A revolt of the deplored? The role of perceived cultural distance in the educational gradient in anti-establishment politics.

Anti-establishment politics have become part of contemporary Western democracies. Less-educated citizens in particular have been found to display political distrust and populist attitudes, support populist parties, and abstain from voting. We advance a novel explanation for these patterns, drawing on extant theoretical insights to hypothesize that less- and more-educated citizens differ in the extent to which they perceive politicians to be culturally distant to them. Informed by our earlier in-depth qualitative research, we developed novel indicators of such perceptions and included them in a survey fielded among a high-quality panel representative of the Dutch population. We found: 1) positive associations between perceived cultural distance to politicians and political distrust, populist attitudes, the intention to vote for a populist party, and non-voting; and 2) that, overall, perceived cultural distance contributes substantially more to the educational gradient in anti-establishment political attitudes and behavior than the conventional rationalist and materialist approaches.

INTRODUCTION

Anti‐establishment politics is prominent in Western democracies. While many recent manifestations, such as the election of Donald Trump or the outcome of the Brexit referendum, attest to this, anti‐establishment politics has been studied for decades. These studies repeatedly highlight that less‐educated citizens in particular are likely to: have low trust in politics (Hakhverdian & Mayne, 2012); hold populist attitudes (Spruyt et al., 2016); support populist parties and politicians (Gidron & Hall, 2017; Mutz, 2018); and not vote (Hadjar & Beck, 2010; Jackson, 1995). Albeit their differences, each of these attitudes and behaviors signal discontent with establishment politics and politicians, most likely among groups that feel “marginalized, undermined and unrepresented by formal political forces, and wanted to make their point against what was generally seen as a political establishment” (Dodd et al., 2017, p. S7).

Two approaches for the educational gradient in anti‐establishment political attitudes and behavior dominate: a rationalist and a materialist one. According to the former, limited knowledge of the political domain, predominantly identified in less‐educated citizens, would produce a distorted evaluation of its functioning, inspiring distrust of (establishment) politics and politicians (see, e.g., Hakhverdian & Mayne, 2012). It may equally limit awareness and understanding of this domain, leading to less electoral participation (see, e.g., Hadjar & Beck, 2010). According to the materialist approach, those experiencing economic problems could blame establishment politicians for issues like job insecurity or unemployment, inspiring distrust in politics, populist attitudes, and support for populist parties (for an overview, see, e.g., Haugsgjerd & Kumlin, 2020; Spruyt et al., 2016). A weak economic position, in which less‐educated citizens are more likely to find themselves, could also constrain electoral participation due to financial concerns that lessen the attention paid to politics or make it more difficult to participate because of limited resources (see, e.g., Rosenstone, 1982).

Notwithstanding their merits, the rationalist and materialist approaches are less sensitive to the “socially structured differences in how people see, relate to, and understand politics” (Laurison, 2016, p. 688). Less‐educated citizens may not only be disadvantaged in politics because of a lack of relevant knowledge or resources, but also because of differences in how they relate to politics and politicians. In an effort to lay bare “the wider sociological factors which may be implicated in political alignments” (Dodd et al., 2017, p. S5), we therefore theorize a novel explanation focusing on such differences and test it empirically, while simultaneously accounting for the rationalist and materialist approaches. We build on studies which illustrate that less‐ and more‐educated citizens have different life‐worlds (see, e.g., Roose et al., 2012) and demonstrate how the higher social strata's contempt, as perceived by lower social‐strata individuals, inspires feelings of misrecognition (see, e.g., Lamont, 2018). We apply these insights to the political domain and theorize that an educational gradient in perceptions of cultural distance to politicians is relevant to understanding the educational gradient in anti‐establishment political attitudes and behavior. While our earlier in‐depth qualitative exploration already hints at this (Noordzij et al., 2021a), the relevance of perceived cultural distance to politicians among the public at large and its role in the educational gradient in anti‐establishment political attitudes and behavior, is unchartered territory.

Therefore, we analyzed data collected from a high‐quality panel formed using a probability sample drawn from the Dutch population register. Based on this analysis, we answer our research question: does perceived cultural distance contribute to the educational gradient in anti‐establishment political attitudes and behavior? Our data include novel items based on insights from our previous in‐depth qualitative research (Noordzij et al., 2021a), while our focus is on the Dutch case because it stands out for its marked educational gradient in anti‐establishment politics (see, e.g., Noordzij et al., 2019, 2021b). Moreover, this context enables us to strictly test the role of perceived cultural distance due to the high proportionality of the Dutch electoral system: a multi‐party system with no threshold to entering parliament other than gaining the number of votes needed to win a seat. This enables the political representation of various identities, which potentially makes the perceived distance between politicians and the electorate smaller than in countries with a disproportionate system. Hence, our finding that, overall, perceived cultural distance is of greater relevance than the conventional rationalist and materialist approaches for the educational gradient in anti‐establishment political attitudes and behavior in the least likely case of the Netherlands, suggests that it is also likely to be relevant in other contexts.

