (2018Mason, Lilliana. 2018. Uncivil Agreement: How Politics Became Our Identity. Chicago: University of Chicago Press.) shows that while white evangelicals overwhelmingly support the Republican Party, African Americans tend to identify with the Democratic Party. As the author discusses, this gradual process of alignment of social identities along partisan lines has turned partisanship into a sort of social identity (social sorting), thus making it easier for partisans to infer generalizations about people on the opposite side, even if such inferences are wrong. In the case of Spain, Torcal (2023Torcal, Mariano. 2023. De votantes a hooligans. La polarización política en España. Madrid: Los Libros de la Catarata.) argues that citizens have built partisan mega-identities, around which several social and ideological identities converge. These mega-identities have been shaped, firstly, around ideological identities of left and right, which have created two opposing blocks of partisans, and secondly, based on territorial identities, which also favour greater social distancing.

The second account of the rise of affective polarization identifies the perception of a growing ideological distance between parties as its main cause. Thus, recent comparative and experimental studies provide evidence that there is a relationship between ideological and affective polarization, meaning that as the parties are becoming more ideologically distant from each other, voters are too, and, hence, they are increasingly willing to display hostility toward those whom they see as ideological opponents (Rogowski and Sutherland, 2016Rogowski, Jon C., and Joseph L.Sutherland. 2016. “How Ideology Fuels Affective Polarization.” Political Behaviour 38(2): 485–508. doi: 10.1007/s11109-015-9323-7; Webster and Abramowitz, 2017Webster, Steven W., and Alan I.Abramowitz. 2017. “The Ideological Foundations of Affective Polarization in the US Electorate.” American Politics Research 45(4): 621–647. doi:10.1177/1532673X17703132). In the same vein, in Spain, studies have found a positive relationship between both types of polarization, so that “the greater the ideological polarization of the parties (measured as the distance in ideological positions perceived by voters), the greater the affective polarization of individuals” (Orriols and León, 2020Orriols, Lluís, and SandraLeón. 2020. “Looking for Affective Polarisation in Spain: PSOE and Podemos from Conflict to Coalition.” South European Society and Politics 25(3-4): 351–379. doi:10.1080/13608746.2021.1911440: 12).

These two accounts of the causes of (rising) polarization can indeed be helpful in understanding the emergence of populist parties in western democracies. Given that the populist agenda is typically adopted by parties located at opposite ends of the ideological spectrum, that is, radical-right and radical-left parties, they benefit from the greater ideological sorting and distancing that national electorates are experiencing today. The relationship between polarization and populism indeed works in a virtuous cycle, as the current scenario of greater ideological and social polarization paves the way for the emergence and electoral success of populist parties, and at the same time, populist parties, once on this stage, fuel and receive high levels of affective polarization against and from (mainstream) political opponents (Harteveld et al., 2022Harteveld, Eelco, PhilippMendoza, and MatthijsRooduijn. 2022. “Affective Polarization and the Populist Radical Right: Creating the Hating?” Government and Opposition 57(4): 703–727. doi:10.1017/gov.2021.31; Davis et al., 2024Davis, Braeden, JayGoodliffe, and KirkHawkins. 2024. “The Two-Way Effects of Populism on Affective Polarization.” Comparative Political Studies 0(0): 1–33. doi:10.1177/00104140241237453). Indeed, as Torcal and Comellas (2022Torcal, Mariano, and Josep M.Comellas. 2022. “Affective Polarisation in Times of Political Instability and Conflict. Spain from a Comparative Perspective.” South European Society and Politics 27(1): 1–26. doi:10.1080/13608746.2022.2044236: 12) have shown, in Spain the emergence of the populist radical-right party Vox was preceded by the existence of high levels of affective polarization that had been produced in the system by the other parties, thus proving that radical right-wing parties tend to prosper and gain visibility in contexts of high polarisation. Bischof and Wagner (2019Bischof, Daniel, and MarkusWagner. 2019. “Do Voters Polarize When Radical Parties Enter Parliament?” American Journal of Political Science 63(4): 888–904. doi:10.1111/ajps.12449), using data from 17 European countries between 1973 and 2016, demonstrated that radical-right parties gaining representation in parliament tends to produce a long-term impact on levels of affective polarization within the electorate. In this sense, populist parties tend to both benefit from, and produce, polarization within the political system.

