Media Attention and Radical Right-Wing Populist Party Sympathy: Longitudinal Evidence From The Netherlands

Media Attention and Radical Right-Wing Populist Party Sympathy: Longitudinal Evidence From The... Abstract This contribution provides evidence for the sources of sympathies for radical right-wing populist (RRP) parties in a longitudinal perspective. We extend previous knowledge by evaluating the impact of media attention on individual changes of RRP party sympathies. To test our hypotheses, we use panel data from The Netherlands and combine it with information on the saliency of RRP parties and their issues in major Dutch newspapers. Drawing on multilevel structural equation models, our findings indicate that media attention positively affects changes in RRP party sympathies. Furthermore, we find that the effect of media attention on RRP party sympathies is enhanced by perceived ethnic threat and Euroscepticism, respectively. In sum, this study shows that media attention to RRP parties is pivotal for the fortune of such parties. Radical right-wing populist (RRP) parties and their electorate received much academic attention in the past two decades. The core and distinguishing feature of these parties is their harsh position on immigration restrictions (Immerzeel, Lubbers, & Coffé, 2015; Ivarsflaten, 2008). These parties are radical, as they oppose pluralism and diversity of ideas, right-wing regarding their stance on (non-)egalitarianism, and finally populist in their distinction between the pure people and the corrupt elite.1 Scholars agree on voters’ concern about negative consequences because of immigration as the most important motivation to prefer an RRP party (Arzheimer, 2008; Cutts, Ford, & Goodwin, 2011; Lubbers, Gijsberts, & Scheepers, 2002; Norris, 2005; Rydgren, 2008; Van der Brug, Fennema, & Tillie, 2000). Previous research mostly focused on explaining who prefers RRP parties; yet, only little is known about why individual sympathies for RRP parties change over time. The scarce evidence for changes in RRP party sympathies either focused on aggregated growth (Boomgaarden & Vliegenthart, 2007; Koopmans & Muis, 2009; Lubbers & Scheepers, 2001; Poznyak, Abts, & Swyngedouw, 2011; Vliegenthart, Boomgaarden, & Van Spanje, 2012; Walgrave & de Swert, 2004) or solely on sociodemographic and structural influences, overlooking attitudinal characteristics (Rink, Phalet, & Swyngedouw, 2009). Our study aims to improve on this gap in the literature. Despite the merits of prior research, the relative neglect of analyzing within-individual RRP party sympathy’s change is unfortunate because the underlying longitudinal mechanism remains unknown. Following scholarly suggestions, the role of mass media attention, that is, news media content, is fruitful in this endeavor (Arzheimer, 2012; Rydgren, 2005; Van der Brug & Fennema, 2007). Drawing on agenda-setting theory, mass media coverage is expected to provide the necessary political opportunity for the rise of RRP parties, perhaps even from the political fringe to government participation. The idea is that RRP parties thrive on visibility and prominence and that they depend on media attention as a breeding ground for electoral mobilization. The growing literature on sympathies for RRP parties paid rather limited attention to salience of RRP relevant news content, and existing evidence is inconclusive (but see Boomgaarden & Vliegenthart, 2007; Koopmans & Muis, 2009; Van der Pas, de Vries, & van der Brug, 2011; Vliegenthart et al., 2012). Furthermore, empirical evidence for the effect of media attention on within-individual change of RRP party sympathy remains missing and so is evidence for the interaction of media attention with the most prominent determinants in RRP party research–perceived ethnic threat and also Euroscepticism. With this study, we assess whether perceived ethnic threat or Euroscepticism reinforces the effect of salience of RRP parties on RRP party sympathy and their issues, such as immigration or the European Union. We aim to contribute to the existing literature in two significant ways. First, we provide evidence for the role of mass media attention in explanations of RRP party sympathies in an individual longitudinal perspective. Second, we investigate if, and if so, to what extent perceived ethnic threat and Euroscepticism moderate the effect of media attention on RRP party sympathy. The site of our study is The Netherlands. For several reasons, The Netherlands is a national setting that is well suited to investigate the nexus of media attention and changes of individual RRP party sympathies: first, The Netherlands witnessed the rise of rather successful and rigorous RRP parties (Vossen, 2011). Most recently, the Partij voor de Vrijheid (PVV, Party for Freedom, founded in 2006) propped up a minority government from 2010 until 2012 and emerged to an important element of the Dutch political space. Its founder and leader Geert Wilders has a reputation for his hard-line anti-Islamic stance and his strong position toward immigration restriction policies, which is in line with core characteristics of the RRP party family (Vossen, 2010, 2011). Unlike others, the PVV has no extreme right origin, but Geert Wilders’ alliance with the Front National’s front runner Le Pen shows that the PVV is no different from other RRP parties (Akkerman, de Lange, & Rooduijn, 2016). Second, The Netherlands are characterized by an increasingly volatile electorate (van der Meer, van Elsas, Lubbe, & van der Brug, 2013), that is, voters are more likely to change their vote intention, which leverages the potential of longitudinal explanations for changes in RRP party sympathies, compared with national settings with a rather stable electorate, which are presumably less likely to be affected by media attention. Theoretical Framework The evolving literature on RRP parties has identified a rather broad set of sources. Previous research finds that not only sociodemographic variables such as gender (Givens, 2004) and education (Elchardus and Spruyt, 2010) affect the appeal to RRP parties but also constructs such as relative deprivation (Elchardus and Spruyt, 2012) or relative wealth (Mols and Jetten, 2016) are important. In current research on RRP party sympathies, group threat and group conflict theory are prominent theoretical frameworks (Rydgren, 2007). Within these frameworks, perceived ethnic threat refers to symbolic issues, such as losing a cultural identity, and to competition over scarce resources, for example, employment and material security (Blalock, 1967; Sherif, Harvey, White, Hood, & Sherif, 1961). Perceptions of these cultural and economic out-group threats are expected to increase the likelihood to vote for RRP parties, as these parties proclaim to serve the interests of the self-defined in-group against migrants and ethnically defined out-groups. Previous research provided abundant evidence of this association (Berning, 2016; Berning & Schlueter, 2016; Ivarsflaten, 2008; Lucassen & Lubbers, 2012). Media Attention Why should media attention contribute to sympathies for RRP parties? Building on agenda-setting theory, issues frequently mentioned in news reports are also salient in the public debate (Yang & Stone, 2003). The media pushes the public agenda, also on information relevant for RRP party sympathies. Thus, media attention provides a fertile ground for longitudinal explanations of changes in RRP party sympathies.2 In our case, we examine media attention to RRP parties themselves and to issues that are most relevant for RRP parties: immigration and the European Union. The salience of RRP parties is expected to affect RRP party sympathies through awareness. News reports about RRP parties, immigration, and the European Union are publicity for such parties, and in turn, the potential electorate will perceive RRP parties and their issues to be more relevant. All parties mobilize their potential support partly through visibility (Oegema & Kleinnijenhuis, 2000). Nonetheless, as RRP parties generally recruit their electorate on the political margin, they largely depend on voters becoming (more) aware of their existence, their stance on immigration and the European Union, or to perceive the RRP party as more popular. The electorate of RRP parties is likely to be less politically sophisticated, informed, and literate in party programs, given their relatively low educated profile; therefore, an increase of media attention to RRP parties may strengthen RRP party sympathies (see also Dalton, 2014: 230). An important proposition is that news reports set the public agenda (McCombs & Shaw, 1972). However, this does not imply that people adapt the connotations of these news reports. Just as Cohen (1963) used to tell: “the news media may not be successful much of the time in telling people what to think, but it is stunningly successful in telling its readers what to think about” (Cohen, 1963: 13). A counterargument may be that, beyond salience and general publicity, negative press would in fact be harmful for the rise of RRP parties. Previous research shows that the tone of news media reports influences vote choice (Beck, Dalton, Greene, & Huckfeldt, 2002). Especially undecided voters are more inclined to vote for a party if the tone of media reports is positive (Hopmann, Vliegenthart, De Vreese, & Albæk, 2010). Nonetheless, Kleinnijenhuis, van Hoof, Oegema, and de Ridder (2007) find that different news types, such as criticism or salience, complement rather than confound each other. While the tone of news reports is not subject of this study and not tested empirically, we build on previous research that shows that criticism of RRP parties in the media is not related to RRP party support (Koopmans and Muis, 2009). Furthermore, the general image of news reports on RRP parties is critical and negative (Schafraad, Wester, & Scheepers, 2013), and there is evidence that in such a controversial news environment media slant has no effect on political attitudes (Gerber, Karlan, & Bergan, 2009), as only a fraction of articles evaluated politicians (Aaldering & Vliegenthart, 2016). Hence, we focus on salience, assuming that the general public debates will pick up on topics that are frequently mentioned in the news via interpersonal communication (Yang & Stone, 2003). Extant research finds that the mere visibility of political actors in the media increases their electoral potential (Oegema & Kleinnijenhuis, 2000; Semetko & Schoenbach, 1994). Salience of the European Union or immigration issues and RRP parties primes voters to strongly relate these issues to such parties and perceive both to be important. The underlying mechanism of a priming effect assumes that individuals will use information that is easily accessible or activated recently (Van der Brug, Semetko, & Valkenburg, 2007). The salience of said issues will cause individuals to use them as evaluation criteria for their electoral preferences (Iyengar & Kinder, 1987). On the one hand, individuals will perceive salient issues to be relevant and important. On the other hand, RRP parties will appear to be more qualified to face these issues. This relates to recent experimental research that provides evidence for the individual-level relationship between media cues and RRP party support (Sheets, Bos, & Boomgaarden, 2016). Thus, the combination of salient RRP parties and RRP issues, such as the European Union or immigration, will increase the support for these parties. To our knowledge, only a few studies address the link between media attention and sympathies for RRP parties. Lubbers and Scheepers (2001) evidenced with repeated cross-sectional survey data a positive effect of media attention to RRP parties on individual preferences for RRP parties in Germany. Vliegenthart et al. (2012) found similar results for Belgium, The Netherlands, and also Germany drawing on aggregated measures for media attention and RRP party support. These findings are contradicting with the Van der Pas et al. (2011) study, which might be because of differences of the trajectory span. In contrast to changes over