PERCEIVED CULTURAL DISTANCE AND THE EDUCATIONAL GRADIENT IN ANTI‐ESTABLISHMENT POLITICS

Along with differences in political expertise and economic standing, educational groups also have starkly different life‐worlds and lifestyles. These encompass contrasting cultural preferences with status overtones, ranging from cosmopolitanism (van der Waal & de Koster, 2015) to leisure activities and cultural consumption (Roose et al., 2012). This matters for perceptions of exclusion and subordination (Ridgeway, 2014): citizens in the lower social strata may perceive that those in the higher social strata deem their lifestyles and worldviews to be inferior, inspiring feelings of threat and sparking opposition (cf. Jarness & Flemmen, 2019; Lamont, 2018). Linked to these experiences are what Lamont (2018, p. 424) terms “recognition gaps,” in which “worth and cultural membership [become associated] with upper‐middle‐ or middle‐class identity, occupation, and lifestyle, attributes that are now out of reach for a growing segment of the population.” Consequently, the lower strata may perceive themselves and are likely “to be perceived by others, as ‘losers’” (Lamont, 2018, p. 424). Similarly, Krogstad and Stark (2020, p. 8) illustrate how a Norwegian newspaper taps into such feelings of misrecognition by describing “the cultural elite as an ‘exclusive circle’ [and] ‘haughty’ individuals who view ordinary people as inferior, or, in the alleged language of the cultural elite, as ‘stupid’, ‘vulgar’ or ‘uneducated morons’.”

In terms of possible political ramifications, we identified various indications that the life‐world of establishment politicians resonates more strongly with those belonging to the higher social strata, and more‐educated citizens in particular. Indeed, more‐educated citizens are currently overrepresented in politics, a phenomenon coined as “diploma democracy” by Bovens and Wille (2017). Consequently, what Ridgeway (2014, p. 11) highlighted for gateway institutions in general can also be seen in politics, namely that the “workplace cultures and practices [of dominant actors] are infused with the implicit but distinctive assumptions, values, and taken‐for‐granted knowledge of the middle class.” This institutionalization of the life‐world of more‐educated citizens comes at the expense of the interests and policy preferences of many less‐educated citizens (cf. Bovens & Wille, 2017). In addition to the impact it has on substantive representation, the diploma democracy likely also has extensive consequences for how citizens relate to the political domain (cf. Laurison, 2015, 2016; Visser et al., 2021). Compared to less‐educated citizens, more‐educated citizens are more likely to feel “like a fish in the water” (Bourdieu & Wacquant, 1992, p. 127) in the political domain because their life‐world is similar to those of many a politician.

Moreover, recent studies have hinted at how many (populist) politicians capitalize on the perception of distance to the cultural tastes and values, and alleged arrogance, of the cultural elite more broadly (see, e.g., Krogstad & Stark, 2020; Westheuser, 2020), and establishment politicians in particular (see e.g., Lamont et al., 2017; Norris & Inglehart, 2019). These politicians would signal their alleged “‘imagined sameness’ with the people” (Krogstad & Stark, 2020, p. 10) and “exploit and deepen mistrust of elites” (Norris & Inglehart, 2019, p. 91). Anecdotal examples of how establishment politicians are often far removed from, and even frown on, the life‐world of the lower social strata equally abound. While former US presidential candidate Hillary Clinton referred to Donald Trump supporters as “deplorables” (TIME, 2016), Martin Schulz, the former President of the European Parliament, called for a “revolt of the decent” to combat populist voters (De Correspondent, 2016). Such statements spark resistance from (populist) politicians who claim to truly represent the “common people,” reflecting the feelings of misrecognition to which such statements give rise. These politicians do not call for a “revolt of the decent,” but a “revolt of the deplored.” Trump, for instance, responded to Clinton's remark by stating that it revealed that she was “a person who looks down on the proud citizens of our country” (The Washington Post, 2016). Similarly, according to a prominent Dutch populist politician, many establishment politicians “look down on the real world, far away [from them], which they do not understand nor wish to understand” (PVV, 2017).

These examples suggest that differences in life‐worlds and the perceived contempt of many politicians for the lives of the lower social strata matter for understanding the latter's discontent with establishment politics. In‐depth interviews with politically discontented (white) working‐class citizens, or rural residents, in a US context (cf. Cramer, 2016; Williams, 2017; Wuthnow, 2018; see, also, Mckenzie, 2017), also hint at this. Williams (2017, p. 121), for instance, argues that “the class culture gap is a huge driver in American politics today” and highlights various lifestyle attributes that signal the distance between establishment politicians and the white working class. The same studies suggest that these perceived distances matter for feeling “left behind” and misrecognized: Lamont (2018, p. 425) has commented that this literature points to “a growing gap between what these workers believe to be their legitimate worth to society [and] the lower status they believe the broader society attributes them—a recognition gap, which generates considerable anger and resentment.” The literature above, however, does not explicate how perceptions of cultural distance impact political attitudes and behavior. Moreover, they primarily focus on class or regional differences, while we specifically focus on educational differences in perceptions of cultural distance to politicians, as level of education is the most salient aspect of stratification for determining political attitudes and behavior in societies like the Netherlands (van der Waal & de Koster, 2015). Tellingly, having a low level of education in such societies is even more strongly frowned upon than, for instance, being poor (Kuppens et al., 2018), which suggests that the feelings of misrecognition found among the lower classes and rural populations in the abovementioned studies also strongly apply to less‐educated citizens.

