• 1 Introduction

  Over the past years, political scientists have highlighted that the electoral support for populist parties, i.e. political actors defined by people-centrist and anti-elitist messages (Rooduijn & Akkerman, 2015; Wauters & Pittoors, 2019) is driven by both ideological and issue-based considerations ( Hobolt & Tilley, 2016; Van der Brug et al., 2000). In addition, it has been maintained that there are other unique drivers at play: protest motivations – defined in different ways, i.e. political disaffection (Norris, 2005), political distrust (Hooghe & Dassonneville, 2018a), political discontent (Rooduijn et al., 2016), dissatisfaction with democracy (Hernández, 2018) – and populist attitudes – defined as people-centrism and anti-elitism attitudes among voters (see, e.g. Geurkink et al., 2019; Van Hauwaert & Van Kessel, 2018). It seems indeed that both issue preferences and motivations based on negative feelings towards politics and the elite (hereafter populism-related motivations) contribute to explaining populist party support, and that these two aspects may reinforce each other’s effects (Passarelli & Tuorto, 2018). This is in line with other studies that showed how, to mobilise voters and become successful, populist parties need to combine their ‘thin ideology’ with existing ‘deep ideologies’ to develop policy-oriented considerations (Mudde & Rovira Kaltwasser, 2012).
  However, we know less about whether all voters rely equally on these two aspects (i.e. issue considerations and populism-related motivations) when casting a vote for a populist party. In this regard, previous research demonstrates that political sophistication works as a source of heterogeneity of the electorate and that it affects political and voting behaviour in various ways (Weisberg & Nawara, 2010). More precisely, the literature has shown that political sophistication is crucial to determine “the way in which attitudes guide political behaviour” (de Vries et al., 2010, p. 3). Political sophisticates are more aware of their ideological, political and policy preferences and are more prone to rely on those in their electoral calculus (de Vries et al., 2010; Lachat, 2008; Sturgis & Tilley, 2004). Yet, there is still only limited evidence of whether this also applies to populist party support. As such, we investigate how individuals’ level of political sophistication moderates the effects of issue considerations and populism-related motivations on support for populist parties. In so doing, we explore the possibility that voting motivations weight differently in the electoral calculus according to voters’ political sophistication.
  We study this using novel survey data gathered in the context of the 2019 Belgian general elections. Our analyses rely on the 2019 EOS Represent survey, which has collected data before (N = 3,420 in Flanders, N = 3,133 in Wallonia) and after (N = 1,978 in Flanders, N = 1,429 Wallonia) the 2019 elections in Wallonia and Flanders.¹xData on the Brussels region has been excluded from the analyses because of the smaller sample and the only few votes cast for populist parties. Belgium is an interesting country to study populism, for several reasons. As Hooghe and Dassonneville (2018a) argued, compulsory voting is an important factor as it reveals those populist votes driven by some sense of frustration that, in other contexts, might have been ‘lost’ in the counts of abstentions. Moreover, we also know that proportional systems like Belgium allow small parties to enter parliament (Spruyt et al., 2016), thereby favouring ‘sincere’ or first-preference votes (see also Schmitt et al., 2020). In the light of this, the Belgian case offers a suitable setting to evaluate what drives populist party support, since its electoral rules allow that a varied set of voters cast votes and that even upcoming populist parties (e.g. Partij van de Arbeid – Parti du Travail de Belgique, hereafter PTB-PVDA) are considered feasible vote choices by many voters. Indeed, Belgium provides, on the one hand, an example of a long-established right-wing populist party – Vlaams Belang, hereafter VB – that represents a “textbook case of populism” (Jagers & Walgrave, 2007, p. 334) but, on the other hand, the 2019 general elections also saw the breakthrough of a newer left-wing populist party, i.e. PTB-PVDA (Wauters & Pittoors, 2019). As such, these elections allow us to test whether our hypotheses hold up across the political spectrum.
  The results indicate that political sophistication enhances the impact of populism-related motivations (i.e. distrust, dissatisfaction and populist attitudes) on populist party support, although the effects are contingent on the party. Moreover, we show that, for issue considerations, the moderation effect only comes into play for VB voters: The impact of anti-immigrant considerations becomes higher at increasing levels of political sophistication.

• 2 The Electoral Support for Populist Parties

  The literature suggests that populist party support is driven by a combination of both structural context shifts and individual-level determinants. At the structural level, it is suggested that worsening economic conditions may foster support for populist parties (Betz, 1994; Jackman & Volpert, 1996), and Mair (2002) suggested that the erosion of parties as intermediaries between the public and politics contributed to the rise of populist parties. As parties shifted more towards responsibility (i.e. governing) rather than responsiveness to public needs, this opened a gap for populist protesters seeking to put ‘the people’ central and favoured a closer link between people and political decision makers (but see Kriesi, 2014). Yet, despite these broader contextual shifts, our focus lies more with the individual-level determinants of populist party support, i.e. issue preferences and other populism-related motivations (protest motivations, studied by means of different indicators, and populist attitudes). With regard to the former, Van der Brug et al. (2000) have stressed the importance of ideological and pragmatic considerations, demonstrating that, especially, attitudes towards immigration are a strong predictor of the vote for anti-immigrant parties. More recently, other studies have confirmed that issue positions do matter in the vote for populist parties (Birch & Dennison, 2017; Hooghe & Dassonneville, 2018b). Importantly, it seems that different issue preferences determine support for right- and left-wing populist parties – in line with the notion that populism is a ‘thin’ ideology that needs to be complemented with a (right- or left-wing) full ideology. Specifically, Hobolt and Tilley (2016) found that policy positions on immigration, European Union (EU) and government spending predict the vote for radical right parties, while economic redistribution counts for the vote for radical left parties. Similarly, based on comparative evidence from nine European countries, Van Hauwaert and Van Kessel (2018) demonstrate that attitudes towards immigration and authoritarianism drive right-wing populist party support, whereas socioeconomic attitudes affect left-wing populist party support. Yet, issue considerations are only one side of the coin. Indeed, a study of the Dutch case illustrates that policy considerations combine with the effects of protest attitudes and evaluation of party leaders in explaining support for populist parties (Schumacher & Rooduijn, 2013).
  With regard to populism-related motivations, studies have highlighted the relevance of indicators of political disaffection (Norris, 2005; Bakker et al., 2020), political distrust (Bélanger & Nadeau, 2005; Hooghe & Dassonneville, 2018a), political discontent (Van Kessel, 2015) and dissatisfaction with the functioning of democracy ( Hernández, 2018; Vidal, 2017). Although there is some discussion regarding the direction of influence – i.e. Rooduijn et al. (2016) suggest that in addition to discontent’s impact on party preference, populist party support may itself foster greater political discontent – the evidence for the role of discontent is well-established. Furthermore, populist attitudes have been also deemed important. Van Hauwaert and Van Kessel (2018) studied the combined effect of issue considerations and populist attitudes, and found that populist attitudes are prominent amongst populist party voters. They also showed that populist attitudes moderate the effect of issue preferences on the support for populist parties and that voters with strong populist attitudes may vote for a populist party whose positions do not match with their issue preferences. Using a sample of Dutch citizens, Geurkink et al. (2019) have also demonstrated that populist attitudes are not ‘old wine in new bottles’: Populist attitudes are distinct from political trust and external political efficacy, not just theoretically but also in the way they relate to populist voting.