years, Van der Pas et al. (2011) focused on weekly changes within 1 year, also drawing on aggregated data. In sum, previous research is limited to evidence of aggregated changes and somewhat inconclusive. The Interaction of Perceived Ethnic Threat and Euroscepticism With Media Attention to RRP Parties Sympathies for RRP parties arise from some people’s perception that immigrants are threatening the interests of their in-group. They are expected to use their vote choice in favor of their policy preference for restrictive immigration legislation (Mughan & Paxton, 2006). Along this line of explanation, Eurosceptic voters are also expected to support RRP parties in rejection of European integration (Werts, Scheepers, & Lubbers, 2012). This motivation, that is, an attitudinal proximity of voters to their party, is by no means different to other parties (Van der Brug et al., 2000). Explanations for RRP party sympathy as nonideological protest find only limited support in the literature (Swyngedouw, 2001; van der Brug & Fennema, 2003, but see Schumacher & Rooduijn, 2013). To strengthen the postulated assumption of a genuine effect of media attention on sympathies for RRP parties, we turn to the interaction of media attention with perceived ethnic threat or Euroscepticism and their relation to sympathies for RRP parties. In fact, testing this interaction is important not only analytically, as such as the nature or strength of the relationship depends on perceived ethnic threat or Euroscepticism, respectively, but also for a more integrated understanding of how these concepts are linked. We assume that media attention does not only affect sympathies for RRP parties directly but also reinforces the effect of perceived ethnic threat and Euroscepticism on RRP party sympathies (Lubbers & Scheepers, 2001: 434). An increase of RRP parties and their message in the news likely heightens the voters’ perceptions that RRP parties’ issues are more relevant and that their perception of ethnic threat and skepticism about the European integration are sensible. This in turn may contribute to an urge to alleviate these concerns by sympathizing with RRP parties. Put differently, the RRP electorate, who perceives immigrants and the European Union as threatening, is likely to feel that they are right about their perception when RRP parties are frequently mentioned in the media. Furthermore, an increase in salience of RRP parties may trigger the link between RRP party sympathies and perceived ethnic threat, as well as Euroscepticism. An increase in awareness of those parties may simply contribute to the feeling that RRP parties are a good representation of people who perceive immigrants as threatening and are against the European Union. Following the priming argument, salience of RRP parties and their issues will especially resonate with individuals who hold Eurosceptic attitudes and perceive immigrants as threatening, as individuals will interpret new information based on their preexisting attitudes (Tesler, 2015). Hypotheses Our theoretical assumptions described above are tested with the following hypotheses: Hypothesis 1: The higher the media attention to RRP parties, the more people sympathize with RRP parties. Hypothesis 2: The higher the media attention to immigration, the more people sympathize with RRP parties. Hypothesis 3: The positive effect of media attention to RRP parties on RRP party sympathies increases, when media attention to immigration is more salient. Hypothesis 4: The positive effect of media attention to RRP parties on RRP party sympathies increases, when perceived ethnic threat is larger. Hypothesis 5: The higher the media attention to the European Union, the more people sympathize with RRP parties. Hypothesis 6: The positive effect of media attention to RRP parties on RRP party sympathies increases, when media attention to the European Union is more salient. Hypothesis 7: The positive effect of media attention to RRP parties on RRP party sympathies increases, when Euroscepticism is larger. Data and Methods Data To examine our theoretical assumptions, we need two different data sources. We combine individual-level multiwave panel data with longitudinal information from computer-assisted content analyses of newspaper articles. Regarding the individual-level data, we make use of the Longitudinal Internet Studies for the Social Sciences (LISS) administered by CentERdata (Tilburg University, The Netherlands). The LISS panel is a representative sample of Dutch citizens, drawn from the population based on a true probability sample (Scherpenzeel & Das, 2010).3 The data are available for six waves from 2008 to 2013. We limit our analyses to respondents who participated at least in two waves of the study period.4 As we focus on attitudes of majority group members, we excluded respondents with a migration background. The data structure follows a long format with Nobs. = 25,110 observations nested in Nind. = 6,184 individuals. Measures RRP party sympathy. We use two items to measure RRP party sympathy.5 For the first item, respondents were asked to indicate their sympathy for the PVV: “What do you think of the Party for Freedom?” The second item gauges the evaluation of the party leader by asking: “What do you think of Geert Wilders?” Response options for both items were given on an 11-point scale and ranged from 0 (very unsympathetic) to 10 (very sympathetic). We then averaged both items to a single index of respondents’ RRP party sympathy, with higher values indicating higher RRP party sympathy. Reliability of this measure was high (Cronbach’s alpha = .96). This measure has several advantages compared with discrete choice conceptualizations. For instance, it allows respondents to sympathize with more than one party, which is especially relevant in The Netherlands with many parties competing for votes. It measures even small differences in party preferences, a momentary measure of party ratings, and it resembles a construct, which is not only relevant during elections but rather captures the underlying party utility function. Conceptually, it refers to party preference, which is the central intervening variable between reasoning about the party and actual voting behavior (Pappi, 1996). Media attention. To assess media attention, we conducted computer-assisted content analyses of five major Dutch newspapers. Specifically, we compiled the number of articles published in De Telegraaf, AD/Algemeen Dagblad, De Volkskrant, NRC Handelsblad, and Trouw for RRP party related news. Both De Telegraaf and AD/Algemeen Dagblad are considered as more right-wing leaning, tabloid kind of newspapers. De Telegraaf is the largest newspaper in the country, with a circulation of half a million-printed copies. Nonetheless, the online version reaches a much larger audience. The other three papers are described as quality papers, with the Volkskrant more left-wing leaning, the NRC Handelsblad more social–liberal, and the Trouw with more attention to religiosity. To gauge the number of articles related to the RRP party, we electronically searched for the terms “PVV” or “Wilders.” To measure salience of immigration related news, we build on previous research and used a search string suggested and tested by Vliegenthart et al. (2012: 335). Finally, we assessed media attention to the European Union, by searching all five newspapers for articles about the “European Union” or the “EU.”6 We analyzed all articles published up to 8 weeks before the interview for each wave and each person. This period is long enough to let news reports influence the public’s agenda and helps to reduce the risk of simultaneity bias (Schlueter and Davidov, 2013).7 Moreover, measuring media attention before the interviews is in line with our assumption regarding the temporal order, that is, media attention precedes RRP party sympathies.8 To probe the validity of our data, we manually reviewed a small number of randomly chosen newspaper articles from the total sample of the newspaper articles we collected. Consistent with our search criteria, the small subset of articles turned out to cover indeed issues pertaining to the PVV/Wilders, immigration, or the European Union, respectively. To aid in interpretation of the unstandardized regression coefficients, we transformed the frequency scores to range from a minimum of 0 to a maximum of 1. Thus, a one-unit change in these variables equals the difference between the lowest and the highest possible score. The individual-level panel data were collected in 6 consecutive years. However, as it is common in survey research, not all interviews were conducted on the same date. In total, the LISS panel consists of 311 different interviews dates (see Supplementary Appendix Figure A). We therefore compiled 933 search strings for all media attention measures, that is, we applied our three media attention measure (PVV/Wilders, immigration, and the European Union) to 311 different time intervals. The individual-level panel data were then merged with the results of our computer-assisted content analyses, that is, we assigned three frequency scores to each interview (date). For example, Respondent A answered the questionnaire on December 3, 2007 and Respondent B did so on March 4, 2008, with both respondents being interviewed during the first wave. Thus, for Respondent A, the results from the content analyses refer to the 8 weeks period before December 3, 2007, while for Respondent B the frequency scores refer to the 8 weeks period before March 4, 2008. Perceived ethnic threat and Euroscepticism. To operationalize perceived ethnic threat, we use three items. For the first two items, respondents were asked to evaluate the following statements on a five-point Likert type scale: “There are too many people of foreign origin or descent in The Netherlands” and “Some sectors of the economy can only continue to function because people of foreign origin or descent work there.” Answer options ranged from 1 (fully disagree) to 5 (fully agree). These items correspond to operationalizations of perceived ethnic threat in related research (Scheepers, Gijsberts, & Coenders, 2002; Schlueter, Meuleman, & Davidov, 2013; Semyonov, Raijman, & Gorodzeisky, 2006). Additionally, respondents were asked: “Where would you place yourself on a scale of 1 to 5, where 1 means that immigrants retain their own culture and 5 means that they should adapt entirely?” After inverting the second item, higher values of all measures reflect higher ethnic threat perception. To gauge the effect of Euroscepticism, we used respondents’ answers to the following question: “Where would you place yourself on a scale from 1 to 5, where 1 means that European unification should go further and 5 means that it has already gone too far?” Control variables. Previous research shows that men are more likely to sympathize with RRP parties (Givens, 2004). Hence, we control for gender using a dichotomous measure with 0 for female and 1 for male. Age is expected to have a negative influence (Coffé and Voorpostel, 2010). We capture this by including year of birth as a continuous measure. There is also evidence that low education increases RRP party sympathy (Betz, 1994). We measure education on a three-point scale with 1 for primary school, vmbo (intermediate secondary education), 2 for havo/vwo (higher secondary education/preparatory university education), mbo (intermediate vocational education), and 3 for hbo (higher vocational education), wo (university). Further, we control for the effect of economic developments with an individual measure of unemployment (Lancee and Pardos-Prado, 2013), measured by a dichotomous variable with 0 (not unemployed) and 1 (unemployed). To control for the potential confounding factor of media exposure, we include a dichotomous item, asking respondents: “Do you follow the news in a bought newspaper or one that you have a subscription to?” Answers were coded as 0 (no) and 1 (yes). Descriptive statics for all control variables are presented in Supplementary Table A. Method We use multilevel structural equation modeling (MLSEM) for