Our earlier in‐depth qualitative research explored the relevance of perceived cultural differences to politicians for understanding less‐educated citizens' political discontents in the Netherlands (Noordzij et al., 2021a). In particular, inspired by Cramer's “method of listening” (2016, p. 26, 39), we conducted in‐depth (group) interviews with dozens of politically discontented, less‐educated citizens, making sure to stress our genuine interest in their perceptions. Our open‐ended interviews uncovered a deeply felt perception of politicians as culturally distant “others”: many were viewed as remote, insensitive or blind to the participants' lived experiences and, therefore, their concerns. One of our interviewees summed this perceived distance up by saying that many politicians “travel in different circles, they're confronted with different things, and so they have different interests. […] And that's why they don't know what it's like for us.” Feelings of misrecognition also came into play: some interviewees accused politicians of regarding themselves and their life‐world as superior to those of the lower social strata, and of looking down on them. Consequently, many felt excluded and misrecognized politically. One of our interviewees commented that she felt “set aside as being deplorable […] because you don't have an education.”

We have translated the insights gained from our earlier research into novel indicators for measuring perceived cultural distance to politicians and testing the hypotheses outlined below. We expected there to be positive associations between perceived cultural distance to politicians and political distrust (hypothesis 1a), populist attitudes (hypothesis 1b), support for populist parties (hypothesis 1c), and non‐voting (hypothesis 1d). First, we took the view that the perception of politicians as culturally distant “others” who look down on those “below” them is likely to motivate distrust in the political domain. Second, populist attitudes center on opposing the political establishment, in which “the elite is seen as ‘evil,’ while the people are seen as ‘good’” (Akkerman et al., 2014, p. 1327). Conceptually, a key difference with perceived cultural distance to politicians is that populist attitudes focus on whether political power should reside with “the people” instead of “the elite.” We assume that those who perceive politicians as far removed from their lived experiences and looking down on them are more likely to agree with that. Third, populist parties are most vocal in opposing the establishment and expressing the primacy of the “general will” of a virtuous and homogenous “people” over that of a corrupt “elite,” claiming to give voice to those who have been overlooked by establishment politics (Akkerman et al., 2014). Last, whether or not citizens decide to abstain from participating in electoral politics out of a deliberate act of resistance, or from disinterest and disengagement (cf. Kemmers, 2017), the perception of politicians as culturally distant “others” may contribute to both.

Our theorizing above suggests that greater support for anti‐establishment politics among less‐educated citizens can be partially explained by their greater sense of cultural distance to politicians. Hence, we expect the educational gradient in the perceived cultural distance to politicians to contribute to the educational gradient in political distrust (hypothesis 2a), populist attitudes (hypothesis 2b), support for populist parties (hypothesis 2c), and non‐voting (hypothesis 2d). Technically speaking, we expect perceived cultural distance to politicians to partly mediate the relationship between education and anti‐establishment political attitudes and behavior.

While our theorizing and the hypotheses derived from it rest on the assumption that perceived cultural distance to politicians causes citizens to hold anti‐establishment political attitudes and behavior, we acknowledge that the causal direction could run both ways. We further reflect on this in the concluding section.

DATA AND MEASURES

We tested our hypotheses using recent survey data on a wide range of social and political attitudes and behavior (van der Waal et al., 2020). These data were obtained from members of the high‐quality Longitudinal Internet Studies for the Social sciences (LISS) panel, which was established through a random sample from the official population register. The process was administered by CentERdata (Tilburg University, the Netherlands) (Scherpenzeel, 2009), and the survey was conducted from March 2–31, 2020 and had a response rate of 80.1%.

While most of the data used in our analyses were collected in March 2020, multiple independent variables were derived from data collected from the same panel shortly before then (de Koster et al., 2020; Elshout, 2019a, 2019b). Prior to linking these data, we removed any respondents who were straightlining (n = 12) in de Koster et al. (2020). We also excluded those who had completed our main survey (n = 38) or that by de Koster et al. (2020) (n = 32) in 10 minutes or less, which is the minimum time needed to provide valid answers. We then appended our main data to that obtained by de Koster et al. (2020) and Elshout (2019a, 2019b) for the respondents included in all three datasets. After this merger process, and given our focus on the educational gradient in anti‐establishment political attitudes and behavior, we removed the data of all of those who had not yet completed their education (n = 64). The resulting dataset comprises 1,574 respondents. Table A1 in the online appendices contains the descriptive statistics of all the variables, as well as information on the datasets from which they were derived. Tables A2 and A3 include additional details on the items, coding, and factor analyses for all the multi‐item measures.