The demand side of populism has also been connected to increasing levels of polarization. According to this view, populism is not only related to what populist parties offer to the public and how they behave (the supply side), but also to the ideas and attitudes of individuals who demand populism (the demand side). In this sense, populism is defined as a set of ideas or ideology (the ideational approach), which is thin-centered, that is, populism cannot stand on its own but must be attached to left- or right-wing ideologies (Mudde, 2004Mudde, Cas. 2004. “The Populist Zeitgeist.” Government and Opposition 39(4): 542–563. doi:10.1111/j.1477-7053.2004.00135; Mudde and Rovira Kaltwasser, 2017Mudde, Cas, and CristobalRovira Kaltwasser. 2017. Populism: A Very Short Introduction. New York: Oxford University Press.). According to Mudde (2004Mudde, Cas. 2004. “The Populist Zeitgeist.” Government and Opposition 39(4): 542–563. doi:10.1111/j.1477-7053.2004.00135), populist parties, both left- and right-wing, share a common understanding of the world, democracy, and political representation, which is based on three essential components: (1) it is people-centered and anti-elite, (2) it entails a clash between the pure people and the corrupt elite, and (3) it claims that politics should be an expression of the general will of the people.

This ideational vision of populism also has the effect of increasing polarization by constructing an anti-establishment political frontier and contesting some of democracy’s basic institutional and procedural norms, such as political representation through parties (Roberts, 2022Roberts, Kenneth M.2022. “Populism and Polarization in Comparative Perspective: Constitutive, Spatial and Institutional Dimensions.” Government and Opposition 57(4): 680–702. doi:10.1017/gov.2021.14: 680). As Roberts (2022Roberts, Kenneth M.2022. “Populism and Polarization in Comparative Perspective: Constitutive, Spatial and Institutional Dimensions.” Government and Opposition 57(4): 680–702. doi:10.1017/gov.2021.14) argues, populism has a “constitutive” logic vis-à-vis affective polarization, in that it is intrinsic to populism to construct an antagonistic frontier between “the people” and a power elite or establishment, which makes populism intrinsically polarizing. Thus, holding populist attitudes is also potentially connected to higher levels of affective polarization, presumably against mainstream political parties or any other body that breaks down the unified popular subject.

Polarization and populism are also closely related to the third phenomena under analysis, namely cultural backlash. A prominent explanation of the reemergence of populism in western democracies over the past decades has connected it to the conformation of a new societal conflict, which has come to substitute the traditional cleavages, between defenders of materialist or authoritarian values on the one hand, and the champions of more liberal or libertarian views on the other (Norris and Inglehart, 2019Norris, Pippa, and RonaldInglehart. 2019. Cultural Backlash. Trump, Brexit, and Authoritarian Populism. Cambridge: Cambridge University Press.). As Norris and Inglehart (2019Norris, Pippa, and RonaldInglehart. 2019. Cultural Backlash. Trump, Brexit, and Authoritarian Populism. Cambridge: Cambridge University Press.: 32) argue, this new cleavage has its origins in the “silent revolution” theory of values change, which contends that the levels of existential security achieved by citizens of consolidated western societies in the post-war period led to an intergenerational shift towards more post-materialist values, such as cosmopolitanism and multiculturalism, and support for new issues such as environmental protection, human rights, and gender equality. However, in recent decades, the impact of globalization in labour markets and the mobility of workers has significantly increased the economic and social insecurity of many citizens, especially those who are socioeconomically disadvantaged but who, for a long time, had enjoyed a hegemonic social status. Thus, the cultural backlash theory suggests that the so-called “losers of globalization” (Teney et al., 2014Teney, Céline, OnawaPromise Lacewell, and Pieter DeWilde. 2014. “Winners and Losers of Globalization in Europe: Attitudes and Ideologies.” European Political Science Review 6(4): 575–595. doi:10.1017/S1755773913000246; Kriesi et al., 2012Kriesi, Hanspeter, EdgarGrande, MartinDolezal, MarcHelbling, DominicHoeglinger, SwenHutter, and BrunoWüest. 2012. Political Conflict in Western Europe. New York: Cambridge University Press.), characterized primarily as elderly, low-educated white males, have staged a backlash against the advancement of progressive values and the displacement of traditional family values. These socially conservative values typically embrace a broad rejection of the “Other”, ranging from immigrants, Muslims, black people, LGBTQ, ethnic minorities, and feminists, among others, and advocate the maintenance of a non-egalitarian status quo by opposing any social transformation that might contribute to social change, such as gender equality or the fight against climate change.