longitudinal data to test our hypotheses (Song, Lee, & Hser, 2008). This approach offers several advantages. First, we can take the nested data structure into account, that is, observations nested within individuals, which yields to adjusted standard errors, and we are able to decompose the total variance of RRP party sympathy into within and between components. The within component refers to intraindividual variance and the between component is based on interindividual variance (Curran & Bauer, 2011). Second, our analytical approach does not compel balanced data, that is, individuals do not require having the same number of observations (Hox, 2002). Third, this technique allows us to measure perceived ethnic threat as a latent variable at both levels, adequately accounting for sampling and measurement error (Marsh et al., 2009). Moreover, simultaneously modeling relationships at two levels enables us to deconflate estimates to between- and within-individual effects (Zhang, Zyphur, & Preacher, 2009). All models are based on full information maximum likelihood estimates, available in Mplus 7 (Muthén & Muthén, 1998-2012). To assess the goodness of fit of our measurement model, we use multiple fit indices: χ2/df ratio, the comparative fit index (CFI), and the root mean square error of approximation (RMSEA), and the standardized root mean square residual (SRMR) (Bentler, 1990; Boomsma, 2000; Hu & Bentler, 1999; Marsh & Hocevar, 1985). We consider models with χ2/df < 5, RMSEA < .06, SRMR(within; between) <.06, and CFI > .95 to have a good fit to the data. Our structural models include random slopes, which imply that the variance of the dependent variable varies across observations, with no single covariance matrix for the overall model. Thus, for our structural models, the model fit indices described above are not available. We measure perceived ethnic threat with three items as one latent factor both at the within- and between-person level of analysis. To ensure that this factor of perceived ethnic threat is in fact the same construct measured on both levels, we need to establish metric cross-level invariance (Marsh et al., 2009). We use Satorra–Bentler scaled χ2 difference tests (Satorra & Bentler, 2001) for model comparison. In favor for interpretation of our structural models, we centered year of birth on its mean and following the suggestions of Enders and Tofighi (2007), we facilitate group mean centering for all other variables, except dichotomous measures. Results We begin presenting our results by describing the trends of media attention and RRP party sympathies in The Netherlands. We then briefly discuss the results of our measurement model for perceived ethnic threat, and continue with the results of our multilevel structural equation models. Media Attention and RRP Party Sympathies in The Netherlands Figure 1 depicts the average frequency of newspaper articles mentioning Geert Wilders or his PVV (dashed-dotted line), newspaper articles on immigration (dashed line), newspaper articles mentioning the European Union (dotted line), as well as the aggregated scores of RRP party sympathies (solid line) per year. The graph shows that sympathies for an RRP party largely follow the trend of media attention to Wilders and the PVV. Both lines start rather low and build up to their peak in 2011. All respondents were interviewed around the turn of the year, that is, the maxima at 2011 were about 2–3 months after the right-wing government formation of the VVD and the Christen-Democratisch Appèl (CDA, Christian Democratic Appel), propped up by the PVV. Figure 1 View largeDownload slide RRP party sympathies and media attention. Solid line depicts RRP party sympathy; dash-dotted line represents media attention to PVV/Wilders; dashed line represents media attention to immigration; dotted line indicates media attention to the European Union Figure 1 View largeDownload slide RRP party sympathies and media attention. Solid line depicts RRP party sympathy; dash-dotted line represents media attention to PVV/Wilders; dashed line represents media attention to immigration; dotted line indicates media attention to the European Union The formation of the government took about 4 months preceding the general elections in June 2010 with Geert Wilders’ PVV winning the third largest share of seats. After both trends rocketed in 2011, they significantly sank in 2012 and plummeted in 2013, right after the government resolved prematurely in mid-2012. In the general elections in September 2012, the PVV once more came in third; yet this time, VVD and the Partij van de Arbeid (PvdA, Labor Party) were able to form a government without the PVV’s support. Generally, media attention to the European Union and immigration is much more balanced throughout the years. Media attention to the European Union peaks in 2012, after Wilders’ anti-European Union campaign for the general elections. Nevertheless, media attention to immigration and the European Union follows a similar trend as RRP party sympathies. Collectively, these descriptive findings suggest that changes of media attention to RRP parties clearly relate to changes in sympathies for such parties. Further, this association appears much more pronounced than the nexus between media attention to immigration or the European Union with RRP party sympathy. However, these conclusions are based on aggregated data only. To achieve a better understanding if and to what extent media attention shapes RRP sympathy within persons over time, we now turn to the results from hypothesis testing using multilevel structural equation modeling. Measurement Model As mentioned earlier, we measure perceived ethnic threat with a latent variable on two levels, that is, within and between respondents. The results show that our measurement model fits the data really well with χ2 = 5.896, df = 2, p-value = .052, χ2/df = 2.948, RMSEA = .008, CFI = .999, SRMRwithinperson = .007, and SRMRbetweenperson = .003. Further, comparing a model with all factor loadings within and between persons restricted to be equal did not fit worse to the data than the initial model with all factor loadings estimated freely (Δχ2 = 5.896; Δdf = 2; n.s.). This establishes metric measurement invariance, which ensures that the factor loadings reflect the same meaning within and between persons. Interestingly, we find that most variance in respondents’ perceived ethnic threat is situated at the between-person level (ICC = .934). This means that only about 6.5% of all variation of perceived ethnic threat is attributable to differences within persons over time. Nevertheless, because the results show significant within-person variance, we proceed with a decomposed measure of perceived ethnic threat. Structural Model We now present the results of our hypotheses tests. An initial model (not presented) without explanatory variables shows that 76.3% of the total variance of RRP party sympathy is located at the individual level. In other words, 23.7% of differences in sympathies for RRP parties are because of differences over time within individuals. In Table 1, we present the results of our structural model testing Hypotheses 1–4. Model 1a depicts the results for the test of Hypotheses 1 and 2. Substantially, we find a significant positive effect of media attention to Wilders/PVV, as well as media attention to immigration on RRP party sympathies. With this, we can reaffirm our bivariate observations in respect to Hypotheses 1 and 2. While standardized effects are not available for random slope models, we reran Model 1a as a random intercept model to compare the standardized effects of media attention with Wilders/PVV and immigration. The data reveal that media attention to Wilders/PVV exerts a considerably stronger influence (β = 0.165) on respondents’ RRP party sympathy than media attention to immigration (β = 0.054). In additional analyses, we separately included the effects of media attention to either Wilders/PVV or immigration to evaluate the proportion of explained variance. We find that media attention to Wilders/PVV accounts for 11.7% of the within-person variance of RRP party sympathy, and media attention to immigration only accounts for 2.3% of the variance. Table 1 Multilevel Structural Equation Models: RRP Party Sympathy Regressed on Media Attention to PVV/Wilders and Media Attention to Immigration Variable Model 1a Model 1b Model 1c b SE b SE b SE Within person     PVV/Wilders media 0.806 0.035*** 0.299 0.110** 0.78 0.040***     Immigration media 0.294 0.032*** 0.199 0.076** 0.4 0.059***     Unemployed 0.028 0.098 0.036 0.098 0.037 0.098     Media exposure −0.032 0.038 −0.036 0.038 −0.037 0.038     Euroscepticism 0.038 0.014** 0.038 0.014** 0.038 0.014*     Perceived threat 1.363 0.165*** 1.362 0.166*** 1.343 0.165***     PVV/Wilders media × Immigration media – 0.928 0.205*** − Between person     Intercept 2.064 0.134*** 2.131 0.159*** 2.268 0.154***     Male 0.371 0.050*** 0.362 0.050*** 0.365 0.050***     Year of birth 0.018 0.002*** 0.018 0.002*** 0.018 0.002***     Low education – – –     Med education −0.114 0.062 −0.115 0.062 −0.118 0.062     High education −0.232 0.064*** −0.245 0.064*** −0.246 0.065***     Unemployed −0.059 0.337 −0.049 0.338 −0.043 0.338     Media exposure −0.165 0.076* −0.177 0.076* −0.177 0.077*     Euroscepticism 0.294 0.032*** 0.236 0.038*** 0.239 0.038***     Perceived threat 2.277 0.056*** 2.308 0.058*** 2.333 0.057***     PVV/Wilders media × perceived threat – – 0.419 0.052*** Variance within 1.413 0.035*** 1.408 0.035*** 1.418 0.035*** Variance between 2.783 0.068*** 2.863 0.129*** 2.764 0.068*** Variance slope PVV/Wilders 1.091 0.136*** 1.055 0.316** 1.018 0.136*** Variance slope Immigration 0.446 0.299 0.492 0.317 0.317 0.295 Variance slope PVV/Wilders × Immigration – 0.158 0.92 – Cov (slope PVV/Wilders, intercept) 0.33 0.47 0.26 Cov (slope Immigration, intercept) 0.04 0.1 0.04 Cov (slope PVV/Wilders × Immigration, intercept) – −0.291 − Log likelihood −156,695.152 −156,643.606 −156,663.151 AIC 313,456.304 313,361.212 313,396.302 BIC 313,724.628 313,662.06 313,680.888 Variable Model 1a Model 1b Model 1c b SE b SE b SE Within person     PVV/Wilders media 0.806 0.035*** 0.299 0.110** 0.78 0.040***     Immigration media 0.294 0.032*** 0.199 0.076** 0.4 0.059***     Unemployed 0.028 0.098 0.036 0.098 0.037 0.098     Media exposure −0.032 0.038 −0.036 0.038 −0.037 0.038     Euroscepticism 0.038 0.014** 0.038 0.014** 0.038 0.014*     Perceived threat 1.363 0.165*** 1.362 0.166*** 1.343 0.165***     PVV/Wilders media × Immigration media – 0.928 0.205*** − Between person     Intercept 2.064 0.134*** 2.131 0.159*** 2.268 0.154***     Male 0.371 0.050*** 0.362 0.050*** 0.365 0.050***     Year of birth 0.018 0.002*** 0.018 0.002*** 0.018 0.002***     Low education – – –     Med education −0.114 0.062 −0.115 0.062 −0.118 0.062     High education −0.232 0.064*** −0.245 0.064*** −0.246 0.065***     Unemployed −0.059 0.337 −0.049 0.338 −0.043 0.338     Media exposure −0.165 0.076* −0.177 0.076* −0.177 0.077*     Euroscepticism 0.294 0.032*** 0.236 0.038*** 0.239 0.038***     Perceived threat 2.277 0.056*** 2.308 0.058*** 2.333 0.057***     PVV/Wilders media × perceived threat – – 0.419 0.052*** Variance within 1.413 0.035*** 1.408 0.035*** 1.418 0.035*** Variance between 2.783 0.068*** 2.863 0.129*** 2.764 0.068*** Variance slope PVV/Wilders 1.091 0.136*** 1.055 0.316** 1.018 0.136*** Variance slope Immigration 0.446 0.299 0.492 0.317 0.317 0.295 Variance slope PVV/Wilders × Immigration – 0.158 0.92 – Cov (slope PVV/Wilders, intercept) 0.33 0.47 0.26 Cov (slope Immigration, intercept) 0.04 0.1 0.04 Cov (slope PVV/Wilders × Immigration, intercept) – −0.291 − Log likelihood −156,695.152 −156,643.606 −156,663.151 AIC 313,456.304 313,361.212 313,396.302 BIC 313,724.628 313,662.06 313,680.888 Note: *p < .05; **p < .01; ***p < .001; PVV/Wilders media = media attention to the PVV/Wilders; Immigration media = media attention to immigration. Table 1 Multilevel Structural Equation Models: RRP Party Sympathy Regressed on Media Attention to PVV/Wilders and Media Attention to Immigration