• 3 Populist Parties in Belgium

  Next, we turn to which parties can be considered as ‘populist’ parties in Belgium. The Belgian party system is split between Flanders and Wallonia (Deschouwer, 2009). The literature about populism has focused especially on the radical right in the Flemish party system, as Vlaams Blok / Vlaams Belang has enjoyed continued electoral success, whereas, until recently, populist parties were less successful in Wallonia (De Cleen & Van Aelst, 2016). Research investigating populist parties has categorised populist parties on the basis of their communication style (Jagers & Walgrave, 2007; Rooduijn & Akkerman, 2015; Wauters & Pittoors, 2019). Two types of messages identify a populist party: on the one hand, a people-centrism message, which lies in the defence of the will of the people (i.e. the vox populi). On the other hand, anti-elitist messages target the corrupted elite as the enemy to fight. Following prior research, it is clear that the extreme-right VB is a textbook case of a populist party (Jagers & Walgrave, 2007; Pauwels, 2014). Furthermore, we also consider the extreme left-wing party PVDA/PTB as populist: Delwit (2012) labelled the party as ‘social populist’, and recent work by Goovaerts et al. (2020) and Wauters and Pittoors (2019) also agreed with this definition.
  Research investigating the drivers of populist voting in Belgium has confirmed that populist voting is explained by the interplay between policy motivations, political discontent and populist attitudes. Elchardus and Spruyt (2016), focusing on Flanders, argued that support for populism is not directly affected by economic insecurity or dissatisfaction, but that it is foremost a consequence of declinism, i.e. a negative view of the evolution of society. In a similar vein, Hooghe and Dassonneville (2018a) observed that those who vote for VB and N-VA in Flanders and PTB, FDF (Front Démocratique des Francophones) and FN (Front National)²xThese analyses do not include FDF and FN because they are smaller parties. in Wallonia had higher levels of political distrust. Further, Spruyt et al. (2016) indicated that populist attitudes are embedded in deep feelings of discontent, with both politics and life in general. Pauwels (2014) has also confirmed that disaffection with the functioning of democracy (expressed by a preference for direct democracy) is an important reason to vote for populist parties in Belgium, Germany and the Netherlands.
  The literature has also investigated whether issue evaluations determine voting for populist parties, with a specific focus on the case of VB. Already in 1995, Billiet and de Witte (1995) argued that the vote for VB is driven by negative attitudes towards immigration. In 2008, they confirmed these results, claiming that everyday racism, i.e. negative attitudes of the majority towards foreigners, works as a predictor of the support for VB. Similarly, the findings of Rink et al. (2009) showed that immigrant population size increases the probability to vote for VB, but its effect is reduced among people with higher levels of education. Although the case of PTB-PVDA has drawn less scholarly attention, we know that the issue of attitudes also matter for left-wing populist parties (see Hobolt & Tilley, 2016). These parties are especially effective in channelling discontent towards the economic situation, being more capable than left-wing mainstream parties in conveying messages in favour of economic equality and welfare spending and against the austerity ( Hobolt & Tilley, 2016; Rooduijn & Akkerman, 2015). As such, issue considerations and political discontent seem to affect populist party support, for both left- and right-wing populist parties. Moreover, research in the Belgian context has served to demonstrate the importance of issue ownership: Although there has been a traditional focus on the right-wing VB, we have good reasons to assume the same holds for PTB/PVDA as well.

• 4 (Populist) Voting and Political Sophistication

  Political sophistication has been studied in the literature as a source of heterogeneity of the electorate (Lachat, 2007; Weisberg & Nawara, 2010). Although populist voters tend to be portrayed in the public debate as poorly informed and politically apathetic, academic research has demonstrated that this portrayal does not hold up under empirical scrutiny (Bischof & Senninger, 2018; Van Hauwaert & van Kessel, 2018). On the one hand, several studies do seem to suggest a connection between lower sophistication and populist voting: Rovira Kaltwasser and Van Hauwaert (2019) found that populism is associated with the sense of disaffection with the way democracy works – which is more common amongst citizens with lower political involvement – both in Europe and Latin America, and that highly educated people are less likely to be populist voters. Studies also showed that supporters of radical right parties have lower levels of political trust and education than radical left voters (Rooduijn, 2018; Van Hauwaert & van Kessel, 2018; Van Kessel et al., 2020). Moreover, scholars have pointed out that populist voters tend to believe more in conspiracy theories (Castanho Silva et al., 2017).
  On the other hand, Van Hauwaert and Van Kessel found that populist parties’ supporters are attentive and politically interested, rather than apathetic citizens, which “directly contradicts interpretations of populist party support as the result of a simple protest mechanism for uninformed voters” (2018, p. 17). Meléndez and Rovira Kaltwasser (2019), using data from Chile, drew similar conclusions showing that populist voters’ political interest does not significantly differ from that of mainstream parties’ supporters. With regard to voters’ political knowledge, Bischof and Senninger (2018) showed that populist parties do use less complex campaign messages and, because of that, individuals are better informed about their positions. Others have illustrated that being informed positively affects the probability to turn out, no matter whether the support is for populist or mainstream parties (Van Kessel et al., 2020).
  In brief, findings suggest that populist voters might generally be less educated and more disaffected, but they are not homogeneously politically apathetic and uninterested. In this sense, political sophistication can be studied as an effective source of heterogeneity of the populist electorate. As de Vries et al. (2010) noticed, “When studying the moderating influence of political sophistication, scholars often distinguish between political interest and political knowledge” (3). In our case, to account for the different aspects of political attentiveness that might characterise the populist electorate, we build upon a multidimensional definition of political sophistication (Lachat, 2007), relying on political knowledge, political interest and political involvement (here understood as campaign attention). Indeed, sophistication should comprise not only factual political knowledge – sophisticated citizens should be informed about political affairs – but also a general tendency to inform oneself about politics, have higher levels of political involvement and be engaged in political activities (Lachat, 2007, p. 56). Yet, this only partially covers the multifaceted concept of political sophistication, which, according to many other conceptualisations, also regards the consistency of attitudes, i.e. the degree of association between opinions (Converse, 1964; Weissberg, 1976). In this respect, Luskin (1987) has clarified that sophistication is composed of size, range and constraint, in the sense that “a person is politically sophisticated to the extent to which his or her PBS (political belief system) is large, wide-ranging, and highly constrained” (860). In other words, political sophisticates are more informed (large), have a more diversified knowledge (wide-ranging) and better structure (highly constrained) of the information they hold. Yet, this article does not include an indicator of attitudinal consistency amongst its predictors, as this would interfere with the main purpose of the analyses, i.e. testing how different sets of attitudes drive voting behaviour. For this reason, it focuses on the indicators of sophistication that have been already used in the literature in relation to the study of the populist electorate (specifically, political knowledge and political interest).
  Regardless of the employed operationalisation, in the literature there is strong evidence that the level of attentiveness to politics affects not only the way people receive and understand messages (Zaller, 1992) and the complexity of their reasoning ( Baldassarri & Schadee, 2006; Gigerenzer & Todd, 1999) but also their voting behaviour. Sophisticated voters evaluate objects in the political world on the basis of more abstract and ideological considerations (Converse, 1964), evaluate candidates according to issue preferences (Lavine & Gschwend, 2007), are more prone to issue voting (Sturgis & Tilley, 2004), more sensitive to contextual factors (Ensley, 2007), more likely to know party positions (Gerber et al., 2015) and to vote for a party that matches their ideological preferences (Lachat, 2008). In addition, sophistication has also been deemed important in mobilising a more reasoned vote since it guides the way voters understand and make use of their political attitudes and preferences. Indeed, voters need to discern their attitudes in order to use them effectively in the electoral choice. In other words, they need to rely on clear preferences to direct their votes towards the party that better represents them. As argued by de Vries et al. (2010), the role of political sophistication is crucial in rendering attitudes more easily accessible: In this way, political sophisticates are more likely to take into account their political attitudes when forming voting preferences and casting a vote. This explains why more sophisticated people are more inclined to act upon their policy preferences while casting a vote. Furthermore, they should be able to make sense of other political preferences or inclinations such as in our case, populist motivations and count on them meaningfully in their electoral decisions. Taking this into account, we expect that both issue considerations and populism-related motivations matter more in the electoral choice of politically sophisticated voters compared to less sophisticated voters:

  H1: The relation between populism-related motivations and vote for populist parties is stronger for highly sophisticated voters, compared to less sophisticated voters.

  H2: The relation between issue considerations and vote for populist parties is stronger for highly sophisticated voters, compared to less sophisticated voters.

  We expect this moderation to work both for radical left and radical right populist party voting (VB and PTB/PVDA). However, as noted earlier, we do expect different issue considerations to matter for left- and right-wing populist parties, with immigration positions affecting the vote for VB, and economic positions affecting the vote for PTB/PVDA. In contrast, populist motivations should drive the vote for populist parties similarly across the spectrum. In terms of causality, we are aware that, with regard to policy preferences, the reverse relationship is also likely to happen, i.e. political sophisticates, counting on a better understanding of politics, can align their positions to those of the party they voted for (Slothuus & de Vreese, 2010). However, in accordance with the literature about the drivers of populist support, in this article we study how policy positions affect the vote for populist parties and not vice versa. To do that, we conduct analyses per region, studying the moderating role of sophistication for the vote for VB (radical right) and PVDA (radical left) in Flanders and PTB (radical left) in Wallonia. The following paragraph provides more information about the data.

• 5 Data

  To test our hypotheses, we use the longitudinal study of the 2019 EOS Represent survey,³xFor a description of the data, see the introduction to this Special Issue. which contains the reported vote choice of respondents (wave 2), several policy positions (wave 1), populism-related indicators (i.e. political distrust and dissatisfaction with democracy) and populist attitudes (wave 1), as well as indicators of political sophistication (i.e. political knowledge, political interest in wave 1 and involvement in the campaign in wave 2). Our dependent variable is the reported vote choice for the federal elections. Respondents were asked, “For which party did you vote for the federal chamber of representatives in the elections of 26 May 2019?” The answer categories consisted of the seven main parties in each region,⁴xCD&V, Groen, N-VA, Open VLD, PVDA, sp.a, Vlaams Belang in Flanders, cdH, DéFI, Ecolo, MR, PP, PS, PTB in Wallonia. blank/invalid vote, did not vote, not eligible to vote, or do not recall. As regards the selection of cases, the most cited example of populist party in Belgium is the Flemish radical right VB, which enjoyed a period of sustained electoral success already in the 1990s, followed by a waned support after its electoral peak in 2004 and emerged victorious once again in 2019. Although less often quoted (probably also because of its more modest electoral success), the radical left PTB-PVDA, which – contrary to VB – runs both in Flanders and in Wallonia, is also characterised by a strong populist rhetoric. Indeed, both VB and PTB-PVDA “fundamentally pitch the people versus the corrupt elite, present both as largely homogenous and want to see policy as the unfiltered expression of the will of the people” (Wauters & Pittoors, 2019, p. 6). Although many scholars have focused exclusively on the right-wing VB, Rooduijn and Akkerman (2015) have demonstrated that the Western Europe radical left is typically populist, which allows us to study populist parties across the spectrum.
  We opted to split the analyses for VB, PVDA and PTB. For the case of Flanders, we recode as 1 the votes for PVDA and 0 the voters for all the other parties (excluding VB); another variable indicates as 1 the votes for VB and 0 the voters for all the other parties (excluding PVDA). This allows us to contrast the vote for a populist party with a vote for other types of parties. Similarly, in Wallonia, we compute one dichotomous variable where 1 indicates the votes for PTB and 0 for all the other parties. In both regions, blank or invalid votes, together with those who did not vote, were not eligible to vote or did not remember the vote, are recoded as missing values. In contrast to a multinomial regression approach, the dichotomous nature of our dependent variable allows us to assert to what extent the independent variables of interest affect populist party voting in contrast to other parties in general, rather than having to contrast the populist party vote against one specific party. The key drawback of using this approach is that we cannot make claims about populist parties versus other individual parties (i.e. populist parties versus nationalist, Christian-democrats and so on). Yet, given our theoretical interest in studying the effect of the different populist parties vis-à-vis other parties in general, we believe that using binominal models is the most fitting approach.
  To measure respondents’ political sophistication, we rely on a multidimensional definition (Lachat, 2007) and we use the index proposed by Lupton et al. (2015), which consists of respondents’ combined scores on the questions concerning political interest, political knowledge and political involvement. In particular, we use the following questions: political interest, which asked respondents, “To what extent are you generally interested in politics?” Respondents answered on an 11-point scale ranging from 0 (Not interested at all) to 10 (Extremely interested). Then, six questions about factual political knowledge⁵xIncorrect and refused answers are recoded together as 0, whilst correct answer is recoded as 1. The questions ranged covered the three categories of knowledge outlined by Delli Carpini and Keeter (1996): the rules of the game (i.e. composition of federal parliament), players of the game (i.e. chairman of federal chamber of representatives) and the substance of politics (i.e. EU involvement in policy domains). (Cronbach’s alpha = 0.82) and campaign intensity as indicators of political involvement (intensity with which respondent followed the last electoral campaign: 1 = Very intensively; 2 = Intensively; 3 = Not very intensively; 4 = Not at all, reversed). These variables are combined in an additive index (Cronbach’s alpha = 0.65), then rescaled to range from 0 to 10, where 0 means ‘Lowest level of political sophistication’ and 10 means ‘Highest level of political sophistication’.
  We expect political sophistication to moderate the impact of two determinants on populist party voting: issue preferences and populism-related motivations. With respect to the first, we take into account two issue positions, i.e. economic attitudes, which are signalled by the radical left populist parties, and anti-immigrant attitudes, which are, instead, relevant to the vote for radical right populist parties. Both variables are measured on 11-point scales: Economic preferences regard respondents’ opinions on the free market (“Some people think that the government must intervene as little as possible in the market. Other people think that the government must intervene as much as possible in the market”, 0-10 scale: 0 = The government must intervene as little as possible; 10 = The government must intervene as much as possible). The other variable regards, instead, cultural position on the issue of ‘immigration’ (“Some people think that non-western immigrants must be able to live in Europe while preserving their own culture. Others think that those immigrants should adapt to the European culture”, 0-10 scale: 0 = Completely preserve their own culture; 10 = Completely adapt to the European culture).
  Moving on to populism-related motivations, we use three indicators. First, we create an additive index of political distrust in four institutions: political parties, federal parliament, politicians and the EU (Cronbach’s alpha = 0.94), measured on a scale from 0 (Absolutely no confidence) to 10 (Complete confidence). This index is reversed and recoded from 0 to 10, where 0 means ‘Low level of political distrust’ and 10 ‘High level of political distrust’. Second, regarding political dissatisfaction, respondents were asked, “Generally speaking, are you rather satisfied or dissatisfied with the way Belgian democracy works?” and they answered on a scale from 1 (Very satisfied) to 5 (Very dissatisfied). Third, we track respondents’ populist attitudes based on a battery of seven populist items, measured on a scale from 1 (Totally disagree) to 5 (Totally agree). The results of a principal component factor analyses with varimax rotation show that four of these items load on an anti-elite dimension (see also Akkerman et al., 2014). Since populism in Belgium has been demonstrated to be characterised by an anti-elite rhetoric (Jagers & Walgrave, 2007), we select these four variables⁶xThe other three items were When making decisions, politicians care about people like me; Politics is the result of compromise and common sense; In general, politics reflect rather well the people’s preferences. Those have not been included in the index of populism since they do not strictly pertain to the anti-elitism dimension (they can be assimilated to the political efficacy and pluralist dimension, see, e.g. Akkerman et al., 2013). to build an indicator of populist attitudes (Cronbach’s alpha = 0.71). The additive index is thus built on the following four items:

  “Politicians must follow the people’s opinion”–“Political opposition is more present between citizens and the elite than between citizens themselves”–“I prefer being represented by an ordinary citizen rather than by a professional politician”–“Rich citizens have a bigger influence on politics than do poor citizens”.

  The index is recoded from 0 ‘Low populist attitudes’ to 10 ‘High populist attitudes’.
  Lastly, we include gender, age, education and respondents’ overall distance to the party as controls. The latter tracks the absolute distance of voters from the average ideological position of the electorate of populist parties (VB, PVDA or PTB, according to the party considered in the model). This enables us to account for the overall distance to the party, to assert whether specific issue considerations exert an effect on the populist party vote beyond overall ideological distance.

• 6 Results

  Table 1 reports the average values of political sophistication, populist attitudes, political distrust, political dissatisfaction and issue positions for the Flemish party electorates, carried out by means of analysis of variance (ANOVA). The F test shows that there is a statistically significant difference between the means of the different parties, although the significance of pairwise comparisons between parties varies. For our purposes, we focus here on the differences between the populist parties (PVDA and VB) and the other parties, as assessed through a Tukey–Cramer post hoc comparison between the other parties and PVDA/VB, respectively. The superscript indexes in the table track which of the other party electorates’ means differ significantly (p < 0.05) from the means of PVDA (index a) and VB (index b): If an entry has a superscript index, its mean varies significantly from that party. Regarding political sophistication, the VB electorate scores significantly lower than almost all electorates, whereas for PVDA the level of sophistication is decidedly more ‘middle of the pack’, with no significant differences from other parties, except VB. PVDA’s supporters also show significantly higher means of populist attitudes, while – surprisingly – this is not the case for the VB electorate. As expected, the PVDA and VB electorates are significantly more distrusting than all other parties, and more discontent than other electorates. The only exception is the difference between the electorates of PVDA and sp.a (Socialistische Partij Anders), which is not significant In terms of ideological preferences on the economy, the variations between party electorates are small, and we do not find significant differences between the populist parties and the other parties. In contrast, for immigration, we do find differences between parties, but these seem to be more structured alongside the left-right axis than between populist and non-populist parties.
  Table 2 presents the data for Wallonia. Voters for PTB are characterised, on average, by significantly higher levels of populist attitudes, political distrust and political dissatisfaction than the other electorates – excepting the electorate from Parti Populaire (PP), which is quite similar in these respects to the PTB electorate. Therefore, we also ran the multivariate analyses excluding the PP electorate from the baseline category in Wallonia, but our substantive findings for the two hypotheses remain unchanged when doing so. With regard to positions on free market, we should expect PTB supporters to take positions in favour of the state intervention, thus to show low scores. Yet, on the free market scale, the scores of all electorates are quite close to each other, and no significant differences emerge when comparing the PTB electorate overall to the other electorates. Indeed, the results are quite surprising. PTB supporters do not score high on positions on free market, and it seems, indeed, that they take more neutral positions about the intervention of the state in the market. This suggests that across all voters who cast votes for populist parties, issue positions do not systematically differ from those of other parties. On immigration, we did not have clear expectations for PTB supporters – and indeed they score pretty average on this scale when compared to the other electorates. Only the PP electorate scores very right-wing, and significantly differs from the PTB electorate.
  Next, we evaluate our hypotheses using binomial logistic regressions. Analyses are weighted on the basis of socio-demographic variables (gender, age, education) and the vote. Before formally testing H1 and H2, we first regress voting for PTB, PVDA and VB on the independent variables of interest, including only their direct effects (Table 3).

  Table 1 Average values of voters’ political sophistication, populist attitudes, political distrust and dissatisfaction and issue preferences by parties in Flanders

  Independent variables	CD&V (N = 202)	Groen (N = 173)	N-VA (N = 518)	Open VLD (N = 172)	sp.a (N = 208)	PVDA (N = 134)	Vlaams Belang (N = 395)
  Political sophistication	6.18b	5.95b	6.35b	5.49	5.67	5.97b	5.22a
  Populist attitudes	6.64a	6.42a	6.45a	6.23a, b	6.97a	7.79b	6.80a
  Political distrust	5.03a, b	5.15a, b	5.47a, b	5.12a, b	6.03a, b	6.98	7.06
  Political dissatisfaction	2.88a, b	2.97a, b	3.17a, b	2.98a, b	3.29b	3.64	3.94
  Free market	6.24	6.25	5.64	5.94	6.21	5.97	5.74
  Immigration	7.08a	5.98b	7.88a	7.11a	6.93b	5.95b	7.72a

  Note: table entries with superscript indexes differ significantly (p < .05) from the scores of PVDA (a) and VB (b).

  Table 2 Average values of voters’ political sophistication, populist attitudes, political distrust, dissatisfaction and issue preferences by parties in Wallonia

  Independent variables	cdH (N = 90)	DeFi (N = 86)	Ecolo (N = 220)	MR (N = 247)	PP (N = 59)	PS (N = 282)	PTB (N = 190)
  Political sophistication	5.28	5.06	5.22	5.69a	4.66	5.12	4.64
  Populist attitudes	6.34 a	6.68 a	6.73 a	6.07 a	7.17	6.61 a	7.50
  Political distrust	5.79 a	6.59 a	6.27 a	5.51 a	8.05	6.28 a	8.04
  Political dissatisfaction	3.18 a	3.54 a	3.33 a	2.87 a	4.15	3.24 a	4.11
  Free market	6.25	6.02	6.15	6.35	5.37	6.41	5.99
  Immigration	6.63	7.47	6.72	7.60 a	8.32 a	6.84	6.69

  Note: table entries with superscript indexes differ significantly (p < .05) from the scores of PTB (^a).