Operationalization of the dependent variables

First, we measured political distrust by constructing a reliable three‐item scale (Cronbach's α = 0.96) consisting of trust in parliament, political parties, and politicians (cf. Marien, 2011). We reverse coded these three items, before recoding the original range from (0) to (10) to (1) to (7) to ensure that all the ordinal dependent variables in our analyses were measured on the same range. Higher scores indicate that a respondent has a greater level of distrust in politics.

Populist attitudes was operationalized using a measure by Akkerman et al. (2014) and recommended by Castanho Silva et al. (2019) because of its high internal and external validity. We used the six items they have identified as tapping into populist attitudes to construct a reliable six‐item scale (Cronbach's α = 0.83). Higher scores on a range from (1) to (7) indicate stronger populist attitudes.

We measured support for left‐ and right‐wing populist parties separately. While voters for parties on both sides of the spectrum share populist attitudes, they differ ideologically and socio‐economically (Akkerman et al., 2017). At the time of the survey, the Dutch parliament contained representatives from one left‐wing populist party (Socialist Party, Socialistische Partij) and two right‐wing populist parties (Forum for Democracy, Forum voor Democratie; the Party for Freedom, Partij voor de Vrijheid) (Rooduijn et al., 2019). We used two variables to measure support for these parties to improve the interpretability of our results: intention to vote for a left‐wing populist party and intention to vote for a right‐wing populist party. For the former, we measured whether a respondent would vote for (1) a left‐wing populist party or (0) a non‐populist party if elections were held today (excluding 161 respondents who would vote for a right‐wing populist party); for the latter, we determined whether a respondent would vote for (1) a right‐wing populist party or (0) a non‐populist party were elections to be held today (excluding 75 respondents who would vote for a left‐wing populist party). We coded the following respondents as missing: those who would not cast a vote for anyone; those who are ineligible to vote; those who do not wish to reveal their voting intention; and those who do not yet know which party they would vote for (n = 298). Finally, we measured the likelihood of non‐voting by asking the respondents whether they would vote if elections for parliament were held today, reverse coding and recoding the original range from (0) to (10) in a range from (1) definitely to (7) definitely not.

Operationalization of the independent variables

We measured education using the minimum number of years needed to complete the reported level of education, with the range from (8) primary education to (18) a university degree.

In our analyses, and in addition to testing the merit of our novel explanation, we also assess the relevance of the rationalist and materialist approaches to anti‐establishment political attitudes and behavior. We included a measure of political knowledge to control for the rationalist approach to anti‐establishment politics. We used literature and data including measurements of political knowledge to construct a set of 15 questions, with an equal number for each identified political‐knowledge domain. We designed both multiple‐choice and correct‐incorrect questions. The survey stressed our intention to explore which facts are familiar to Dutch citizens (cf. Delli Caprini & Keeter, 1996). We explicitly asked the respondents to not look up or guess an answer if they did not know the correct response to a question, but to instead use the option I don't know. After recoding the answers as (1) correct and (0) incorrect (including I don't know), we counted the total number of correct responses for each respondent. Higher scores in the range (0) to (15) indicate more political knowledge.

We included three variables pertaining to the materialist explanation (cf. Haugsgjerd & Kumlin, 2020). Income was measured by taking the log of a respondent's net monthly household income because this distribution was skewed. Labor‐market position was measured by dummy variables indicating whether a respondent is (1) not in the labor market (either exempt from seeking work, housekeeping, retired or disabled); (2) unemployed; or (3) employed. Job insecurity was measured by asking how likely it is during the next 12 months that (1) those employed in the labor market become involuntarily unemployed; (2) freelancers are not able to find work; or (3) those who are unemployed cannot find work. The answer categories range from (1) a very small chance to (7) a very high chance. Respondents who are not part of the labor market were coded as (0).

In terms of control variables, we included variables used in studies on political attitudes and behavior (see, e.g., Akkerman et al., 2017; Gidron & Hall, 2017; Jackson, 1995; Noordzij et al., 2019, 2021b; Spruyt et al., 2016). In particular, we asked whether the respondent (1) is or (0) is not female and for their age in years. We controlled for being non‐native by asking whether a respondent (1) does or (0) does not have at least one parent who was born abroad, which is the standard demarcation used by Statistics Netherlands. In addition, we asked about religious denomination, which was coded as (1) none; (2) Protestant; (3) Catholic; or (4) other. We measured the frequency of attendance at religious services by asking whether a respondent (1) does not attend; (2) occasionally attends (once or a few times a year, or at least once a month); or (3) frequently attends religious services (once a week, more than once a week, or daily). We also controlled for cohabitation, asking whether the head of the household is living (0) alone or (1) with a partner. Information about the urbanity of where respondent lives was coded from (1) not urban to (5) very strongly urban. We also controlled for (1) whether or (0) not there are children living in the household. To account for the possible confounding influence of political attitudes, we controlled for ethnocentrism and economic egalitarianism. Both are especially prominent among less‐educated citizens (van der Waal & de Koster, 2015), while the former fuels support for right‐wing populists and the latter for left‐wing populists (Akkerman et al., 2017; van der Waal & de Koster, 2018). Inspired by prior studies that measure these attitudes (see, e.g., Akkerman et al., 2017; van der Waal & de Koster, 2018), we used four items to construct a reliable scale for ethnocentrism (Cronbach's α = 0.75) and four to construct a scale for economic egalitarianism (Cronbach's α = 0.75). They range from (1) to (7), and higher scores indicate stronger ethnocentrism and egalitarianism, respectively.