Variable Model 1a Model 1b Model 1c b SE b SE b SE Within person     PVV/Wilders media 0.806 0.035*** 0.299 0.110** 0.78 0.040***     Immigration media 0.294 0.032*** 0.199 0.076** 0.4 0.059***     Unemployed 0.028 0.098 0.036 0.098 0.037 0.098     Media exposure −0.032 0.038 −0.036 0.038 −0.037 0.038     Euroscepticism 0.038 0.014** 0.038 0.014** 0.038 0.014*     Perceived threat 1.363 0.165*** 1.362 0.166*** 1.343 0.165***     PVV/Wilders media × Immigration media – 0.928 0.205*** − Between person     Intercept 2.064 0.134*** 2.131 0.159*** 2.268 0.154***     Male 0.371 0.050*** 0.362 0.050*** 0.365 0.050***     Year of birth 0.018 0.002*** 0.018 0.002*** 0.018 0.002***     Low education – – –     Med education −0.114 0.062 −0.115 0.062 −0.118 0.062     High education −0.232 0.064*** −0.245 0.064*** −0.246 0.065***     Unemployed −0.059 0.337 −0.049 0.338 −0.043 0.338     Media exposure −0.165 0.076* −0.177 0.076* −0.177 0.077*     Euroscepticism 0.294 0.032*** 0.236 0.038*** 0.239 0.038***     Perceived threat 2.277 0.056*** 2.308 0.058*** 2.333 0.057***     PVV/Wilders media × perceived threat – – 0.419 0.052*** Variance within 1.413 0.035*** 1.408 0.035*** 1.418 0.035*** Variance between 2.783 0.068*** 2.863 0.129*** 2.764 0.068*** Variance slope PVV/Wilders 1.091 0.136*** 1.055 0.316** 1.018 0.136*** Variance slope Immigration 0.446 0.299 0.492 0.317 0.317 0.295 Variance slope PVV/Wilders × Immigration – 0.158 0.92 – Cov (slope PVV/Wilders, intercept) 0.33 0.47 0.26 Cov (slope Immigration, intercept) 0.04 0.1 0.04 Cov (slope PVV/Wilders × Immigration, intercept) – −0.291 − Log likelihood −156,695.152 −156,643.606 −156,663.151 AIC 313,456.304 313,361.212 313,396.302 BIC 313,724.628 313,662.06 313,680.888 Variable Model 1a Model 1b Model 1c b SE b SE b SE Within person     PVV/Wilders media 0.806 0.035*** 0.299 0.110** 0.78 0.040***     Immigration media 0.294 0.032*** 0.199 0.076** 0.4 0.059***     Unemployed 0.028 0.098 0.036 0.098 0.037 0.098     Media exposure −0.032 0.038 −0.036 0.038 −0.037 0.038     Euroscepticism 0.038 0.014** 0.038 0.014** 0.038 0.014*     Perceived threat 1.363 0.165*** 1.362 0.166*** 1.343 0.165***     PVV/Wilders media × Immigration media – 0.928 0.205*** − Between person     Intercept 2.064 0.134*** 2.131 0.159*** 2.268 0.154***     Male 0.371 0.050*** 0.362 0.050*** 0.365 0.050***     Year of birth 0.018 0.002*** 0.018 0.002*** 0.018 0.002***     Low education – – –     Med education −0.114 0.062 −0.115 0.062 −0.118 0.062     High education −0.232 0.064*** −0.245 0.064*** −0.246 0.065***     Unemployed −0.059 0.337 −0.049 0.338 −0.043 0.338     Media exposure −0.165 0.076* −0.177 0.076* −0.177 0.077*     Euroscepticism 0.294 0.032*** 0.236 0.038*** 0.239 0.038***     Perceived threat 2.277 0.056*** 2.308 0.058*** 2.333 0.057***     PVV/Wilders media × perceived threat – – 0.419 0.052*** Variance within 1.413 0.035*** 1.408 0.035*** 1.418 0.035*** Variance between 2.783 0.068*** 2.863 0.129*** 2.764 0.068*** Variance slope PVV/Wilders 1.091 0.136*** 1.055 0.316** 1.018 0.136*** Variance slope Immigration 0.446 0.299 0.492 0.317 0.317 0.295 Variance slope PVV/Wilders × Immigration – 0.158 0.92 – Cov (slope PVV/Wilders, intercept) 0.33 0.47 0.26 Cov (slope Immigration, intercept) 0.04 0.1 0.04 Cov (slope PVV/Wilders × Immigration, intercept) – −0.291 − Log likelihood −156,695.152 −156,643.606 −156,663.151 AIC 313,456.304 313,361.212 313,396.302 BIC 313,724.628 313,662.06 313,680.888 Note: *p < .05; **p < .01; ***p < .001; PVV/Wilders media = media attention to the PVV/Wilders; Immigration media = media attention to immigration. We include an interaction term of media attention to Wilders/PVV and immigration in Model 1b, to test our expectation regarding their joint effect on RRP party sympathies (Hypothesis 3). The results reveal a significant interaction effect. Figure 2 presents the marginal effects of media attention to immigration by media attention to Wilders/PVV. The results show that the positive effect of media attention to immigration is only significant when media attention to Wilders/PVV is at least at an average level. The effect increases with the level of media attention to Wilders/PVV. These findings support Hypothesis 3. Including the interaction between media attention to Wilders/PVV and immigration decreases the within-person variance of RRP party sympathy by only 0.3%. Figure 2 View largeDownload slide Marginal effect of media attention to immigration on RRP party sympathy by media attention to PVV/Wilders (Model 1b). Note: Dashed lines are 95% confidence intervals Figure 2 View largeDownload slide Marginal effect of media attention to immigration on RRP party sympathy by media attention to PVV/Wilders (Model 1b). Note: Dashed lines are 95% confidence intervals In Models 1a and 1b, we find that the effect of media attention to Wilders/PVV significantly varies across individuals. Thus, we proceed by testing Hypothesis 4, the interaction of media attention to Wilders/PVV with perceived ethnic threat on RRP party sympathies. Thereby, we are testing if the variability of media attention’s effect is depending on an individual level of ethnic threat perception, that is, the between-person level variable. Notably, neither the effect of media attention to immigration nor the interaction between both salience measures shows significant slope variance. Model 1c reports the interaction effect between media attention to Wilders/PVV and perceived ethnic threat on RRP party sympathies. The results show a positive and significant interaction. To interpret the underlying effect, we turn to the marginal effects plot presented in Figure 3. The plot shows that the positive effect of perceived ethnic threat is larger, when media attention to Wilders/PVV increases, as we postulated in Hypothesis 4. The interaction effect between media attention to Wilders/PVV and perceived ethnic threat explains 6.7% of the slope variation of the effect of media attention to Wilders/PVV. In other words, 6.7% of the differences of the effect of media attention to Wilders/PVV on RRP party sympathies are explained by perceived ethnic threat. Because conventional measures of variance explained are considered less suitable to gauge the relevance of cross-level interactions, we consider the differences of BIC between a model without versus a model with the cross-level interaction. Comparing Model 1a with Model 1c, we find a rather large difference of ΔBIC = 43.74, which is strong evidence that Model 1c is the preferred model (Raftery, 1995). Figure 3 View largeDownload slide Marginal effect of perceived ethnic threat on RRP party sympathy by media attention to PVV/Wilders (Model 1c). Note: Dashed lines are 95% confidence intervals Figure 3 View largeDownload slide Marginal effect of perceived ethnic threat on RRP party sympathy by media attention to PVV/Wilders (Model 1c). Note: Dashed lines are 95% confidence intervals Table 2 presents the results for our tests of Hypotheses 5–7. In Model 2a, we find that media attention to the European Union has a significant positive effect on RRP party sympathy. As in Table 1, we find a significant positive effect of media attention to Wilders/PVV on RRP party sympathy. The explanatory contribution of media attention to the European Union is only marginal. To compare the strength of the effects of media attention with Wilders/PVV and the European Union, we reran Model 2a with fixed slopes. We find that the standardized effects of media attention to the European Union on respondents’ RRP party sympathy (β = 0.019) are much weaker than the effect of media attention to Wilders/PVV (β = 0.185). Including the effect of media attention to the European Union into the model explains only 0.9% of the within individual variance of RRP party sympathy. Turning to the interaction effect between media attention to the European Union and media attention to Wilders/PVV, tested in Model 2b, we find a significant negative effect. The findings of the interaction effect are presented in Figure 4. Contrary to our expectations of Hypothesis 6, the marginal effects plot shows that the effect of media attention to the European Union slightly decreases, when media attention to Wilders/PVV increases. The interaction effect between media attention to Wilders/PVV and the European Union decreases the variance of RRP party sympathy by only 0.1%. We find no significant slope variance for media attention to the European Union and thus proceed with the test of the interaction effect between media attention to Wilders/PVV and Euroscepticism (Hypothesis 7). The results of Model 2c show a significant positive interaction between Euroscepticism and media attention to Wilders and his PVV. Figure 5 displays the marginal effects of Euroscepticism by media attention to Wilders/PVV. We find that the positive effect of Euroscepticism is larger, when media attention to Wilders/PVV increases. The explanatory power of the interaction effect between media attention to Wilders/PVV and Euroscepticism is only limited. Including the moderating effect decreases the slope variance of the effect of media attention to Wilders/PVV on RRP party sympathy by only 3.0%. However, a BIC difference of ΔBIC = 11.92 between Model 2a and Model 2c suggests that including the interaction substantively increases the model fit. Table 2 Multilevel Structural Equation Models: RRP Party Sympathy Regressed on Media Attention to PVV/Wilders and Media Attention to the European Union Variable Model 2a Model 2b Model 2c b SE b SE b SE Within person     PVV/Wilders media 0.866 0.035*** 1.112 0.105*** 0.201 0.14     EU media 0.171 0.053** 0.348 0.100** 0.169 0.053**     Unemployed 0.013 0.097 0.019 0.097 0.012 0.097     Media exposure −0.023 0.038 −0.029 0.038 −0.022 0.038     Euroscepticism 0.036 0.014* 0.039 0.014* 0.034 0.014*     Perceived threat 1.392 0.166*** 1.38 0.166*** 1.398 0.166***     PVV/Wilders media × EU media – −0.427 0.180* – Between person     Intercept 2.272 0.154*** 2.385 0.159*** 2.232 0.154***     Male 0.364 0.050*** 0.364 0.050*** 0.364 0.050***     Year of birth 0.018 0.002*** 0.018 0.002*** 0.018 0.002***     Low education – – –     Med education −0.115 0.062 −0.119 0.062 −0.115 0.062     High education −0.243 0.064*** −0.249 0.064*** −0.242 0.064***     Unemployed −0.037 0.337 −0.053 0.338 −0.036 0.338     Media exposure −0.179 0.076* −0.169 0.076* −0.179 0.077*     Euroscepticism 0.238 0.038*** 0.23 0.038*** 0.25 0.038***     Perceived threat 2.312 0.057*** 2.316 0.057*** 2.312 0.057*** Media attention × Euroscepticism – – 0.192 0.041*** Variance within 1.433 0.033*** 1.431 0.033*** 1.433 0.033*** Variance between 2.766 0.068*** 2.929 0.135*** 2.765 0.068*** Variance slope PVV/Wilders 1.066 0.135*** 1.008 0.256*** 1.034 0.133*** Variance slope EU 0.064 0.476 0.039 0.504 0.063 0.474 Variance slope PVV/Wilders × EU – 0.149 0.522 – Cov (slope PVV/Wilders, intercept) 0.36 0.57 0.35 Cov (slope EU, intercept) −0.11 −0.01 −0.11 Cov (slope PVV/Wilders × EU, intercept) – −0.39 – Log likelihood −156,790.602 −156,740.085 −156,779.57 AIC 313,649.203 313,554.171 313,629.141 BIC 313,925.658 313,855.018 313,913.726 Variable Model 2a Model 2b Model 2c b SE b SE b SE Within person     PVV/Wilders media 0.866 0.035*** 1.112 0.105*** 0.201 0.14     EU media 0.171 0.053** 0.348 0.100** 0.169 0.053**     Unemployed 0.013 0.097 0.019 0.097 0.012 0.097     Media exposure −0.023 0.038 −0.029 0.038 −0.022 0.038     Euroscepticism 0.036 0.014* 0.039 0.014* 0.034 0.014*     Perceived threat 1.392 0.166*** 1.38 0.166*** 1.398 0.166***     PVV/Wilders media × EU media – −0.427 0.180* – Between person     Intercept 2.272 0.154*** 2.385 0.159*** 2.232 0.154***     Male 0.364 0.050*** 0.364 0.050*** 0.364 0.050***     Year of birth 0.018 0.002*** 0.018 0.002*** 0.018 0.002***     Low education – – –     Med education −0.115 0.062 −0.119 0.062 −0.115 0.062     High education −0.243 0.064*** −0.249 0.064*** −0.242 0.064***     Unemployed −0.037 0.337 −0.053 0.338 −0.036 0.338     Media exposure −0.179 0.076* −0.169 0.076* −0.179 0.077*     Euroscepticism 0.238 0.038*** 0.23 0.038*** 0.25 0.038***     