  Table 3 Baseline models (binomial logistic regression)

  	Wallonia	Flanders
  	PTB	PVDA	VB
  Political sophistication	0.0851*	0.108*	-0.0533
  	(0.0482)	(0.0633)	(0.0370)
  Populist attitudes	0.170**	0.309***	-0.112**
  	(0.0675)	(0.0825)	(0.0519)
  Political distrust	0.266***	0.258***	0.198***
  	(0.0720)	(0.0725)	(0.0460)
  Dissatisfaction with democracy	0.236	0.191	0.670***
  	(0.158)	(0.138)	(0.0894)
  Positions on free market	-0.0173	-0.0457	0.0352
  	(0.0508)	(0.0531)	(0.0401)
  Positions on immigration	-0.0739*	-0.241***	0.0756*
  	(0.0439)	(0.0479)	(0.0390)
  Gender	-0.0799	-0.100	-0.350**
  	(0.225)	(0.271)	(0.171)
  Age	-0.0239***	-0.0242**	-0.0265***
  	(0.00750)	(0.00954)	(0.00538)
  Education	-0.463***	-0.473***	-0.469***
  	(0.120)	(0.156)	(0.0875)
  Ideological distance to PTB	-0.0673
  	(0.0777)
  Ideological distance to PVDA		-0.404***
  		(0.0920)
  Ideological distance to VB			-0.282***
  			(0.0563)
  Constant	-2.759***	-2.126*	-0.955
  	(0.890)	(1.207)	(0.712)
  Observations	1,205	1,417	1,676
  R-squared	0.13	0.20	0.17

  *** p<0.01, ** p<0.05, * p<0.1

  In line with prior research, we find that sophistication does not exert a negative effect on voting for populist parties: Indeed, this is the case for none of the parties under study here, suggesting that the image of populist party voters is less sophisticated and is inaccurate. Instead, for the left-wing populist parties PTB and PVDA, we even find a slightly positive effect of sophistication. When examining populist attitudes, the impact is positive, as expected, for PTB and PVDA, but negative for VB. We do not have a quick and easy explanation for this counterintuitive effect, but it lines up with the bivariate evidence which has already shown that VB voters do not hold significantly more populist attitudes than many other electorates. In contrast, distrust and dissatisfaction exert consistently positive effects on voting for all populist parties, although the impact of dissatisfaction is only significant for VB. For issue considerations, we find an interesting pattern: In line with the bivariate results, free market considerations do not exert an effect on voting for any populist party. In contrast, immigration attitudes do affect populist party voting – and not just VB but also PTB/PVDA. In line with recent work by Walgrave et al. (forthcoming), this seems to suggest that attitudes related to the cultural left-right were more influential than attitudes on the economic left-right.
  To evaluate H1 and H2, we specify interaction models. Given that interaction effects are hard to evaluate based on the estimated coefficients (see, e.g. Brambor et al., 2006), we provide the full models in the appendix, and focus our attention here on the graphed conditional marginal effects. H1 proposed that the impact of populism-related attitudes on populist party voting would increase at higher levels of political sophistication. Three measures capture the populist attitudes of voters: political distrust, political dissatisfaction and populist attitudes. As such, we estimated models with two-way interaction terms for each of these attitudes and political sophistication. Although the significance of the interaction term varies, we always plot the marginal effect since even non-significant interaction terms may entail a meaningful moderation (see Brambor et al., 2006). Figure 1 presents the marginal effect of populist attitudes, distrust and dissatisfaction for PTB, PVDA and VB, respectively.
  Overall, the findings are in line with H1: All curves slope upward, suggesting that the marginal effects increase as the level of sophistication rises. Yet, there do seem to be party-specific patterns at play as well. Turning first to the findings for PTB, the results suggest that at higher levels of sophistication, distrust and dissatisfaction exert a significant effect on the probability to vote PTB compared to the other parties. For distrust, the effect becomes significant already at lower levels of sophistication compared to dissatisfaction, however. For populist attitudes, the curve is relatively flat, suggesting no moderation by sophistication. For PVDA, the moderation effect is only minor: On the one hand, populist attitudes seem to increase the chances of voting PVDA across the whole range of sophistication, and, on the other hand, dissatisfaction seems to never play a role. For distrust, we find a very minor pattern of moderation, but overall it seems that sophistication did not exert a strong moderating influence for voting for the PVDA. Finally, the curves for VB do suggest strong moderation: First, populist attitudes exert a negative effect on VB voting at lower levels of sophistication, which runs counter to our expectation. At higher levels of sophistication, this effect becomes insignificant. Note that the baseline category in this regression excludes PVDA, suggesting that even compared to ‘mainstream’ parties VB did not attract more populist voters. For distrust and dissatisfaction, we find that their positive effect on voting for VB only comes into play at moderate and high levels of sophistication, as expected. In sum, the evidence suggests an overall moderation effect of populism-related motivations by political sophistication for VB and PTB, and only a very weak moderation for PVDA. As such, we can neither clearly confirm nor reject H1. Indeed, it seems that this moderation is party-specific, rather than a pattern that emerges for all populist parties under study.

  Conditional marginal effects of populism-related motivations on the probability to vote for populist parties according to levels of political sophistication (*)
  (*) The area between the dashed lines represents the confidence interval (at 95%). The zero line is reported in grey.

  We also expected that the impact of issue considerations would be moderated by respondents’ level of sophistication (Figure 2). Here, it is apparent that we only find such a moderation for VB. Respondents’ position on the free market does not exert an effect on PTB/PVDA voting, and this holds at all levels of political sophistication. This is not too surprising, given that the bivariate results already demonstrated that, contrary to our expectations, the attitudes of the PVDA/PTB electorates on the economy did not diverge all that much from the other electorates. In contrast, for VB, we do find meaningful moderation in the expected direction: Immigration attitudes only exert a positive effect once respondents’ level of sophistication exceeds the middle point on the sophistication scale. Again then, both the impact of issue attitudes and the moderation by sophistication seem to be party-specific, rather than a generalisable pattern across populist parties. We reflect on these findings in the conclusion.

  Conditional marginal effects of issue considerations on the probability to vote for populist parties according to levels of political sophistication (*)
  (*) The area between the dashed lines represents the confidence interval (at 95%). The zero line is reported in grey.