Finally, the Dutch government introduced lockdown measures regarding the COVID‐19 pandemic during the data collection, which affected citizens' institutional trust (Oude Groeniger et al., 2021). Hence, we included a control variable that coded whether the respondents began to complete the survey (0) before or (1) after the start of these COVID‐19 measures.

Perceived cultural distance to politicians

We translated the principal insights from our earlier qualitative research (Noordzij et al., 2021a) into five novel indicators. These were then used to construct a reliable five‐item scale (Cronbach's α = 0.88) measuring perceived cultural distance to politicians. We presented our respondents with a set of statements, with answer categories ranging from (1) completely disagree to (7) completely agree. These statements were: most politicians… (a) …are far removed from people like me; (b) … do not know what is going on for people like me; (c) … live lives completely different from those of people like me ; (d) …want to impose their way of life on people like me; and (e) …look down on people like me. The first three statements tap into the perception of politicians as culturally distant “others” with a different life‐world, while the latter two resonate with experienced feelings of misrecognition. Confirmatory factor analyses demonstrate that these novel survey items measure a latent construct that is empirically distinct from political distrust and populist attitudes (see Online Supplementary Material 1 in the online appendices).

Moreover, we assessed the association between our measure of perceived cultural distance to politicians and both education and relevant status indicators. First, we expected that less‐educated citizens would perceive politicians to be more culturally distant than would their more‐educated counterparts. This is confirmed by a negative correlation between education and perceived cultural distance to politicians (r = −.27; p < .001; n = 1,531). Second, we expected that the respondents who perceive politicians to be culturally distant would have little affinity with elite culture. The latter was measured with a reliable scale (Cronbach's α = 0.76) consisting of five standardized items on highbrow cultural participation, which has been the conventional measure for cultural capital from the 1980s onward (Jæger & Breen, 2016). As anticipated, we identify a negative correlation between perceived cultural distance to politicians and highbrow cultural participation (r = −.26; p < .001; n = 1,562). Last, we expected our scale to be positively associated with feelings of societal subordination. To measure this, we presented our respondents with an image of a vertical ladder with seven steps and the text: “In our society, there are groups which tend to be located at the top and groups which tend to be located at the bottom of the social ladder. Below is a ladder that represents the social ladder” (cf. Gidron & Hall, 2017). We constructed a reliable two‐item scale (the Spearman‐Brown coefficient, to be used for assessing the reliability of two‐item scales, equals 0.76) asking the respondents where they think (1) others in society locate them and (2) politicians locate them. A higher score on a range from (1) to (7) indicates that a respondent experiences greater societal subordination. The positive correlation between this measure and perceived cultural distance (r = .39; p < .001; n = 1,570) further indicates that our measure of the latter is valid.

RESULTS

For assessing our first set of hypotheses, we have visualized the key findings of regression analyses ran for: (1) political distrust (linear regression; details in Table A4, panel one, in the online appendices); (2) populist attitudes (linear regression; Table A4, panel two); (3) intention to vote for a left‐ and right‐wing populist party (logistic regression; Table A5); and (4) likelihood of non‐voting (linear regression; Table A4, panel three). Tables A4 and A5 in the online appendices are constructed as follows: after a descriptive model only including education (Model 0), we added all control variables (Model 1) and subsequently included perceived cultural distance to politicians together with indicators of the rationalist and materialist approaches (Model 2).

To describe the width of the education gap in anti‐establishment political attitudes and behavior, Model 0 of Tables A4 and A5 estimates the effect of education on our dependent variables without taking control variables into account. It is negative and significant in each case. Calculating the difference between the average predicted values for the least and highest educated for Model 0, we see that, compared to the least educated, the highest educated score substantially lower on political distrust (0.97), populist attitudes (1.32), and the likelihood of non‐voting (1.24)—all three variables range from (1) to (7) and have standard deviations of 1.25; 1.05; and 1.76, respectively. For the dichotomous variables measuring voting for a left‐ and right‐wing populist party compared to a non‐populist party, the average probability of the highest educated is, respectively, 0.11 and 0.27 lower than that of the least educated.