Perceived threat 2.312 0.057*** 2.316 0.057*** 2.312 0.057*** Media attention × Euroscepticism – – 0.192 0.041*** Variance within 1.433 0.033*** 1.431 0.033*** 1.433 0.033*** Variance between 2.766 0.068*** 2.929 0.135*** 2.765 0.068*** Variance slope PVV/Wilders 1.066 0.135*** 1.008 0.256*** 1.034 0.133*** Variance slope EU 0.064 0.476 0.039 0.504 0.063 0.474 Variance slope PVV/Wilders × EU – 0.149 0.522 – Cov (slope PVV/Wilders, intercept) 0.36 0.57 0.35 Cov (slope EU, intercept) −0.11 −0.01 −0.11 Cov (slope PVV/Wilders × EU, intercept) – −0.39 – Log likelihood −156,790.602 −156,740.085 −156,779.57 AIC 313,649.203 313,554.171 313,629.141 BIC 313,925.658 313,855.018 313,913.726 Note: *p < .05; **p < .01; ***p < .001; PVV/Wilders media = media attention to the PVV/Wilders; EU media = media attention to the European Union. Table 2 Multilevel Structural Equation Models: RRP Party Sympathy Regressed on Media Attention to PVV/Wilders and Media Attention to the European Union Variable Model 2a Model 2b Model 2c b SE b SE b SE Within person     PVV/Wilders media 0.866 0.035*** 1.112 0.105*** 0.201 0.14     EU media 0.171 0.053** 0.348 0.100** 0.169 0.053**     Unemployed 0.013 0.097 0.019 0.097 0.012 0.097     Media exposure −0.023 0.038 −0.029 0.038 −0.022 0.038     Euroscepticism 0.036 0.014* 0.039 0.014* 0.034 0.014*     Perceived threat 1.392 0.166*** 1.38 0.166*** 1.398 0.166***     PVV/Wilders media × EU media – −0.427 0.180* – Between person     Intercept 2.272 0.154*** 2.385 0.159*** 2.232 0.154***     Male 0.364 0.050*** 0.364 0.050*** 0.364 0.050***     Year of birth 0.018 0.002*** 0.018 0.002*** 0.018 0.002***     Low education – – –     Med education −0.115 0.062 −0.119 0.062 −0.115 0.062     High education −0.243 0.064*** −0.249 0.064*** −0.242 0.064***     Unemployed −0.037 0.337 −0.053 0.338 −0.036 0.338     Media exposure −0.179 0.076* −0.169 0.076* −0.179 0.077*     Euroscepticism 0.238 0.038*** 0.23 0.038*** 0.25 0.038***     Perceived threat 2.312 0.057*** 2.316 0.057*** 2.312 0.057*** Media attention × Euroscepticism – – 0.192 0.041*** Variance within 1.433 0.033*** 1.431 0.033*** 1.433 0.033*** Variance between 2.766 0.068*** 2.929 0.135*** 2.765 0.068*** Variance slope PVV/Wilders 1.066 0.135*** 1.008 0.256*** 1.034 0.133*** Variance slope EU 0.064 0.476 0.039 0.504 0.063 0.474 Variance slope PVV/Wilders × EU – 0.149 0.522 – Cov (slope PVV/Wilders, intercept) 0.36 0.57 0.35 Cov (slope EU, intercept) −0.11 −0.01 −0.11 Cov (slope PVV/Wilders × EU, intercept) – −0.39 – Log likelihood −156,790.602 −156,740.085 −156,779.57 AIC 313,649.203 313,554.171 313,629.141 BIC 313,925.658 313,855.018 313,913.726 Variable Model 2a Model 2b Model 2c b SE b SE b SE Within person     PVV/Wilders media 0.866 0.035*** 1.112 0.105*** 0.201 0.14     EU media 0.171 0.053** 0.348 0.100** 0.169 0.053**     Unemployed 0.013 0.097 0.019 0.097 0.012 0.097     Media exposure −0.023 0.038 −0.029 0.038 −0.022 0.038     Euroscepticism 0.036 0.014* 0.039 0.014* 0.034 0.014*     Perceived threat 1.392 0.166*** 1.38 0.166*** 1.398 0.166***     PVV/Wilders media × EU media – −0.427 0.180* – Between person     Intercept 2.272 0.154*** 2.385 0.159*** 2.232 0.154***     Male 0.364 0.050*** 0.364 0.050*** 0.364 0.050***     Year of birth 0.018 0.002*** 0.018 0.002*** 0.018 0.002***     Low education – – –     Med education −0.115 0.062 −0.119 0.062 −0.115 0.062     High education −0.243 0.064*** −0.249 0.064*** −0.242 0.064***     Unemployed −0.037 0.337 −0.053 0.338 −0.036 0.338     Media exposure −0.179 0.076* −0.169 0.076* −0.179 0.077*     Euroscepticism 0.238 0.038*** 0.23 0.038*** 0.25 0.038***     Perceived threat 2.312 0.057*** 2.316 0.057*** 2.312 0.057*** Media attention × Euroscepticism – – 0.192 0.041*** Variance within 1.433 0.033*** 1.431 0.033*** 1.433 0.033*** Variance between 2.766 0.068*** 2.929 0.135*** 2.765 0.068*** Variance slope PVV/Wilders 1.066 0.135*** 1.008 0.256*** 1.034 0.133*** Variance slope EU 0.064 0.476 0.039 0.504 0.063 0.474 Variance slope PVV/Wilders × EU – 0.149 0.522 – Cov (slope PVV/Wilders, intercept) 0.36 0.57 0.35 Cov (slope EU, intercept) −0.11 −0.01 −0.11 Cov (slope PVV/Wilders × EU, intercept) – −0.39 – Log likelihood −156,790.602 −156,740.085 −156,779.57 AIC 313,649.203 313,554.171 313,629.141 BIC 313,925.658 313,855.018 313,913.726 Note: *p < .05; **p < .01; ***p < .001; PVV/Wilders media = media attention to the PVV/Wilders; EU media = media attention to the European Union. Figure 4 View largeDownload slide Marginal effect of media attention to the European Union on RRP party sympathy by media attention to PVV/Wilders (Model 2b). Note: Dashed lines are 95% confidence intervals Figure 4 View largeDownload slide Marginal effect of media attention to the European Union on RRP party sympathy by media attention to PVV/Wilders (Model 2b). Note: Dashed lines are 95% confidence intervals Figure 5 View largeDownload slide Marginal effect of Euroscepticism on RRP party sympathy by media attention to PVV/Wilders (Model 2c). Note: Dashed lines are 95% confidence intervals Figure 5 View largeDownload slide Marginal effect of Euroscepticism on RRP party sympathy by media attention to PVV/Wilders (Model 2c). Note: Dashed lines are 95% confidence intervals To account for potential confounding factors, we included a set of control variables in all models presented above. The results show that the effects are generally in line with previous research. We find that men have higher levels of RRP party sympathies, as well as younger respondents. Moreover, we can show that education decreases RRP party sympathies. We do not find an effect of unemployment.9 Nonetheless, accounting for these factors does not alter our findings.10 An alternative conceptualization of media attention to RRP parties and issues that are (perceived as) relevant to those parties is the co-occurrence of the RRP party and the European Union or immigration in the same article. We conducted additional analyses for the congruent media attention to RRP parties and to the European Union or to immigration, respectively (Supplementary Appendix Tables B and C). The findings support the results presented here. The main effects and interactions are positive and statistically significant. The effect of congruent media attention to RRP parties and immigration, as well as the interaction with perceived ethnic threat, is considerably larger than the effect of congruent media attention to RRP parties and the European Union, and the interaction with Euroscepticism. The correlation between the explanatory variables and the dependent variable might result from some unobserved, time trending factor (Wooldridge, 2009: 363). Thus, we reran our analysis controlling for survey year measured with a continuous variable ranging from 0 (2008) to 5 (2013). Furthermore, the relationship might be linked to the perceived party size or actual immigration inflow. We therefore reran our analyses including polling results11 and migration rates of nonwestern immigrants. The immigration rates between 2008 and 2013 follow a strict linear trend and highly correlate with survey years. In addition, we reran all full models including a lagged dependent variable, to control for the autoregressive effect of RRP party sympathies. The effect of media attention to the European Union turns out to be not significant after including party sizes, survey year, and a lagged dependent variable. The media attention to the European Union has only limited variance over time and might be linked to an unobserved time trending factor. Nonetheless, the results are generally in line with those presented in the paper. All robustness checks are presented in Supplementary Appendix Table D. To further focus on the within-person variation, we also estimate a fixed-effects panel regression model. The results are shown in Supplementary Appendix Table E. We find no substantial differences to our findings based on the multilevel structural equation model presented in Table 1 and Table 2. To test if the effect of media attention depends on whether respondents have read the newspaper, we included an interaction with media exposure (not shown). However, we find no significant interaction between exposure and media attention, which further supports the notion that issues frequently mentioned in the news are also salient in the public debate (Yang and Stone, 2003). Discussion The goal of this research was to examine individual changes of RRP party sympathies and how they relate to media attention. In particular, we first analyzed to what extent media attention relates to changes in RRP party sympathies. Second, we assessed if, and if so, to what extent individual attitudes interact with the effects of media attention on RRP party sympathies. To investigate our theoretical assumptions, we used a multilevel structural equation approach drawing on six annual waves of a representative Dutch panel survey from 2008 to 2013. To gauge the effects of media attention, we combined the panel data with results from a computer-assisted content analysis of media attention to Wilders/PVV, immigration, and the European Union. Our results show that media attention is positively related to individual sympathies for RRP parties. We find that while the explanatory contribution of media attention to immigration or to the European Union is only marginal, the proportion of explained variance by media attention to Wilders/PVV is rather substantial. Further analyses reveal that perceived ethnic threat and Euroscepticism moderate the effect of media attention on individual RRP party sympathies. With these findings, we show that visibility of RRP parties does not only mobilize the potential RRP electorate but interacts with attitudinal predictors of RRP party sympathy. In particular, among people who perceive stronger ethnic threats, more media attention to the RRP party is associated with more sympathy for this party. Similarly, Euroscepticism strengthens the effect of media attention to the RRP party on sympathy for the PVV. It is critical to note that ethnic threat explains significantly more variance of the effect of media attention to PVV/Wilders on RRP party sympathy compared to Euroscepticism. These findings extend previous research in several significant ways. First, our findings expand on earlier research on the effect of media attention on changes in RRP party sympathies. These findings are particularly important, as antecedents for within-individual changes of RRP party sympathies is largely underresearched. Second, the interaction of perceived ethnic threat and Euroscepticism with media attention is crucial to shed light on the puzzling longitudinal relationship of rather stable attitudes and volatile RRP party sympathies. Third, this study contributes to an evolving line of research, which shows that populism is increasingly part of newspaper reports and the public debate (Rooduijn, 2014). In sum, we extend our theoretical knowledge of the underlying mechanism at work, especially since group threat theory postulates a dynamic relationship. The interpretations of our findings come with certain limitations. We only draw inference on the effects of media attention in the sense of general salience. We refrain from analyses of the news reports’ valence. On the one hand, we believe that it is first and most importantly the visibility of topics that affects RRP sympathies, rather than the tone of news content. On the other hand, we have no information on the news a respondent consumed. Furthermore, one can question the generalizability of our findings beyond the Dutch context. Nevertheless, previous research shows similar effects of media attention on RRP party sympathies across different RRP parties and contexts (Vliegenthart et al., 2012). That said, we encourage future research to investigate to what extent the longitudinal mechanism presented here is potentially moderated by cross-country differences. Relatedly, we believe that future research might identify potentially moderating conditions by integrating the findings presented here into a cross-party comparison. To conclude, the present study provides insights on the longitudinal explanation underlying sympathies for RRP parties. Our study shows that media attention affects individual changes of RRP party sympathies. We substantiate this effect with evidence that the relationship between media attention and RRP party sympathies positively depends on individual perceived ethnic threat and Euroscepticism. Supplementary Data Supplementary Data are available at IJPOR online. Acknowledgment The LISS panel data were collected by CentERdata (Tilburg University, The Netherlands) through its MESS project funded by The Netherlands Organization for Scientific Research. Footnotes 1In this study, we use the term radical RRP parties. This term goes back to the seminal work of Betz (1993) but is also repeatedly used in more recent work especially for the party we study (Akkerman et al. 2016; Berning and Schlueter, 2016). We note that the classification and labeling of this party family were and still are subjects to rigorous scientific debate (Eger and Valdez, 2015; Mudde, 1996). 