• 7 Discussion and Conclusion

  This article set out to investigate the moderating role of political sophistication in voters’ choice for left- and right-wing populist parties in the 2019 Belgian elections. Previous studies have stressed that populist voting might be fuelled by a highly heterogeneous set of factors, including issue considerations, populist attitudes and political discontent (see, e.g.: Hernández & Kriesi, 2016; Passarelli & Tuorto, 2018). Yet, we know that different levels of individual engagement and attentiveness to politics can entail a more or less elaborate reasoning process. Against this background, we wanted to test whether political sophistication shapes the motivations to support a populist party. By differentiating the drivers of populist party support amongst higher and lower levels of sophistication, such an analysis provides insight into the different patterns of determinants that may underlie populist party support. In particular, we argued that higher levels of sophistication should enhance an issue-based vote and the impact of populism-related motivations. To test these hypotheses, we relied on novel survey data that was gathered in the context of the 2019 general elections in Flanders and Wallonia. The party systems in both regions contain successful populist parties – VB and PVDA in Flanders, PTB in Wallonia – and as such are well suited to evaluate our hypotheses.
  The results indicate that the image of populist voters as unsophisticated citizens does not hold up to empirical scrutiny: Sophistication had either no (VB) or a positive effect (PVDA/PTB) on populist party voting. These findings show that lower political skills do not increase the probability of a populist voting: Voters of populist parties were, at least in the 2019 Belgian general elections, not substantially less sophisticated than other electorates. Furthermore, we found that political distrust and dissatisfaction with democracy increase the likelihood to vote for a populist party in both Belgian regions, but that the effects do vary between parties. In Flanders, populist attitudes drive voting for PVDA, but not for VB, while distrust positively affected voting for both PVDA and VB, and dissatisfaction affected VB voting, but not PVDA voting. Similarly, for PTB we found that distrust exerted the expected positive effect, as did populist attitudes, and dissatisfaction was not significant.
  The core of the analysis examined the moderation of the aforementioned attitudes by political sophistication. We expected populist attitudes and issue considerations to exert a greater effect amongst the more sophisticated voters. Although we found some evidence that suggests moderation by political sophistication, the effects are contingent on the party and we did not find evidence of a generalised pattern of moderation. Regarding populist considerations, the overall pattern suggests moderation, but the strength of the moderation varies across attitudes and parties. For distrust, we find the most robust evidence of moderation, with the effect only exerting a significant impact for moderate to highly sophisticated voters. For populist attitudes, the effect seems to vary substantially across parties, and the moderation by sophistication is only substantively impactful for VB, where the negative effect of populist attitudes disappears at higher levels of sophistication. For dissatisfaction, then, we find evidence of moderation amongst PTB and VB voters, but not PVDA. For both PTB and VB, we show that, as expected, these only affect the votes at moderate to high levels of sophistication. For issue considerations, the moderation effect only came into play for VB voters, but we suspect that this is partly due to the fact that free market and economic attitudes simply had a lesser impact on the vote in general – at least in the 2019 elections (see also Walgrave et al., forthcoming). Consequently, we did not find evidence that economic attitudes affected the propensity to vote PTB or PVDA at any level of sophistication.
  In sum, the evidence suggests that while political sophistication may moderate the impact of issue considerations and populist attitudes, the patterns are less straightforward than one would expect: Different attitudes shape the vote for different populist parties, and, consequently, the moderating effect of sophistication also varies across parties and attitudes. Yet, our evidence does suggest that future work would do well to consider the moderating role of sophistication: For various attitudes – i.e. immigration attitudes for VB voting, distrust and dissatisfaction for PTB and VB – we found evidence that these mattered to explain populist party voting. Importantly, in some cases, these effects only became apparent once we accounted for the moderation by sophistication: We did not find an overall significant effect of dissatisfaction for PTB voting, e.g., yet these attitudes did matter – but only for highly sophisticated voters who integrated them into their electoral calculus. A possible explanation in this respect is that, as populist parties’ rhetoric is not homogeneous, but rather relies on both a policy and more purely anti-elite messages (Mudde & Rovira Kaltwasser, 2012), it is most likely that the most receptive (i.e. sophisticated) sectors of the electorate will be the only ones where both types of information resonate and affect electoral decisions at a sufficiently high level. Of course, this also implies that, for such mechanisms to work, public debate between parties needs to provide a mixture of policy and populist contents able to mobilise voters based on such considerations (Johnston et al., 2014). Future research should further explore the role of political sophistication in moderating the effect of usual drivers of populist party support by looking more extensively at the role that contents of party competition, and more specifically of political debates, play in the whole process.

References

• Akkerman, A., Mudde, C. & Zaslove, A. (2013). How populist are the people? Measuring populist attitudes in voters. Comparative Political Studies, 47(9), 1324 -1353.

• Bakker, R., Jolly, S. & Polk, J. (2020). Multidimensional incongruence, political disaffection, and support for anti-establishment parties. Journal of European Public Policy, 27(2), 292-309.

• Baldassarri, D. & Schadee, H. (2006). Voter heuristics and political cognition in Italy: An empirical typology. Electoral Studies, 25(3), 448-466.

• Bélanger, E. & Nadeau, R. (2005). Political trust and the vote in multiparty elections: The Canadian case. European Journal of Political Research, 44(1), 121-146.

• Betz, H. G. (1994). Radical right-wing populism in Western Europe. Palgrave Macmillan.

• Billiet, J. & de Witte, H. (1995). Attitudinal dispositions to vote for a ‘new’ extreme right-wing party: The case of ‘Vlaams Blok’. European Journal of Political Research, 27(2), 181-202.

• Birch, S. & Dennison, J. (2017). How protest voters choose. Party Politics, 25(2), 110-125.

• Bischof, D. & Senninger, R. (2018). Simple politics for the people? Complexity in campaign messages and political knowledge. European Journal of Political Research, 57(2), 473-495.

• Brambor, T., Clark, W. R. & Golder, M. (2006). Understanding interaction models: Improving empirical analyses. Political Analysis, 14(1), 63-82.

• Castanho Silva, B., Vegetti, F. & Littvay, L. (2017). The elite is up to something: Exploring the relation between populism and belief in conspiracy theories. Swiss Political Science Review, 23(4), 423-443.

• Converse, P. E. (1964). The nature of belief systems in mass publics. Critical Review: A Journal of Politics and Society, 18(1-3), 1-74.

• Delwit, P. (2012). La vie politique en Belgique de 1830 à nos jours. Editions de l’Université de Bruxelles .

• de Vries, C. E., van der Brug, W., van Egmond, M. H. & van der Eijk, C. (2010). Individual and contextual variation in EU issue voting: The role of political information. Electoral Studies, 30(1), 16-28.

• De Cleen, B. & Van Aelst, P. (2016). Research on populism in Belgium. In: T. Aalberg, C. De Vreese, F. Esser, J. Strombäck & C. Reinemann (Eds.), Populist political communication in Europe. A cross-national analysis of 27 European countries (pp. 99-110). Routledge.

• Delli Carpini, M. & Keeter, S. (1996). What Americans know about politics and why it matters . Yale University Press.

• Deschouwer, K. (2009). The politics of Belgium: Governing a divided society, Palgrave Macmillan.

• Elchardus, M. & Spruyt, B. (2016). Populism, persistent republicanism and declinism: An empirical analysis of populism as a thin ideology. Government and Opposition, 51(1), 111-133.

• Ensley, M. J. (2007). Candidate divergence, ideology, and vote choice in U.S. Senate elections. American Politics Research, 35(1), 103-122.

• Gerber, D., Nicolet, S. & Sciarini, P. (2015). Voters are not fools, or are they? Party profile, individual sophistication and party choice. European Political Science Review, 7(1), 145-165.

• Geurkink, B., Zaslove, A., Sluiter, R. & Jacobs, K. (2019). Populist attitudes, political trust, and external political efficacy: Old wine in new bottles? Political Studies, 68(1), 247-267.

• Gigerenzer, G., Todd, P. M. & the ABC Research Group. (1999). Simple heuristics that make us smart. Oxford University Press .

• Goovaerts, I., Kern, A., Van Haute, E. & Marien, S. (2020). Proteststem of ideologische stem? De determinanten van het stemgedrag tijdens de verkiezingen van 2019 in België. In J. B. Pilet, A. Kern, J. Lefevere & K. Deschouwer (Eds.), De Belgen verheffen hun stem. Een analyse van het stemgedrag op 26 mei 2019, pp. 64-82. Louvain, Belgium: UCLouvain Presse.

• Hernández, E. (2018). Democratic discontent and support for mainstream and challenger parties: Democratic protest voting. European Union Politics, 19(3), 458-480.

• Hernández, E. & Kriesi, H. (2016). The electoral consequences of the financial and economic crisis in Europe. European Journal of Political Research, 55(2), 203-224.