Next, we assessed the relevance of perceived cultural distance to politicians for anti‐establishment politics. In Figure 1, we plotted the average predicted levels of political distrust, populist attitudes, and the likelihood of non‐voting for the different levels of perceived cultural distance; we did the same in Figure 2 for intention to vote for a populist party. These figures show the associations between perceived cultural distance to politicians and anti‐establishment political attitudes and behavior while also including indicators of the rationalist and materialist approaches and all the control variables (Model 2). The figures demonstrate that citizens who view politicians as culturally distant are more likely to distrust politics (corroborating hypothesis 1a), display populist attitudes (corroborating hypothesis 1b), vote for left‐ or right‐wing populist parties (compared to voting for a non‐populist party, corroborating hypothesis 1c), and not vote (corroborating hypothesis 1d).

FIGURE 1

Predicted level of political distrust, populist attitudes, and likelihood of non‐voting for different levels of perceived cultural distance to politicians (based on Table A4, Model 2), including 95% confidence intervals. Perceived cultural distance is significantly associated with all three types of anti‐establishment political attitudes and behavior (p < .001)

FIGURE 2

Predicted level of intention to vote for a left‐wing populist party compared to a non‐populist party and for a right‐wing populist party compared to a non‐populist party for different levels of perceived cultural distance to politicians (based on Table A5, Model 2), including 95% confidence intervals. Perceived cultural distance is significantly associated with intention to vote for each of the two types of populist party (p < .001)

Does perceived cultural distance to politicians, in addition to the indicators of the rationalist and materialist approaches, also account for the educational gradient in anti‐establishment political attitudes and behavior? In order to answer this question, we need to test our mediation hypotheses, which stipulate that the educational gradient in perceived cultural distance to politicians contributes to the educational gradients in anti‐establishment political attitudes and behavior. To this end, we performed decomposition analyses using the KHB method (Breen et al., 2013), which also aids in more precisely determining the relevance of our explanation vis‐à‐vis the rationalist and materialist approaches. Table 1 contains these results. The total effects of education on anti‐establishment political attitudes and behavior reflect the effects of education in the model that includes all the control variables (Model 1 of Tables A4 and A5). The direct effects of education are those that remain after adding indicators of both the rationalist and the materialist approach and perceived cultural distance (Model 2 of Tables A4 and A5). Below these direct effects are the indirect effects, that is, the effects of education that are mediated by indicators of the rationalist (political knowledge) and materialist (income, labor‐market position, and job insecurity) approaches and perceived cultural distance to politicians, and below these are the percentages specifying the extent of the indirect effects’ contribution to the decline in the total effects of education.

TABLE 1

Decomposition of the total effect of education on anti‐establishment political attitudes and behavior into direct and indirect effects via indicators of the rationalist and materialist approaches and via perceived cultural distance to politicians

	Political distrust	Populist attitudes	Intention to vote for a populist party a		Likelihood of non‐voting
			Left‐wing populist	Right‐wing populist
Total effect of education	−0.05***	−0.08***	−0.23***	−0.19***	−0.15***
	(0.01)	(0.01)	(0.07)	(0.06)	(0.02)
Direct effect of education	−0.00	−0.03**	−0.07	−0.15*	−0.05*
	(0.01)	(0.01)	(0.07)	(0.06)	(0.02)
Indirect effect of education via…
Political knowledge	−0.01	−0.01***	−0.00	0.02	−0.09***
	(0.00)	(0.00)	(0.03)	(0.02)	(0.01)
Income	0.00	−0.00	−0.09***	0.00	0.01
	(0.00)	(0.00)	(0.03)	(0.02)	(0.01)
Being unemployed versus not in the labor market	0.00	0.00	–	0.00	−0.00
	(0.00)	(0.00)	–	(0.01)	(0.00)
Being employed versus not in the labor market	0.00	−0.00	−0.01	0.01	−0.01*
	(0.00)	(0.00)	(0.01)	(0.01)	(0.00)
Job insecurity	0.00	0.00	0.01	−0.00	0.00
	(0.00)	(0.00)	(0.01)	(0.00)	(0.00)
Perceived cultural distance to politicians	−0.04***	−0.04***	−0.06**	−0.07***	−0.01***
	(0.01)	(0.01)	(0.02)	(0.02)	(0.00)
Explained percentage of the total effect of education by…
Political knowledge	11.0%	16.0%	0.0%	−8.0%	58.9%
Income	−3.4%	0.3%	41.3%	−0.4%	−5.0%
Being unemployed versus not in the labor market	−1.1%	−0.1%	–	−0.7%	0.4%
Being employed versus not in the labor market	−2.0%	0.8%	4.9%	−4.6%	4.9%
Job insecurity	−0.3%	−1.1%	−2.7%	0.8%	−1.8%
Perceived cultural distance to politicians	86.0%	44.9%	24.9%	36.5%	9.5%
Total explained percentage of the total effect of education by all variables	90.1%	60.8%	68.5%	23.7%	66.8%
N	1,293	1,293	854	933	1,293

Unstandardized coefficients; standard errors in parentheses. Controlling for all the control variables.