According to Norris and Inglehart (2019Norris, Pippa, and RonaldInglehart. 2019. Cultural Backlash. Trump, Brexit, and Authoritarian Populism. Cambridge: Cambridge University Press.), the progressive value change experienced by modern societies has led socially conservative individuals—i.e. those who feel threatened by the breakdown of the status quo—to transfer their votes to (authoritarian)right-wing populist parties, who promise to defend the old order. In this sense, the emergence of radical-right populism in most countries across western Europe and the recent success of right-wing populist leaders such as Jair Bolsonaro in Brazil or Donald Trump in the US (Norris and Inglehart, 2019Norris, Pippa, and RonaldInglehart. 2019. Cultural Backlash. Trump, Brexit, and Authoritarian Populism. Cambridge: Cambridge University Press.; Wejnert, 2023Wejnert, Barbara. 2023. “Cultural Backlash: The Long-Term Damage of Trump’s Legacy to American Democracy and Global Politics.” Pp. 73–94 in The Perils of Populism: The End of the American Century?, edited by AdebowaleAkande. Cham: Springer Nature Switzerland.; Rennó, 2020Rennó, Lucio R.2020. “The Bolsonaro Voter: Issue Positions and Vote Choice in the 2018 Brazilian Presidential Elections.” Latin American Politics and Society 62(4): 1–23. doi:10.1017/lap.2020.13) can be seen as the result of a socially conservative counter-revolution prompted by the expansion of cultural liberalism. At the same time, radical-right populist parties have embraced the defence of conservative values as a hallmark, thus sharpening the salience of the cultural cleavage in party competition and fostering polarization among those endorsing socially conservative and socially liberal values. To be clear, however, it is important to note that populism is not essentially connected to a right- or left-wing niche ideology, given that supply-side populism associates with radical platforms both of the left and the right, and not all populist parties advocate for conservative cultural values. Indeed, radical-left parties are advocates of radical liberal values, such as radical feminism and elimination of national frontiers, and their criticism toward democracy tends to focus on the political and/or economic elites within that system rather than the whole liberal system as such (Rooduijn and Akkerman, 2017Rooduijn, Matthijs, and TjitskeAkkerman. 2017. “Flank Attacks: Populism and Left-Right Radicalism in Western Europe.” Party Politics 23(3): 193–204. doi:10.1177/1354068815596514).

Hypotheses

The review of the literature above has shown that support for populist attitudes, cultural backlash, and affective polarization have common attitudinal roots and display mutual connections with each other. From a cognitive perspective, the three phenomena share a common propensity to divide the world into in-groups and out-groups, fuelling a binary classification of the world which might permeate across different domains. Thus, populism hinges on the division between “the good people” and a vilified elite (Mudde, 2004Mudde, Cas. 2004. “The Populist Zeitgeist.” Government and Opposition 39(4): 542–563. doi:10.1111/j.1477-7053.2004.00135); cultural backlash, and more particularly its two core manifestations of nativism and sexism concern the antagonism between natives and non-natives (Mudde, 2007Mudde, Cas. 2007. Populist Radical Right Parties in Europe. Cambridge: Cambridge University Press.), and between men and (feminist) women (Anduiza and Rico, 2022Anduiza, Eva, and GuillemRico. 2022. “Sexism and the Far-Right Vote: The Individual Dynamics of Gender Backlash.” American Journal of Political Science 68(2): 478–493. doi:10.1111/ajps.12759), respectively; and affective polarization reflects the tendency for an individual to strongly like their own party and strongly dislike the opposing parties (Iyengar et al., 2019Iyengar, Shanto, YphtachLelkes, MatthewLevendusky, NeilMalhotra, and Sean J.Westwood. 2019. “The Origins and Consequences of Affective Polarization in the United States.” Annual Review of Political Science 22: 129–146. doi:10.1146/nnurev-polisci-051117-073034). In this sense, there is a conventional assumption in the literature that the three attitudes tend to be combined at the attitudinal level.