2In contrast to the relation mentioned above, one could at least argue on the reverse causal direction, that is, an increase in sympathies for RRP parties will lead to an increase in media attention. In the present study, we refrain from causal interpretation; yet, using a conceptualization consistent with related research (Schlueter and Davidov, 2013; Vliegenthart et al., 2012). Moreover, Van der Pas et al. (2011) have shown that increased support for the PVV was related to less visibility of the party in the media a week later. 3The LISS panel was conducted via Internet. Recent research shows that in the case of LISS, the interview mode had no effect on data quality (Revilla and Saris, 2013). All respondents without access to the Internet were provided with a PC and adequate devices. Further information in the LISS panel can be found at www.lissdata.nl. 4Of all respondents, 50.3% participated in all six waves, 16.5% in five waves, 16.2% only in four waves, 8.3% in three waves, and 8.7% participated only in two waves. We reran our analyses limited to individuals who were interviewed in all consecutive survey rounds and found that the results were substantively identical. 5We replicated all analyses with only single items for RRP sympathies, that is, sympathy for the leader and sympathy for the party, and found no substantial differences in the results, because of high correlation between both items (r = .92). 6Specifically, to measure media attention to PVV/Wilders, we searched for “DATE(<[Date of interview]) AND DATE(>=[Date of interview - 8 weeks]) AND (PVV OR Wilders).” To measure articles related to immigration, we searched for “DATE(<[Date of interview]) AND DATE(>=[Date of interview - 8 weeks]) AND (discrim OR (haat w/5 aanzet) OR inburgering* OR (scholing or (cursus* OR les* OR onderwijs) w/10 (immi* OR alloch* OR asiel* OR buitenl*)) OR taalcur* OR taalles* OR taalonderw* OR gezinsherenig* OR schijnhuw* OR nephuw* OR uithuw* OR immig* OR alloch* OR multicult* OR (verpaup* AND (buurt* or wijk*)) OR moslim* OR islam* OR asiel* OR uitgeproc* OR verblijfs* OR (grondwet w/10 artikel 7) OR importbruid OR (bruid* AND buitenland) OR (inkomenseis w/20 trouw*) OR pluriform* OR asielzoeker* OR vluchteling* OR (generaal pardon) OR pardonregeling)” and to measure media attention to the European Union, we searched for: “DATE(<[Date of interview]) AND DATE(>=[Date of interview - 8 weeks]) AND (europese pre/1 unie) OR EU.” 7The different media attention measures correlate with each other. PVV/Wilders media with Immigration mediar = .42; PVV/Wilders media with EU mediar =.32; EU media with Immigration mediar = .56. 8In subsequent analyses, we used a subset of interview dates and matched it to media attention scores based on articles published 6 weeks and 10 weeks before the interview. The shorter and longer intervals did not alter our results. 9We also tested an interaction of media attention to Wilders/PVV and perceived ethnic threat at the within-person level. Controlling for the within-level interaction led to estimation problems because of only small variance of perceived ethnic threat at the lower level. To reduce model complexity given the rather small number of time points, we decided to only allow the slopes of media attention to vary across individuals. Subsequential analyses showed that setting slopes of other within measures at random did not alter our results. 10Previous research further introduced political dissatisfaction as a motivation to prefer RRP parties (Mudde, 2007; but Zhirkov, 2013). Aside the merits of this factor as an additional motivation to prefer an RRP party, the longitudinal argument of political satisfaction to predict sympathies for a RRP party who joined the government is theoretically inconclusive. Nevertheless, to test the robustness of our findings, we reran our analyses including the potential confounding factor. The results were substantially concordant. 11We use the polling results from “de politieke barometer” conducted by Ipsos Synovate. References Aaldering L. , Vliegenthart R. ( 2016 ). Political leaders and the media: Can we measure political leadership images in newspapers using computer-assisted content analysis? Quality and Quantity , 50 , 1871 – 1905 .. doi:10.1007/s11135-015-0242-9 Google Scholar CrossRef Search ADS PubMed Akkerman T. , de Lange S. L. , Rooduijn M. ( 2016 ). Radical right-wing populist parties in Western Europe. Into the mainstream ? Oxford : Routledge . Arzheimer K. ( 2008 ). Protest, neo-liberalism or anti-immigrant sentiment: What motivates the voters of the extreme right in Western Europe? Comparative Governance and Politics , 2 , 173 – 197 .. doi:10.1007/s12286-008-0011-4 Arzheimer K. ( 2012 ). Electoral sociology: Who votes for the extreme right and why - and when? In Backes U. , Morrow D. (Eds.), The extreme right in Europe. Current trends and perspectives (pp. 35 – 50 ). Göttingen : Vendenhoeck & Ruprecht . Beck P. A. , Dalton R. J. , Greene S. , Huckfeldt R. ( 2002 ). The social calculus of voting: Interpersonal, media, and organizational influence on presidential choice . American Political Science Review , 96 ( 1 ), 37 – 73 . Google Scholar CrossRef Search ADS Bentler P. M. ( 1990 ). Comparative fit indexes in structural models . Psychological Bulletin , 107 , 238 – 246 . doi:10.1037/0033-2909.107.2.238 Google Scholar CrossRef Search ADS PubMed Berning C. C. ( 2016 ). Contextual perceived group threat and radical right-wing populist party preferences: Evidence from Switzerland . Research and Politics , 3 ( 1 ). Retrieved from https://doi.org/10.1177/2053168016635670 Berning C. C. , Schlueter E. ( 2016 ). The dynamics of radical right-wing populist party preferences and perceived group threat: A comparative panel analysis of three competing hypotheses in The Netherlands and Germany . Social Science Research , 55 , 83 – 93 . doi: 10.1016/j.ssresearch.2015.09.003 Google Scholar CrossRef Search ADS PubMed Betz H. G. ( 1993 ). The New Politics of Resentment. Radical Right-Wing Populist Parties in Western Europe . Comparative Politics , 25 , 413 – 427 . Google Scholar CrossRef Search ADS Betz H. G. ( 1994 ). Radical right-wing populism in Western Europe . London : Macmillan Basingstoke . Google Scholar CrossRef Search ADS Blalock H. M. ( 1967 ). Toward a theory of minority-group relations . New York, NY : Wiley . Boomgaarden H. G. , Vliegenthart R. ( 2007 ). Explaining the rise of anti-immigrant parties: The role of news media content . Electoral Studies , 26 , 404 – 417 . doi:10.1016/j.electstud.2006.10.018 Google Scholar CrossRef Search ADS Boomsma A. ( 2000 ). Reporting analyses of covariance structures . Structural Equation Modeling , 7 , 461 – 483 . doi:10.1207/S15328007SEM0703_6 Google Scholar CrossRef Search ADS Coffé H. , Voorpostel M. ( 2010 ). Young people, parents and radical right voting. The case of the Swiss people’s party . Electoral Studies , 29 , 435 – 443 . doi:10.1016/j.electstud.2010.03.015 Google Scholar CrossRef Search ADS Cohen B. C. ( 1963 ). The press and foreign policy . Princeton, NJ : Princeton University Press . Curran P. J. , Bauer D. J. ( 2011 ). The disaggregation of within-person and between-person effects in longitudinal models of change . Annual Review of Psychology , 62 ( 1 ), 583 – 619 . doi:10.1146/annurev.psych.093008.100356 Google Scholar CrossRef Search ADS PubMed Cutts D. C. , Ford R. , Goodwin M. J. ( 2011 ). Anti-immigrant, politically disaffected or still racist after all? Examining the attitudinal drivers of extreme right support in Britain in the 2009 European elections . European Journal of Political Research , 50 , 418 – 440 . doi: 10.1111/j.1475-6765.2010.01936.x Google Scholar CrossRef Search ADS Dalton R. J. ( 2014 ). Citizen politics: Public opinion and political parties in advanced industrial democracies ( 6th ed. ). Thousand Oaks, CA : CQ Press . Eger M. A. , Valdez S. ( 2015 ). Neo-nationalism in Western Europe . European Sociological Review , 31 ( 1 ), 115 – 130 . doi: 10.1093/esr/jcu087 Google Scholar CrossRef Search ADS Elchardus M. , Spruyt B. ( 2010 ). Does Higher Education Influence the Attitudes with Regard to the Extreme Right? European Journal of Social Sciences , 18 , 181 – 195 . Elchardus M. , Spruyt B. ( 2012 ). The contemporary contradictions of egalitarianism: An empirical analysis of the relationship between the old and new left/right alignments . European Political Sciences Review , 4 , 217 – 239 . doi:10.1017/S1755773911000178 Google Scholar CrossRef Search ADS Enders C. K. , Tofighi D. ( 2007 ). Centering predictor variables in cross-sectional multilevel models: a new look at an old issue . Psychological Methods , 12 , 121 – 138 . doi:10.1037/1082-989X.12.2.121 Google Scholar CrossRef Search ADS PubMed Gerber A. S. , Karlan D. , Bergan D. ( 2009 ). Does the media matter? A field experiment measuring the effect of newspapers on voting behavior and political opinions . American Economic Journal: Applied Economics , 1 , 35 – 52 . Google Scholar CrossRef Search ADS Givens T. E. ( 2004 ). The radical right gender gap . Comparative Political Studies , 37 ( 1 ), 30 – 54 . doi:10.1177/0010414003260124 Google Scholar CrossRef Search ADS Hopmann D. N. , Vliegenthart R. , De Vreese C. , Albæk E. ( 2010 ). Effects of election news coverage: How visibility and tone influence party choice . Political Communication , 27 , 389 – 405 . doi: 10.1080/10584609.2010.516798 Google Scholar CrossRef Search ADS Hox J. ( 2002 ). Multilevel analysis: Techniques and applications . Mahwah, NJ : Lawrence Erlbaum Associates . Hu L. , Bentler P. M. ( 1999 ). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives . Structural Equation Modeling , 6 ( 1 ), 1 – 55 . doi:10.1080/10705519909540118 Google Scholar CrossRef Search ADS Immerzeel T. , Lubbers M. , Coffé H. ( 2015 ). Competing with the radical right: Distances between the European radical right and other parties on typical radical right issues . Party Politics , 22 , 823 – 834 . Published online before print January 15, 2015. doi:10.1177/1354068814567975 Google Scholar CrossRef Search ADS Ivarsflaten E. ( 2008 ). What unites right-wing populists in Western Europe? Re-examining grievance mobilization models in seven successful cases . Comparative Political Studies , 41 ( 1 ), 3 – 23 . doi:10.1177/0010414006294168 Google Scholar CrossRef Search ADS Iyengar S. , Kinder D. R. ( 1987 ). News that matters: Television and American opinion . Chicago : University of Chicago Press . Kleinnijenhuis J. , van Hoof A. M. , Oegema D. , de Ridder J. ( 2007 ). A test of rivaling approaches to explain news effects: News on issue positions of parties, real-world developments, support and criticism, and success and failure . Journal of Communication , 57 , 366 – 384 . doi: 10.1111/j.1460-2466.2007.00347.x Google Scholar CrossRef Search ADS Koopmans R. , Muis J. ( 2009 ). The rise of right-wing populist Pim Fortuyn in The Netherlands: A discursive opportunity approach . European Journal of Political Research , 48 , 642 – 664 . doi: 10.1111/j.1475-6765.2009.00846.x Google Scholar CrossRef Search ADS Lancee B. , Pardos-Prado S. ( 2013 ). Group conflict theory in a longitudinal perspective: Analyzing the dynamic side of ethnic competition . International Migration Review , 47 ( 1 ), 106 – 131 . doi: 10.1111/imre.12015 Google Scholar CrossRef Search ADS Lubbers M. , Gijsberts M. , Scheepers P. ( 2002 ). Extreme right-wing voting in Western Europe . European Journal of Political Research , 41 , 345 – 378 . doi: 10.1111/1475-6765.00015 Google Scholar CrossRef Search ADS Lubbers M. , Scheepers P. ( 2001 ). Explaining the trend in extreme right-wing voting. Germany 1989-1998 . European Sociological Review , 17 , 431 – 449 . doi:10.1093/esr/17.4.431 Google Scholar CrossRef Search ADS Lucassen G. , Lubbers M. ( 2012 ). Who fears what? Explaining far-right-wing preference in Europe by distinguishing perceived cultural and economic ethnic threats . Comparative Political Studies , 45 , 547 – 574 . doi: 10.1177/0010414011427851 Google Scholar CrossRef Search ADS Marsh H. W. , Hocevar D. ( 1985 ). Application of confirmatory factor analysis to the study of self-concept: First- and higher order factor models and their invariance across groups . Psychological Bulletin , 97 , 562 – 582 . doi: 10.1037/0033-2909.97.3.562 Google Scholar CrossRef Search ADS Marsh H. W. , Lüdtke O. , Robitzsch A. , Trautwein U. , Asparouhov T. , Muthén B., , Nagengast B. ( 2009 ). Doubly-latent models of school contextual effects: Integrating multilevel and structural equation approaches to control measurement and sampling error . Multivariate Behavioral Research , 44 , 764 – 802 . doi:10.1080/00273170903333665 Google Scholar CrossRef Search ADS PubMed McCombs M. E. , Shaw D. L. ( 1972 ). The agenda-setting function of mass media . Public Opinion Quarterly , 36 , 176 – 187 . doi: 10.1086/267990 Google Scholar CrossRef Search ADS Mols F. , Jetten J. ( 2016 ). Explaining the appeal of populist right-wing parties in times of economic prosperity . Political Psychology , 37 , 275 – 292 . doi: 10.1111/pops.12258 Google Scholar CrossRef Search ADS Mudde C. ( 1996 ). The war of words defining the extreme right party family . West European Politics , 19 , 225 – 248 . Google Scholar CrossRef Search ADS Mudde C. ( 2007 ). Populist radical right parties in Europe . Cambridge : Cambridge University Press . Google Scholar CrossRef Search ADS Mughan A. , Paxton P. ( 2006 ). Anti-immigrant sentiment, policy preferences and populist party voting in Australia . British Journal of Political Science , 36 , 341 – 358 . doi:10.1017/S0007123406000184 Google Scholar CrossRef Search ADS Muthén L. K. , Muthén B. O. ( 1998 -2012). Mplus user’s guide ( 7th ed. ). Los Angeles, CA : Muthén & Muthén . Norris P. ( 2005 ). Radical right. Voters and parties in the electoral market . Cambridge; New York, NY : Cambridge University Press . Google Scholar CrossRef Search ADS Oegema D. , Kleinnijenhuis J. ( 2000 ). Personalization in political television news: A 13-wave survey study to assess effects of text and footage . Communications , 25 ( 1 ), 43 – 60 . doi: 10.1515/comm.2000.25.1.43 Google Scholar CrossRef Search ADS Pappi F. U. ( 1996 ). Political behavior: Reasoning voters and multi-party systems. In Goodin R. E. , Klingemann H. D. (Eds.), A new handbook of political science (pp. 255 – 275 ). Oxford : Oxford University Press . Poznyak D. , Abts K. , Swyngedouw M. ( 2011 ). The dynamics of the extreme right support: A growth curve model of the populist vote in Flanders-Belgium in 1987-2007 . Electoral Studies , 30 , 672 – 688 . doi: 10.1016/j.electstud.2011.06.011 Google Scholar CrossRef Search ADS Raftery A. ( 1995 ). Bayesian model selection in social research . Sociological Methodology , 25 , 111 – 163 . doi:10.2307/271063 Google Scholar CrossRef Search ADS Revilla M. A. , Saris W. E. ( 2013 ). A comparison of the quality of questions in a face-to-face and a web survey . International Journal of Public Opinion Research , 25 , 242 – 253 . doi:10.1093/ijpor/eds007 Google Scholar CrossRef Search ADS 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 , 411 – 424 . doi:10.1093/esr/jcn028 Google Scholar CrossRef Search ADS Rooduijn M. ( 2014 ). The mesmerising message: The diffusion of populism in public debates in Western European media . Political Studies , 62 , 726 – 744 . doi: 10.1111/1467-9248.12074 Google Scholar CrossRef Search ADS Rydgren J. ( 2005 ). Is extreme right-wing populism contagious? Explaining the emergence of a new party family . European Journal of Political Research , 44 , 413 – 437 . doi:10.1111/j.1475-6765.2005.00233.x Google Scholar CrossRef Search ADS Rydgren J. ( 2007 ). The sociology of the radical right . Annual Review of Sociology , 33 ( 1 ), 241 – 262 . doi: 10.1146/annurev.soc.33.040406.131752 Google Scholar CrossRef Search ADS Rydgren J. ( 2008 ). Immigration sceptics, xenophobes or racists? Radical right-wing voting in six West European countries . European Journal of Political Research , 47 , 737 – 765 . doi:10.1111/j.1475-6765.2008.00784.x Google Scholar CrossRef Search ADS Satorra A. , Bentler P. M. ( 2001 ). A scaled difference chi-square test statistic for moment structure analysis . Psychometrika , 66 , 507 – 514 . doi: 10.1007/BF02296192 Google Scholar CrossRef Search ADS Schafraad P. , Wester F. , Scheepers P. ( 2013 ). Media attention to the far-right in three Dutch newspapers 1986-2004. Characteristics and trends . Observatorio , 7 , 1 – 31 . Scheepers P. , Gijsberts M. , Coenders M. ( 2002 ). Ethnic exclusionism in European countries: public opposition to civil rights for legal migrants as a response to perceived ethnic threat . European Sociological Review , 18 ( 1 ), 17 – 34 . doi:10.1093/esr/18.1.17 Google Scholar CrossRef Search ADS Scherpenzeel A. C. , Das M. ( 2010 ). True longitudinal and probability-based Internet panels: Evidence from the Netherlands. In: M. Das, P. Ester, & L. Kaczmirek (Eds.), Social and Behavioral Research and the Internet: Advances in Applied Methods and Research Strategies. New York: Taylor & Francis, Routledge. Schlueter E. , Davidov E. ( 2013 ). Contextual sources of perceived group threat: Negative immigration-related news reports, immigrant group size and their interaction, Spain 1996–2007 . European Sociological Review , 29 , 179 – 191 . doi:10.1093/esr/jcr054 Google Scholar CrossRef Search ADS Schlueter E. , Meuleman B. , Davidov E. ( 2013 ). Immigrant integration policies and perceived group threat: A multilevel study of 27 Western and Eastern European countries . Social Science Research , 42 , 670 – 682 . doi:10.1016/j.ssresearch.2012.12.001 Google Scholar CrossRef Search ADS PubMed 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 . doi:10.1016/j.electstud.2012.11.003 Google Scholar CrossRef Search ADS Semetko H. , Schoenbach K. ( 1994 ). Germany’s “unity election”—Voters and the media . Cresskill, NJ : Hampton Press . Semyonov M. , Raijman R. , Gorodzeisky A. ( 2006 ). The rise of anti-foreigner sentiment in European societies, 1988-2000 . American Sociological Review , 71 , 426 – 449 . doi:10.1177/000312240607100304 Google Scholar CrossRef Search ADS Sheets P. , Bos L. , Boomgaarden H. G. ( 2016 ). Media cues and citizen support for right-wing populist parties . International Journal of Public Opinion Research , 28 , 307 – 330 . doi:10.1093/ijpor/edv014 Google Scholar CrossRef Search ADS Sherif M. , Harvey O. J. , White J. B. , Hood W. R. , Sherif C. W. ( 1961 ). Intergroup conflict and cooperation. The robbers cave experiment [1954] . Norman : University of Oklahoma . Song X. Y. , Lee S. Y. , Hser Y. I. ( 2008 ). A two-level structural equation model approach for analyzing multivariate longitudinal responses . Statistics in Medicine , 27 , 3017 – 3041 . doi: 10.1002/sim.3266 Google Scholar CrossRef Search ADS PubMed Swyngedouw M. ( 2001 ). The subjective cognitive and affective map of extreme right voters: using open-ended questions in exit polls . Electoral Studies , 20 , 217 – 241 . doi:10.1016/S0261-3794(00)00010-X Google Scholar CrossRef Search ADS Tesler M. ( 2015 ). Priming predispositions and changing policy positions: An account of when mass opinion is primed or changed . American Journal of Political Science , 59 , 806 – 824 . doi: 10.1111/ajps.12157 Google Scholar CrossRef Search ADS Van der Brug W. , Fennema M. ( 2003 ). Protest or mainstream? How the European anti-immigrant parties developed into two separate groups by 1999 . European Journal of Political Research , 42 ( 1 ), 55 – 76 . doi: 10.1111/1475-6765.00074 Google Scholar CrossRef Search ADS Van der Brug W. , Fennema M. ( 2007 ). Causes of voting for the radical right . International Journal of Public Opinion Research , 19 , 474 – 487 . doi:10.1093/ijpor/edm031 Google Scholar CrossRef Search ADS 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 .. doi:10.1111/1475-6765.00505 Van der Brug W. , Semetko H. A. , Valkenburg P. M. ( 2007 ). Media priming in a multi-party context: A controlled naturalistic study in political communication . Political Behavior , 29 , 115 – 141 . doi:10.1007/s11109-006-9020-7 Google Scholar CrossRef Search ADS Van der Meer T. W. G. , van Elsas E. , Lubbe R. , van der Brug W. ( 2013 ). Are volatile voters erratic, whimsical or seriously picky? A panel study of 58 waves into the nature of electoral volatility (The Netherlands 2006–2010) . Party Politics , 21 ( 1 ), 100 – 114 . doi:10.1177/1354068812472570 Google Scholar CrossRef Search ADS Van der Pas D. , de Vries C. , van der Brug W. ( 2011 ). A leader without a party: Exploring the relationship between Geert Wilders’ leadership performance in the media and his electoral success . Party Politics , 19 , 458 – 476 . doi: 10.1177/1354068811407579 Vliegenthart R. , Boomgaarden H. G. , Van Spanje J. ( 2012 ). Anti-immigrant party support and media visibility: A cross-party, over-time perspective . Journal of Elections, Public Opinion and Parties , 22 , 315 – 358 . doi:10.1080/17457289.2012.693933 Google Scholar CrossRef Search ADS Vossen K. ( 2010 ). Populism in The Netherlands after Fortuyn: Rita Verdonk and Geert Wilders compared . Perspectives on European Politics and Society , 11 ( 1 ), 22 – 38 . doi:10.1080/15705850903553521 Google Scholar CrossRef Search ADS Vossen K. ( 2011 ). Classifying Wilders: The ideological development of Geert Wilders and his party for freedom . Politics , 31 , 179 – 189 . doi:10.1111/j.1467-9256.2011.01417.x Google Scholar CrossRef Search ADS Walgrave S. , de Swert K. ( 2004 ). The making of the (issues of the) Vlaams Blok . Political Communication , 21 , 479 – 500 . doi:10.1080/10584600490522743 Google Scholar CrossRef Search ADS Werts H. , Scheepers P. , Lubbers M. ( 2012 ). Euro-scepticism and radical right-wing voting in Europe, 2002–2008: Social cleavages, socio-political attitudes and contextual characteristics determining voting for the radical right . European Union Politics , 14 , 183 – 205 . doi: 10.1177/1465116512469287 Google Scholar CrossRef Search ADS Wooldridge J. M. ( 2009 ). Introductory econometrics: a modern approach ( 4th ed .). Independence : South-Western . Yang J. , Stone G. ( 2003 ). The powerful role of interpersonal communication in agenda setting . Mass Communication and Society , 6 ( 1 ), 57 – 74 . doi:10.1207/S15327825MCS0601_5 Google Scholar CrossRef Search ADS Zhang Z. , Zyphur M. J. , Preacher K. J. ( 2009 ). Testing multilevel mediation using hierarchical linear models: Problems and solutions . Organizational Research Methods , 12 , 695 – 719 . doi: 10.1177/1094428108327450 Google Scholar CrossRef Search ADS Zhirkov K. ( 2013 ). Nativist but not alienated: A comparative perspective on the radical right vote in Western Europe . Party Politics , 20 , 286 – 296 . doi: 10.1177/1354068813511379 Google Scholar CrossRef Search ADS © The Author(s) 2018. Published by Oxford University Press on behalf of The World Association for Public Opinion Research. All rights reserved. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png International Journal of Public Opinion Research Oxford University Press

Media Attention and Radical Right-Wing Populist Party Sympathy: Longitudinal Evidence From The Netherlands

Loading next page...