The reported estimates of the total effect of education on the intention to vote for a populist party are slightly different from Model 1 in Table A5. This is because the KHB method is designed to validly decompose effects in logistic regressions by correcting for the rescaling that occurs between different models (Karlson et al., 2012). The reported direct effects of education correspond with Model 2 in Table A5.

p < .05

p < .01

p < .001.

The rationalist and materialist approaches somewhat contribute to the decline in the educational gradient in anti‐establishment politics: the former only in case of populist attitudes (16.0%) and the likelihood of non‐voting (58.9%), and the latter only in case of the intention to vote for left‐wing populist voting (via income: 41.3%) and the likelihood of non‐voting (via being employed: 4.9%). Our novel explanation contributed to the decline in the educational gradients for all political attitudes and behavior included in our study, namely political distrust (86.0%—corroborating hypothesis 2a); populist attitudes (44.9%—corroborating hypothesis 2b); and, compared to the intention to vote for a non‐populist party, the intent to vote for a left‐ (24.9%) or a right‐wing populist party (36.5%—corroborating hypothesis 2c). The smallest, yet still significant, contribution to the decline in the effect of education regards the likelihood of non‐voting (9.5%—corroborating hypothesis 2d). Clearly, the greater cultural distance to politicians perceived by less‐educated citizens substantially contributes to the educational gradient in anti‐establishment political attitudes and behavior. Except for the intention to vote for a left‐wing populist party and the likelihood of non‐voting, it does so considerably more than conventional rationalist and materialist approaches.

CONCLUSION AND DISCUSSION

This study aimed to provide a novel explanation of why less‐educated citizens are more likely than their more‐educated counterparts to display anti‐establishment politics, with the focus on political distrust, populist attitudes, support for populist parties, and the likelihood of non‐voting. Adding to the conventionally applied rationalist and materialist approaches, we developed and tested an explanation centering on education‐based differences in lifestyles and worldviews. We theorized that, because many less‐educated citizens do not recognize themselves in the life‐world of countless politicians and feel looked down on by them, they are likely to favor anti‐establishment politics.

Focusing on the Netherlands, which has an exceptionally large educational gradient in anti‐establishment politics (Noordzij et al., 2019, 2021b), we developed a multi‐item scale for perceived cultural distance to politicians based on insights from previous in‐depth qualitative work (Noordzij et al., 2021a). Using data collected from a high‐quality panel representative of the Dutch population, we found that citizens who perceive politicians to be culturally distant are more likely to display political distrust and populist attitudes, vote for a left‐ or right‐wing populist party instead of a non‐populist one, or not vote at all. Our analyses also revealed that the perception among less‐educated citizens that politicians are culturally distant and look down on them substantially accounts for the educational gradient in anti‐establishment politics and, except for the intention to vote for a left‐wing populist party and the likelihood of non‐voting, substantially more so than indicators of the rationalist and materialist approaches do.

As we are the first to measure perceived cultural distance and explore its role in anti‐establishment attitudes and behavior in a population‐based survey, it is not possible to assess whether it has grown in salience in recent decades, nor whether it is responsible for the increased educational gradient in some of those attitudes and behavior witnessed in various countries (Bovens & Wille, 2017). Yet, due to the combination of the increased salience of the life‐world of more‐educated citizens in politics—reflecting the notion of “diploma democracy” (Bovens & Wille, 2017)—and increasing segregation between less‐ and more‐educated citizens (Bovens et al., 2014), we consider both scenarios likely.

Future research could explore why perceived cultural distance to establishment politicians is more consequential for anti‐establishment attitudes than for anti‐establishment behavior. Additionally, the existence of various types of contemporary anti‐establishment political parties and politicians begs the question as to which of them are perceived to be less culturally distant, especially by the less educated, and why. Future studies could also uncover why a perceived cultural distance to politicians inspires some citizens to vote for a populist party and others to not vote. Kemmers' research (2017) on power orientations could provide a stepping stone: perceived cultural distance may inspire voting for a populist party among those convinced that political power is located in parliament and government, while it may inspire non‐voting among those who believe that the real power resides outside the conventional political domain, for example, with multinationals or global elites. Last, a key avenue for further research is to explore the causal relationship between perceived cultural distance and the political attitudes and behavior studied. Whereas our paper posits that the former inspires the latter, we acknowledge that it could also run the other way around. Exposure to populist rhetoric voiced by populist parties could, for instance, strengthen the perception of cultural distance among their electorate because it emphasizes the importance of “common people.” Another possibility is that politically distrusting citizens could be more sensitive to cultural differences between themselves and politicians. Testing the extent to which perceived cultural distance underlies anti‐establishment political attitudes and behavior and vice versa could, for instance, be done if relevant measures are included in a panel design. And (survey) experiments including stimuli that differ in the extent to which they inspire perceived cultural distance would aid rigorously assessing whether perceived cultural distance causally affects anti‐establishment political attitudes and behavior.