Yet, the three attitudinal orientations are not necessarily combined in all circumstances. Indeed, for instance, extreme-left populists, such as supporters of Unidas Podemos, are unlikely to hold nativist or sexist attitudes. Similarly, someone who holds strong conservative values is not necessarily affectively polarized as well. Thus, we expect to find a subgroup of individuals in our sample who hold opinions consistently in favour of populist principles, cultural backlash, and affective polarization, and others some combination thereof. However, we expect only the combination of strong populist attitudes, cultural backlash values, and strong affective polarization to be detrimental for democratic legitimacy. Thus, we test the following hypothesis:

Independent variables

The dependent variable: support for democracy.

The main dependent variable of our study is support for the ideal of democracy as the best form of government. This indicator taps into what has been termed in the literature as diffuse (Easton, 1975Easton, David1975. “A Re-Assessment of the Concept of Political Support.” British Journal of Political Science 5(4): 435–457. doi:10.1017/S0007123400008309) or “idealistic” (Inglehart, 2003Inglehart, Ronald. 2003. “How Solid Is Mass Support for Democracy—And How Can We Measure It?” PS: Political Science and Politic 36(1): 51–57. doi:10.1017/S1049096503001689) support for democracy, that is, an attitudinal positive predisposition toward democracy itself, independently of its particular day-to-day performance, as opposed to supporting other non-democratic or illiberal forms of government under certain circumstances, such as a strong leader who might make his own decisions without having to resort to elections. To be sure, this way of measuring regime preferences could be expressing no more than lip service, as respondents might be conveying support for democracy as a valence issue—i.e. something positive and valuable—though devoid of any specific content (Schedler and Sarsfield, 2007Schedler, Andreas, and RodolfoSarsfield. 2007. “Democrats with Adjectives: Linking Direct and Indirect Measures of Democratic Support.” European Journal of Political Research 46(5): 637–659. doi: 10.1111/j.1475-6765.2007.00708: 638-9). In other words, respondents might express support for democracy as an abstract principle, while in practice they might reject some of its underpinning values, such as minority rights or media freedom. In this sense, in recent years, scholars have suggested that in order to better grasp what people really mean by the word “democracy” when they express support for it, more detailed batteries of questions measuring support for specific values and institutions of democracy should be used (see e.g. Classen et al., 2023Claassen, Christopher et al.2023. “Conceptualizing and measuring support for democracy: A new approach.” Working Paper.; Ferrín and Kriesi, 2016Ferrín, Mónica, and HanspeterKriesi (Eds.). 2016. How Europeans View and Evaluate Democracy. Oxford: Oxford University Press., 2025Ferrín, Mónica, and HanspeterKriesi (Eds.). 2025. How Europeans View and Evaluate Democracy Revisited. Ten years later. Oxford: Oxford University Press.). Notwithstanding the narrower understanding of democratic legitimacy that our survey indicator conveys, it is useful to understand whether individuals are indeed supportive of the principle of democracy at any time and/or under any circumstance.

To measure abstract democratic support, respondents had to position themselves on a 0 to 10 scale, where 0 represents unconditional support for democracy and 10 support for a strong leader. The survey statements read as follows:

To facilitate the interpretation of the results, the variable of democratic support has been reversed, so that higher values represent support for the democratic ideal, whereas lower values represent support for a strong leader.

In terms of methods, we deploy a two-fold analysis. First, we use exploratory factor analysis to explore whether our survey items effectively measure the three underlying dimensions they seek to encapsulate, namely populist attitudes, cultural backlash values, and affective polarization. Next, we test the effect of the three latent factors on support for democracy as the best form of government by means of an OLS analysis. Second, beyond knowing whether the three underlying factors do exist independently from one another, we seek to analyze how they are connected at the attitudinal level, that is, how individuals can be classified according to their attitudes toward (ideational) populism, cultural backlash values, and affective polarization. To this end, we implement a latent class analysis (LCA). Finally, we analyze the extent to which membership in each social subgroup identified by the LCA can be predicted by their higher level of openness to support a non-democratic form of government using OLS.

The association between populist attitudes, cultural backlash, affective polarization, and democratic support

We start our analysis by exploring whether the seven items included in our survey effectively measure the three underlying attitudinal orientations they purport to capture—namely populist attitudes, support for cultural backlash values, and affective polarization—and how they are connected to democratic legitimacy. In other words, do populist attitudes, cultural backlash, and affective polarization exist as three analytically different attitudes and do they all have the same (negative) impact on democratic legitimacy?