• Hobolt, S. B. & Tilley, J. (2016). Fleeing the centre: The rise of challenger parties in the aftermath of the euro crisis. West European Politics, 39(5), 971-991.

• Hooghe, M. & Dassonneville, R. (2018a). A spiral of distrust: A panel study on the relation between political distrust and protest voting in Belgium. Government and Opposition, 53(1), 104-130.

• Hooghe, M. & Dassonneville, R. (2018b). Explaining the trump vote: The effect of racist resentment and anti-immigrant sentiments. PS: Political Science & Politics, 51(03), 528-534.

• Jackman, R. W. & Volpert, K. (1996). Conditions favouring parties of the extreme right in Western Europe. British Journal of Political Science, 26(4), 501-521.

• Jagers, J. & Walgrave, S. (2007). Populism as political communication style: An empirical study of political parties’ discourse in Belgium. European Journal of Political Research, 46(3), 319-345.

• Johnston, R., Partheymüller, J. & Schmitt-Beck. R. (2014). Activation of fundamentals in German campaigns. In B. Wessels, H. Rattinger, S. Rossteutscher & R. Schmitt-Beck (Eds.), Voters on the move or on the run? (pp. 217-237). Oxford University Press.

• Kriesi, H. (2014). The populist challenge. West European Politics, 37(2), 361-378.

• Lachat, R. (2007). A heterogeneous electorate political sophistication, predisposition strength, and the voting decision process. Nomos. doi: 10.5771/9783845204895.

• Lachat, R. (2008). The impact of party polarization on ideological voting. Electoral Studies, 27(4), 687-698.

• Lavine, H. & Gschwend, T. (2007). Issues, party and character: The moderating role of ideological thinking on candidate evaluation. British Journal of Political Science, 37(1), 139-163.

• Lupton, R., Myers, W. & Thornton, J. (2015). Political sophistication and the dimensionality of elite and mass attitudes, 1980-2004. The Journal of Politics, 77(2), 368-380.

• Luskin, R. (1987). Measuring political sophistication. American Journal of Political Science, 31, 856-899.

• Mair, P. (2002). Populist democracy vs party democracy. In Y. Mény & Y. Surel (Eds.), Democracies and the populist challenge (pp. 81-98). Palgrave.

• Meléndez, C. & Rovira Kaltwasser, C. (2019). Political identities: The missing link in the study of populism. Party Politics, 25(4), 520-533.

• Mudde, C. & Rovira Kaltwasser, C. (2012). Populism in Europe and the Americas. In C. Mudde & C. Rovira Kaltwasser (Eds.), Populism in Europe and the Americas: Threat or corrective for democracy? Cambridge University Press.

• Norris, P. (2005). Radical right: Voters and parties in the electoral market, Cambridge University Press.

• Passarelli, G. & Tuorto, D. (2018). The five star movement: Purely a matter of protest? The rise of a new party between political discontent and reasoned voting. Party Politics, 24(2), 129-140.

• Pauwels, T. (2014). Populism in Western Europe. Populism in Western Europe. Routledge.

• Rink, N., Phalet, K. & Swyngedouw, M. (2009). The effects of immigrant population size, unemployment, and individual characteristics on voting for the Vlaams Blok in Flanders 1991-1999. European Sociological Review, 25(4), 411-424.

• Rooduijn, M. (2018). What unites the voter bases of populist parties? Comparing the electorates of 15 populist parties. European Political Science Review, 10(3), 351-368.

• Rooduijn, M. & Akkerman, T. (2015). Flank attacks: Populism and left-right radicalism in Western Europe. Party Politics, 23(3), 193-204.

• Rooduijn, M., van der Brug, W. & de Lange, S. L. (2016). Expressing or fuelling discontent? The relationship between populist voting and political discontent. Electoral Studies, 43, 32-40.

• Rovira Kaltwasser, C. & Van Hauwaert, S. M. (2019). The populist citizen: Empirical evidence from Europe and Latin America. European Political Science Review, 12(1), 1-18.

• Schmitt, H., Sanz, A., & Braun, D. (2020), ‘It all happens at once: understanding electoral behaviour in second-order elections’, Politics and Governance, 8(1), 6-18.

• Schumacher, G. & Rooduijn, M. (2013). Sympathy for the ‘devil’? Voting for populists in the 2006 and 2010 Dutch general elections. Electoral Studies, 32(1), 124-133.

• Slothuus, R. & de Vreese, C. (2010). Political parties, motivated reasoning, and issue framing effects. The Journal of Politics, 72(3), 630-645.

• Spruyt, B., Keppens, G. & Van Droogenbroeck, F. (2016). Who supports populism and what attracts people to it? Political Research Quarterly, 69(2), 335-346.

• Sturgis, P. & Tilley, J. (2004). Political sophistication and issue voting: An intra-individual level analysis. ESRC Research Methods Programme, Working Paper No. 15.

• Van Der Brug, W., Fennema, M. & Tillie, J. (2000). Anti-immigrant parties in Europe: Ideological or protest vote? European Journal of Political Research, 37(1), 77-102.

• Van Hauwaert, S. M. & Van Kessel, S. (2018). Beyond protest and discontent: A cross-national analysis of the effect of populist attitudes and issue positions on populist party support. European Journal of Political Research, 57(1), 68-92.

• Van Kessel, S. (2015). Populist parties in Europe. Agents of discontent? Palgrave Macmillan.

• Van Kessel, S., Sajuria, J. & Van Hauwaert, S. M. (2020). Informed, uninformed or misinformed? A cross-national analysis of populist party supporters across European democracies. West European Politics, 1-26. DOI: 10.1080/01402382.2019.1700448.

• Vidal, G. (2017). Challenging business as usual? The rise of new parties in Spain in times of economic crisis. West European Politics, 41(2), 261-286.

• Walgrave, S., van Erkel, P., Jennart, I., Lefevere, J. & Baudewyns. P. (forthcoming). Hoe het belang van thema’s kiezers naar links of rechts drijft. In J. B. Pilet, A. Kern, J. Lefevere & K. Deschouwer (Eds.), De Belgen verheffen hun stem. Een analyse van het stemgedrag op 26 mei 2019. UCLouvainPresse .

• Wauters, B. & Pittoors, G. (2019). Populist party leaders in Belgium: An analysis of VB and PVDA-PTB. Polish Political Science Review, 7(1), 1-23.

• Weisberg, H. & Nawara, S. (2010). How sophistication affected the 2000 presidential vote: Traditional sophistication measures versus conceptualization. Politial Behavior, 32, 547-565.

• Weissberg, R. (1976). Consensual attitudes and attitude structure. Public Opinion Quarterly, 40(3), 349-359.

• Zaller, J. (1992). The nature and origins of mass opinion. Cambridge University Press.