Our findings also have broader implications. First, whereas a weak economic position and/or having limited political knowledge scarcely contribute to the educational gradient in anti‐establishment politics, they could still be relevant for anti‐establishment politics by driving it in addition to level of education. Interestingly, and in addition to the mechanisms that are central to the rationalist and materialist approaches that focus on citizens' (lack of) knowledge and resources, they could do so via perceived cultural distance to politicians. From this point of view, our explanation of perceived cultural distance does not only provide an addition to the rationalist and materialist approaches in providing another mechanism through which education inspires anti‐establishment politics, but could also provide an interpretation of how knowledge and resources themselves drive anti‐establishment political attitudes and behavior besides education. While the materialist approach centers on constraints by and blame for financial concerns and the rationalist approach on distorted evaluations of politics, having limited knowledge of politics and few economic resources could also add to the perception that politicians are distant to one's life‐world. Importantly, this implies that destigmatization (cf. Lamont, 2018), in addition to addressing their financial concerns and knowledge limitations, could abate lower‐strata individuals' anti‐establishment political attitudes and behavior. Generally, we are convinced our explanation provides a valuable novel contribution to various debates on stratified patterns in political attitudes and behaviors, by being more sensitive to how lower‐strata individuals perceive and relate to the (political) world themselves (cf. Laurison, 2015, 2016; Visser et al., 2021).

Second, various recent studies also hint at the relevance of status for anti‐establishment politics, although they focus on it in different ways: education‐based status following meritocratic beliefs (Kuppens et al., 2018; Sandel, 2020); subjective social status and social esteem (see, e.g., Gidron & Hall, 2017); having an affinity with elite culture (see, e.g., Noordzij et al., 2019; Spruyt et al., 2016); or experiencing status threat (see, e.g., Mutz, 2018). Our conceptualization and operationalization of perceived cultural distance are more specific than these examples, as they are informed by our earlier in‐depth qualitative research on why less‐educated citizens are politically discontented (Noordzij et al., 2021a). Future studies could uncover how feeling culturally remote from the life‐world of politicians and misrecognized by them relates to more general status indicators, and their association with anti‐establishment political attitudes and behavior.

Third, there is a dominant notion that the rise of anti‐establishment politics, and support for right‐wing populist parties in particular, is a backlash in which discontent over globalization‐induced economic changes, immigration, and related policy decisions inspires the lower strata to support the nativist and authoritarian agendas of right‐wing populist parties (see, e.g., Kriesi et al., 2008; cf. Golder, 2016). Accordingly, the idea that these “voters are instrumentally motivated and guided by the same types of ideological and pragmatic considerations as voters of other parties” (Golder, 2016, p. 483) dominates. Our findings provide an additional argument as to how a rise in support for right‐wing populist parties could come about, namely by way of the increased salience of aforementioned diploma democracy (Bovens & Wille, 2017) and the feelings of misrecognition this inspires among less‐educated citizens. This hints at a revolt of the deplored, fuelled by colliding life‐worlds and feelings of misrecognition, in addition to a backlash by the losers of globalization due to globalization‐induced economic insecurities and immigration. In addition, these processes could interact: the nativist and authoritarian preferences of the losers of globalization are nowadays infused with notions of inferiority, which could increase the salience of the mechanism implied by the revolt of the deplored. In case of a backlash by the losers of globalization, the stances voiced by the lower strata are likely to be dismissed by the higher strata, fuelling feelings of misrecognition among “the deplored.” Future research could attempt to unearth the relevance of the abovementioned mechanisms of backlash and revolt for the rise of anti‐establishment politics in recent decades, and their interrelationship.

Fourth, we focused on level of education as a highly salient stratification indicator shaping political attitudes and behaviors in countries like the Netherlands. Future research could focus on the relevance of our explanation for understanding political discontents among working‐class citizens and rural residents (see, e.g., Cramer, 2016). Considering class, it stands to reason that it plays a more prominent role in shaping perceived cultural distance, and its political relevance, in more economically unequal societies. Moreover, as even in a very small and densely populated country like the Netherlands people in the larger cities have grown further apart culturally from the rest of the population (Huijsmans et al., 2021), living in rural regions in countries covering vaster territories is also likely to inspire perceived cultural distance to those living in culturally dominant cities, not only when it comes to the general population but also concerning politicians, who tend to live and work in the cultural center instead of the periphery.

Last, future studies could explore whether perceived cultural distance is also relevant for understanding educational gradients in distrust in judges and scientists (Achterberg et al., 2017; Noordzij et al., 2019), given that their (perceived) lifestyles and worldviews also differ substantially from those of many less‐educated citizens (cf. Frank, 2020).

In summary, we have theorized and demonstrated that perceived cultural distance to politicians substantially contributes to the well‐established educational gradient in anti‐establishment politics—and considerably more so than materialist and rationalist approaches. Future studies could uncover how far our findings travel beyond the strategically selected Dutch case, and whether they are also relevant for other social divides in anti‐establishment politics in both the political and other institutional domains.

CONFLICT OF INTEREST

No potential conflict of interest was reported by the authors.

ETHICS APPROVAL STATEMENT

Our study complies with all relevant ethical regulations.

Supplementary Material

Click here for additional data file.