For that purpose, we first conduct an exploratory factor analysis following principal component factoring and oblique rotation. In doing so, we include the seven individual items on populist attitudes, support for conservative cultural values, and affective polarization, without specifying to which dimension each item (theoretically) “belongs”. As can be seen from Table 1, the results show that the different items load highly (>0.6) on the corresponding factors. Furthermore, there are no cross-loadings, as none of the items load sufficiently (>0.5) on the other dimensions. We can therefore clearly conclude that our full list of items can effectively be summarized into three separate components, namely populist attitudes, cultural backlash values, and affective polarization.

Figure 1 evaluates graphically whether an association exists between each of the latent factors estimated from the previous analysis and democratic support. The graph plots each of the latent dimensions on populist attitudes, cultural backlash, and affective polarization against the variable of democratic support and we add a linear fit with confidence intervals to the latter to better grasp the existing association between the variables. Panel A shows that, contrary to expectations, populist attitudes are positively related to respondents’ support for democracy. By contrast, the latent variable of cultural backlash (Panel B) shows a negative relationship with democratic support, whereas affective polarization (Panel C) appears to be unrelated to democratic legitimacy.

Figure 1 Relationship between populist attitudes, cultural backlash, affective polarization, and democratic support. 

To cross-test these bivariate relationships, we run a linear regression model (OLS) of the three latent variables on populist attitudes, cultural backslash, and affective polarization on democratic support. The model controls for respondents’ sex, age, level of education, and ideology. Figure 2 presents the OLS coefficients of the regression analysis.²Table A.1 in the Appendix presents the full regression output. The results show that the two latent factors on support for cultural backlash values and populist attitudes have a statistically significant effect on support for democracy as the best form of government, though in opposite directions. In line with our expectations and the bivariate analysis above, support for cultural backlash values has a negative effect on democratic support, which means that, other socio-demographic traits being equal, those who score lower on the cultural backlash latent factor tend to display a higher level of democratic support—or, to put it another way, a greater preference for conservative cultural battles leads individuals to hold undemocratic preferences of governance. Similarly in line with the results of the bivariate analysis, the effect of the latent factor on populist attitudes on democratic support is positive, thus indicating that respondents with stronger populist attitudes tend to be more supportive of democracy than those with weaker populist ideas. While this latter finding runs counter to our theoretical expectations, it is in line with previous evidence which has found populism to have a positive effect on support for the principle of democracy (Zaslove and Meijers, 2023Zaslove, Andrej, and MauritsMeijers. 2023. “Populist Democrats? Unpacking the Relationship Between Populist and Democratic Attitudes at the Citizen Level.” Political Studies 0(0). https://doi:10.1177/00323217231173800; Zaslove et al., 2021Zaslove, Andrej, BramGeurkink, KristofJacobs, and AgnesAkkerman. 2021. “Power to the People? Populism, Democracy, and Political Participation: A Citizen’s Perspective.” West European Politics 44(4): 727–751. doi:10.1080/01402382.2020.1776490; Kaltwasser and Hauwaert, 2020Kaltwasser, Cristóbal Rovira, and Steven M.Van Hauwaert. 2020. “The Populist Citizen: Empirical Evidence from Europe and Latin America.” European Political Science Review 12 (1): 1–18. doi:10.1017/S1755773919000262), and/or that populism has no effect on the rejection of key tenets of liberal democracy, such as political pluralism (Ellenbroek et al., 2023Ellenbroek, Victor, Maurits J.Meijers, and AndréKrouwel. 2023. “Populist but Pluralist? Populist Attitudes and Preferences for Political Pluralism in Parliament and Government.” Parliamentary Affairs 76(1): 125–145. doi:10.1093/pa/gsab041), the rule of law (Wuttke et al., 2023Wuttke, Alexander, ChristianSchimpf, and HaraldSchoen. 2023. “Populist Citizens in Four European Countries: Widespread Dissatisfaction Goes with Contradictory but Pro-democratic Regime Preferences.” Swiss Political Science Review 29(2), 246–257. doi:10.1111/spsr.1254), minority rights, and the separation of powers (Zaslove and Meijers, 2023Zaslove, Andrej, and MauritsMeijers. 2023. “Populist Democrats? Unpacking the Relationship Between Populist and Democratic Attitudes at the Citizen Level.” Political Studies 0(0). https://doi:10.1177/00323217231173800), to name but a few. As Zaslove and Meijers (2023Zaslove, Andrej, and MauritsMeijers. 2023. “Populist Democrats? Unpacking the Relationship Between Populist and Democratic Attitudes at the Citizen Level.” Political Studies 0(0). https://doi:10.1177/00323217231173800: 19) discuss, citizens with more populist attitudes even seem to be better or more passionate democrats than those with lower populist attitudes, i.e. they tend to adhere to the democratic ideal more intensely, but this can be explained by their need to highlight their despair about the way status quo practices subvert those principles and/or their willingness to profile themselves as purer democrats. However, as these authors point out, even when individuals with stronger populist attitudes tend to strongly support democracy in principle, they are highly critical of mediated forms of political representation which are essential to liberal democracy, such as political parties and judges, and think that majority rule should not be constrained in any way. However, as pointed out above, this contradictory support for the ideal of democracy while rejecting some of its core underpinning institutions cannot be captured by our survey item.