Appendix: Full Regression Models
Table 4 Interactions between populism-related motivations and political sophistication for the case of PTB – Wallonia (binomial logistic regression)

	Wallonia – PTB
Political sophistication* Populist attitudes	-0.00251
	(0.0219)
Political sophistication*Distrust		0.0280
		(0.0201)
Political sophistication* Dissatisfaction			0.0753*
			(0.0435)
Populist attitudes	0.179*	0.173**	0.168**
	(0.0968)	(0.0683)	(0.0677)
Political sophistication	0.104	-0.140	-0.216
	(0.174)	(0.164)	(0.179)
Political distrust	0.267***	0.134	0.261***
	(0.0736)	(0.120)	(0.0723)
Dissatisfaction with democracy	0.239	0.227	-0.117
	(0.159)	(0.158)	(0.230)
Positions on free market	-0.0177	-0.0186	-0.0178
	(0.0508)	(0.0512)	(0.0507)
Positions on immigration	-0.0744*	-0.0757*	-0.0766*
	(0.0443)	(0.0443)	(0.0440)
Gender	-0.0794	-0.0924	-0.0909
	(0.225)	(0.224)	(0.224)
Age	-0.0239***	-0.0242***	-0.0240***
	(0.00755)	(0.00754)	(0.00751)
Education	-0.464***	-0.457***	-0.465***
	(0.121)	(0.119)	(0.120)
Ideological distance PTB	-0.0672	-0.0647	-0.0692
	(0.0776)	(0.0787)	(0.0785)
Constant	-2.830***	-1.614	-1.214
	(1.070)	(1.260)	(1.144)
Observations	1205	1205	1205
R-squared	0.13	0.13	0.14

*** p<0.01, ** p<0.05, * p<0.1

Table 5 Interactions between populism-related motivations and political sophistication for the case of PVDA – Flanders (binomial logistic regression)

	Flanders – PVDA
Political sophistication* Populist attitudes	0.00501
	(0.0311)
Political sophistication*Distrust		0.0103
		(0.0295)
Political sophistication* Dissatisfaction			0.0332
			(0.0465)
Populist attitudes	0.285	0.309***	0.316***
	(0.174)	(0.0831)	(0.0825)
Political sophistication	0.0697	0.0314	-0.0128
	(0.265)	(0.244)	(0.183)
Political distrust	0.256***	0.205	0.257***
	(0.0754)	(0.184)	(0.0723)
Dissatisfaction with democracy	0.192	0.186	0.00570
	(0.139)	(0.138)	(0.301)
Positions on free market	-0.0436	-0.0432	-0.0430
	(0.0532)	(0.0532)	(0.0531)
Positions on immigration	-0.241***	-0.241***	-0.243***
	(0.0475)	(0.0478)	(0.0480)
Gender	-0.0992	-0.104	-0.0984
	(0.271)	(0.271)	(0.271)
Age	-0.0242**	-0.0240**	-0.0244**
	(0.00957)	(0.00967)	(0.00955)
Education	-0.472***	-0.471***	-0.473***
	(0.157)	(0.156)	(0.156)
Ideological distance PVDA	-0.405***	-0.409***	-0.407***
	(0.0925)	(0.0945)	(0.0927)
Constant	-1.957	-1.713	-1.477
	(1.722)	(1.858)	(1.567)
Observations	1417	1417	1417
R-squared	0.20	0.20	0.20

*** p<0.01, ** p<0.05, * p<0.1

Table 6 Interactions between populism-related motivations and political sophistication for the case of VB – Flanders (binomial logistic regression)

	Flanders – VB
Political sophistication* Populist attitudes	0.0540***
	(0.0203)
Political sophistication*Distrust		0.0372**
		(0.0160)
Political sophistication* Dissatisfaction			0.139***
			(0.0330)
Populist attitudes	-0.365***	-0.123**	-0.117**
	(0.122)	(0.0518)	(0.0516)
Political sophistication	-0.430***	-0.317***	-0.581***
	(0.148)	(0.119)	(0.132)
Political distrust	0.183***	0.0187	0.199***
	(0.0465)	(0.0927)	(0.0459)
Dissatisfaction with democracy	0.674***	0.660***	-0.0404
	(0.0920)	(0.0914)	(0.200)
Positions on free market	0.0449	0.0440	0.0530
	(0.0397)	(0.0394)	(0.0390)
Positions on immigration	0.0832**	0.0753*	0.0811**
	(0.0403)	(0.0392)	(0.0390)
Gender	-0.356**	-0.366**	-0.347**
	(0.173)	(0.172)	(0.172)
Age	-0.0261***	-0.0254***	-0.0269***
	(0.00543)	(0.00547)	(0.00552)
Education	-0.456***	-0.456***	-0.455***
	(0.0877)	(0.0871)	(0.0878)
Ideological distance to VB	-0.285***	-0.284***	-0.283***
	(0.0567)	(0.0573)	(0.0558)
Constant	0.717	0.352	1.582
	(1.059)	(0.957)	(0.995)
Observations	1676	1676	1676
R-squared	0.18	0.18	0.19

*** p<0.01, ** p<0.05, * p<0.1

Table 7 Interactions between issue considerations and political sophistication (binomial logistic regression)

	Wallonia	Flanders
	PTB	PVDA	VB
Political sophistication*Positions on free market	0.0119	0.0413**
	(0.0164)	(0.0201)
Political sophistication*Positions on immigration			0.0418***
			(0.0153)
Populist attitudes	0.176**	0.337***	-0.103*
	(0.0690)	(0.0815)	(0.0529)
Political sophistication	0.0154	-0.122	-0.376***
	(0.104)	(0.129)	(0.130)
Political distrust	0.265***	0.251***	0.187***
	(0.0721)	(0.0728)	(0.0463)
Dissatisfaction with democracy	0.234	0.195	0.676***
	(0.158)	(0.135)	(0.0908)
Positions on free market	-0.0722	-0.280**	0.0448
	(0.0959)	(0.132)	(0.0400)
Positions on immigration	-0.0706	-0.241***	-0.109
	(0.0436)	(0.0472)	(0.0778)
Gender	-0.0725	-0.101	-0.360**
	(0.225)	(0.269)	(0.174)
Age	-0.0240***	-0.0242**	-0.0270***
	(0.00750)	(0.00968)	(0.00550)
Education	-0.461***	-0.469***	-0.459***
	(0.120)	(0.157)	(0.0873)
Ideological distance to PTB	-0.0651
	(0.0777)
Ideological distance to PVDA		-0.383***
		(0.0890)
Ideological distance to VB			-0.277***
			(0.0562)
Constant	-2.511***	-1.068	0.361
	(0.921)	(1.283)	(0.953)
Observations	1205	1417	1676
R-squared	0.13	0.20	0.18

*** p<0.01, ** p<0.05, * p<0.1

Noten

• 1 Data on the Brussels region has been excluded from the analyses because of the smaller sample and the only few votes cast for populist parties.

• 2 These analyses do not include FDF and FN because they are smaller parties.

• 3 For a description of the data, see the introduction to this Special Issue.

• 4 CD&V, Groen, N-VA, Open VLD, PVDA, sp.a, Vlaams Belang in Flanders, cdH, DéFI, Ecolo, MR, PP, PS, PTB in Wallonia.

• 5 Incorrect and refused answers are recoded together as 0, whilst correct answer is recoded as 1. The questions ranged covered the three categories of knowledge outlined by Delli Carpini and Keeter (1996): the rules of the game (i.e. composition of federal parliament), players of the game (i.e. chairman of federal chamber of representatives) and the substance of politics (i.e. EU involvement in policy domains).

• 6 The other three items were When making decisions, politicians care about people like me; Politics is the result of compromise and common sense; In general, politics reflect rather well the people’s preferences. Those have not been included in the index of populism since they do not strictly pertain to the anti-elitism dimension (they can be assimilated to the political efficacy and pluralist dimension, see, e.g. Akkerman et al., 2013).