Figure 2 OLS coefficients of populist attitudes, cultural backlash, and affective polarization on democratic support. 

To finish, we see that the effect of affective polarization is non-significant at conventional levels, so that there are no relevant differences in levels of democratic support between individuals who are affectively polarized and those who are not polarized. Thus, at least when considering individuals’ support for democracy in principle, there are no reasons to think that the increasing levels of affective polarization that studies are identifying today are endangering democratic legitimacy in Spain.

Latent class analysis

 ⌅

In the previous section we have reported the effect that populist attitudes, cultural backlash, and affective polarization have on democratic support when they are considered as analytically distinct attitudes. However, as we have seen in the theoretical section above, the three attitudes share common roots and tend to mutually influence each other, so that they are often combined in practice. In this section, therefore, we focus on exploring the social groups that exist in the Spanish electorate according to their attitudes toward populism, cultural backlash, and like/dislike toward out-group parties. By means of this analysis we seek to identify whether the prototypical voter who simultaneously combines strong populist views, support for conservative cultural values, and high levels of affective polarization can be clearly identified, and whether membership in this social subgroup can be predicted by their greater openness to support a non-democratic form of government.

To this end, we implement a latent class analysis (LCA) (Magidson and Vermunt, 2004Magidson, Jay and Jeroen K.Vermunt. 2004. “Latent Class Models.” Pp. 549–553 in The Sage Handbook of Quantitative Methodology for the Social Sciences, edited by D.Kaplan. London: Sage.; Weller et al., 2020Weller, Bridget E., Natasha K.Bowen, and Sarah J.Faubert. 2020. “Latent Class Analysis: A Guide to Best Practice.” Journal of Black Psychology 46(4): 287–311. doi:10.1177/0095798420930932). This technique is used to identify different subgroups within a population that share certain characteristics which are recognized based on the similarity of the subgroup’s responses to the survey questions. Our latent class analysis involves two steps. First, we take the seven original survey items on populist attitudes, cultural backlash, and affective polarization and identify latent subgroups of respondents that share similar responses, and then present the size of each subgroup. Next, we test whether democratic support can be a predictor of membership in any of the subgroups, particularly in the subgroup of respondents that shares favourable views toward populism and cultural backlash and is affectively polarized. We also identify other socio-demographic traits that define each subgroup.

The results of the LCA in the seven items on populist attitudes, cultural backlash, and affective polarization uncover four subgroups of respondents in our population. To delimit the number of groups, we have considered both the Bayesian Information Criterion (BIC) and the theoretical interpretability of the groups (Weller et al., 2020Weller, Bridget E., Natasha K.Bowen, and Sarah J.Faubert. 2020. “Latent Class Analysis: A Guide to Best Practice.” Journal of Black Psychology 46(4): 287–311. doi:10.1177/0095798420930932). Figures A1–A2 in the Appendix show that the evidence basically remains unchanged when we apply a three-class or a five-class model.

Figure 3 plots the marginal means of the seven observed variables, conditional on being in the four classes identified. Table 2, below, displays the proportion of respondents in each latent subgroup. The first group (G1) presents strong support for the two populist attitudes and moderated adherence to cultural conservative values, although they are not affectively polarized against other parties and leaders and display only a minor opposition to their children getting married to someone from the out-party group. A potential explanation for the (almost) nonexistent levels of affective polarization of this group could be that the members of this group have a lower level of party identification; hence, we have named G1 ‘Non-partisan conservative populists.’ This group constitutes 15% of the sample, as Table 2 shows. The second group (G2) shares the same views as the first group, but more intensely in all the variables and quite particularly in the items on affective polarization. We have named G2 ‘Polarized conservative populists,’ and this group accounts for 47% of the sample.

Figure 3 Latent class analysis. Marginal means of the 7 survey items conditional on being in the 4 classes. 

Table 2 Proportion of respondents in each latent group. 4-class analysis 

Latent class	Size (As proportion of the total survey sample)
G1 – Non-partisan conservative populists	15%
G2 – Polarized conservative populists	47%
G3 – Polarized strong conservative populists	19%
G4 – Polarized liberal populists	19%

The third group (G3), which constitutes 19% of the sample, displays the clearest profile: they present strong favorable opinions toward the two items on populism, are highly supportive of the two cultural conservative values—quite a lot more strongly than the respondents in G1 and G2—and are affectively polarized against out-group parties, leaders, and voters. This group thus reflects very well the profile of citizens sharing populist principles, support for cultural backlash values, and high levels of affective polarization that we seek to identify in this paper. We have named G3 ‘Polarized strong conservative populists.’ Finally, the fourth group (G4) presents strong support for the two populist attitudes and is affectively polarized against out-group parties, leaders, and voters, but is not in favor of backlash values. This group represents 19% of the sample and we have named it ‘Polarized liberal populists.’

Once we have identified the four classes that best capture the different combinations of populist attitudes, cultural backlash values, and affective polarization in our sample of respondents, we then explore the level of democratic support and the socio-demographic characteristics that best predict being in each of the classes. In LCA, every individual is assigned a probability³This probability is a continuous variable that ranges from 0 to 1. of being part of one group, which can be used as the outcome or dependent variable in a regression analysis. Therefore, we estimate four separate OLS models—i.e. one for each subgroup or class—using the probability of membership in each class as the dependent variable, and as predictors we include the variable of democratic support and several socio-demographic variables, namely sex, age, level of education, ideology, and vote for the main radical-right party in Spain, Vox.

Figure 4 presents the coefficients of the four OLS models.⁴Table A.2 in the Appendix presents the full regression output. The evidence suggests that our main independent variable, respondents’ support for democracy, is only a clear predictor of being a polarized conservative populist (G2) and a polarized liberal populist (G4), but not a polarized strong conservative populist (G3) as anticipated. Indeed, the coefficients show that supporting democracy as the best form of government has a positive and statistically significant impact on the probability of being a polarized liberal populist (G4), whereas having a greater preference for a strong leader is a predictor of being a polarized conservative populist (G2). This means that individuals in the group of polarized liberal populists are identified by their stronger support for a democratic regime as an ideal form of government, whereas individuals in the group of polarized conservative populist are characterized by their tendency to support a strong leader. In turn, membership in the other two classes (namely the non-partisan conservative populists and the polarized strong conservative populists) is not associated with differential levels of democratic support.

Figure 4 Probability of membership in each latent group (OLS models) 

Focusing on the two classes that are better predicted by their distinct level of support for democracy, it is interesting to note that age represents a relevant predictor with, again, an opposing effect in the two subgroups. Thus, as the graph shows, the polarized conservative populists (G2) are more likely to be younger voters, whereas the polarized liberal populists (G4) are more likely to be the elderly. This indicates that holding populist views, conservative cultural values, and being affectively polarized is particularly harmful for democracy when this attitudinal pattern is maintained by younger individuals. This is consistent with the fact that the group of non-partisan conservative populists (G1) is also associated with younger age groups, and the group of polarized strong conservative populists (G3) is associated with older age groups, although, as already noted, neither of these two groups are predicted by their differential level of democratic support.

Other sociodemographic characteristics are also of interest. For instance, the polarized conservative populists (G2) are not distinguished from the rest of the sample by any particular ideological leaning, and they show a lower tendency to support Vox. This sociodemographic profile is indeed shared with the group of non-partisan conservative populists, who are also more likely to be from younger age groups. The polarized strong conservative populists are the only group with a higher tendency to vote for Vox and, in line with this, are situated further to the right of the ideological scale than any other social group. To finish, the polarized liberal populists are more left-wing and highly educated than the rest.