 
/lp/ou_press/media-attention-and-radical-right-wing-populist-party-sympathy-QViOzsTLPf
Publisher
Oxford University Press
Copyright
© The Author(s) 2018. Published by Oxford University Press on behalf of The World Association for Public Opinion Research. All rights reserved.
ISSN
0954-2892
eISSN
1471-6909
D.O.I.
10.1093/ijpor/edy001
Publisher site
See Article on Publisher Site

Abstract

Abstract This contribution provides evidence for the sources of sympathies for radical right-wing populist (RRP) parties in a longitudinal perspective. We extend previous knowledge by evaluating the impact of media attention on individual changes of RRP party sympathies. To test our hypotheses, we use panel data from The Netherlands and combine it with information on the saliency of RRP parties and their issues in major Dutch newspapers. Drawing on multilevel structural equation models, our findings indicate that media attention positively affects changes in RRP party sympathies. Furthermore, we find that the effect of media attention on RRP party sympathies is enhanced by perceived ethnic threat and Euroscepticism, respectively. In sum, this study shows that media attention to RRP parties is pivotal for the fortune of such parties. Radical right-wing populist (RRP) parties and their electorate received much academic attention in the past two decades. The core and distinguishing feature of these parties is their harsh position on immigration restrictions (Immerzeel, Lubbers, & Coffé, 2015; Ivarsflaten, 2008). These parties are radical, as they oppose pluralism and diversity of ideas, right-wing regarding their stance on (non-)egalitarianism, and finally populist in their distinction between the pure people and the corrupt elite.1 Scholars agree on voters’ concern about negative consequences because of immigration as the most important motivation to prefer an RRP party (Arzheimer, 2008; Cutts, Ford, & Goodwin, 2011; Lubbers, Gijsberts, & Scheepers, 2002; Norris, 2005; Rydgren, 2008; Van der Brug, Fennema, & Tillie, 2000). Previous research mostly focused on explaining who prefers RRP parties; yet, only little is known about why individual sympathies for RRP parties change over time. The scarce evidence for changes in RRP party sympathies either focused on aggregated growth (Boomgaarden & Vliegenthart, 2007; Koopmans & Muis, 2009; Lubbers & Scheepers, 2001; Poznyak, Abts, & Swyngedouw, 2011; Vliegenthart, Boomgaarden, & Van Spanje, 2012; Walgrave & de Swert, 2004) or solely on sociodemographic and structural influences, overlooking attitudinal characteristics (Rink, Phalet, & Swyngedouw, 2009). Our study aims to improve on this gap in the literature. Despite the merits of prior research, the relative neglect of analyzing within-individual RRP party sympathy’s change is unfortunate because the underlying longitudinal mechanism remains unknown. Following scholarly suggestions, the role of mass media attention, that is, news media content, is fruitful in this endeavor (Arzheimer, 2012; Rydgren, 2005; Van der Brug & Fennema, 2007). Drawing on agenda-setting theory, mass media coverage is expected to provide the necessary political opportunity for the rise of RRP parties, perhaps even from the political fringe to government participation. The idea is that RRP parties thrive on visibility and prominence and that they depend on media attention as a breeding ground for electoral mobilization. The growing literature on sympathies for RRP parties paid rather limited attention to salience of RRP relevant news content, and existing evidence is inconclusive (but see Boomgaarden & Vliegenthart, 2007; Koopmans & Muis, 2009; Van der Pas, de Vries, & van der Brug, 2011; Vliegenthart et al., 2012). Furthermore, empirical evidence for the effect of media attention on within-individual change of RRP party sympathy remains missing and so is evidence for the interaction of media attention with the most prominent determinants in RRP party research–perceived ethnic threat and also Euroscepticism. With this study, we assess whether perceived ethnic threat or Euroscepticism reinforces the effect of salience of RRP parties on RRP party sympathy and their issues, such as immigration or the European Union. We aim to contribute to the existing literature in two significant ways. First, we provide evidence for the role of mass media attention in explanations of RRP party sympathies in an individual longitudinal perspective. Second, we investigate if, and if so, to what extent perceived ethnic threat and Euroscepticism moderate the effect of media attention on RRP party sympathy. The site of our study is The Netherlands. For several reasons, The Netherlands is a national setting that is well suited to investigate the nexus of media attention and changes of individual RRP party sympathies: first, The Netherlands witnessed the rise of rather successful and rigorous RRP parties (Vossen, 2011). Most recently, the Partij voor de Vrijheid (PVV, Party for Freedom, founded in 2006) propped up a minority government from 2010 until 2012 and emerged to an important element of the Dutch political space. Its founder and leader Geert Wilders has a reputation for his hard-line anti-Islamic stance and his strong position toward immigration restriction policies, which is in line with core characteristics of the RRP party family (Vossen, 2010, 2011). Unlike others, the PVV has no extreme right origin, but Geert Wilders’ alliance with the Front National’s front runner Le Pen shows that the PVV is no different from other RRP parties (Akkerman, de Lange, & Rooduijn, 2016). Second, The Netherlands are characterized by an increasingly volatile electorate (van der Meer, van Elsas, Lubbe, & van der Brug, 2013), that is, voters are more likely to change their vote intention, which leverages the potential of longitudinal explanations for changes in RRP party sympathies, compared with national settings with a rather stable electorate, which are presumably less likely to be affected by media attention. Theoretical Framework The evolving literature on RRP parties has identified a rather broad set of sources. Previous research finds that not only sociodemographic variables such as gender (Givens, 2004) and education (Elchardus and Spruyt, 2010) affect the appeal to RRP parties but also constructs such as relative deprivation (Elchardus and Spruyt, 2012) or relative wealth (Mols and Jetten, 2016) are important. In current research on RRP party sympathies, group threat and group conflict theory are prominent theoretical frameworks (Rydgren, 2007). Within these frameworks, perceived ethnic threat refers to symbolic issues, such as losing a cultural identity, and to competition over scarce resources, for example, employment and material security (Blalock, 1967; Sherif, Harvey, White, Hood, & Sherif, 1961). Perceptions of these cultural and economic out-group threats are expected to increase the likelihood to vote for RRP parties, as these parties proclaim to serve the interests of the self-defined in-group against migrants and ethnically defined out-groups. Previous research provided abundant evidence of this association (Berning, 2016; Berning & Schlueter, 2016; Ivarsflaten, 2008; Lucassen & Lubbers, 2012). Media Attention Why should media attention contribute to sympathies for RRP parties? Building on agenda-setting theory, issues frequently mentioned in news reports are also salient in the public debate (Yang & Stone, 2003). The media pushes the public agenda, also on information relevant for RRP party sympathies. Thus, media attention provides a fertile ground for longitudinal explanations of changes in RRP party sympathies.2 In our case, we examine media attention to RRP parties themselves and to issues that are most relevant for RRP parties: immigration and the European Union. The salience of RRP parties is expected to affect RRP party sympathies through awareness. News reports about RRP parties, immigration, and the European Union are publicity for such parties, and in turn, the potential electorate will perceive RRP parties and their issues to be more relevant. All parties mobilize their potential support partly through visibility (Oegema & Kleinnijenhuis, 2000). Nonetheless, as RRP parties generally recruit their electorate on the political margin, they largely depend on voters becoming (more) aware of their existence, their stance on immigration and the European Union, or to perceive the RRP party as more popular. The electorate of RRP parties is likely to be less politically sophisticated, informed, and literate in party programs, given their relatively low educated profile; therefore, an increase of media attention to RRP parties may strengthen RRP party sympathies (see also Dalton, 2014: 230). An important proposition is that news reports set the public agenda (McCombs & Shaw, 1972). However, this does not imply that people adapt the connotations of these news reports. Just as Cohen (1963) used to tell: “the news media may not be successful much of the time in telling people what to think, but it is stunningly successful in telling its readers what to think about” (Cohen, 1963: 13). A counterargument may be that, beyond salience and general publicity, negative press would in fact be harmful for the rise of RRP parties. Previous research shows that the tone of news media reports influences vote choice (Beck, Dalton, Greene, & Huckfeldt, 2002). Especially undecided voters are more inclined to vote for a party if the tone of media reports is positive (Hopmann, Vliegenthart, De Vreese, & Albæk, 2010). Nonetheless, Kleinnijenhuis, van Hoof, Oegema, and de Ridder (2007) find that different news types, such as criticism or salience, complement rather than confound each other. While the tone of news reports is not subject of this study and not tested empirically, we build on previous research that shows that criticism of RRP parties in the media is not related to RRP party support (Koopmans and Muis, 2009). Furthermore, the general image of news reports on RRP parties is critical and negative (Schafraad, Wester, & Scheepers, 2013), and there is evidence that in such a controversial news environment media slant has no effect on political attitudes (Gerber, Karlan, & Bergan, 2009), as only a fraction of articles evaluated politicians (Aaldering & Vliegenthart, 2016). Hence, we focus on salience, assuming that the general public debates will pick up on topics that are frequently mentioned in the news via interpersonal communication (Yang & Stone, 2003). Extant research finds that the mere visibility of political actors in the media increases their electoral potential (Oegema & Kleinnijenhuis, 2000; Semetko & Schoenbach, 1994). Salience of the European Union or immigration issues and RRP parties primes voters to strongly relate these issues to such parties and perceive both to be important. The underlying mechanism of a priming effect assumes that individuals will use information that is easily accessible or activated recently (Van der Brug, Semetko, & Valkenburg, 2007). The salience of said issues will cause individuals to use them as evaluation criteria for their electoral preferences (Iyengar & Kinder, 1987). On the one hand, individuals will perceive salient issues to be relevant and important. On the other hand, RRP parties will appear to be more qualified to face these issues. This relates to recent experimental research that provides evidence for the individual-level relationship between media cues and RRP party support (Sheets, Bos, & Boomgaarden, 2016). Thus, the combination of salient RRP parties and RRP issues, such as the European Union or immigration, will increase the support for these parties. To our knowledge, only a few studies address the link between media attention and sympathies for RRP parties. Lubbers and Scheepers (2001) evidenced with repeated cross-sectional survey data a positive effect of media attention to RRP parties on individual preferences for RRP parties in Germany. Vliegenthart et al. (2012) found similar results for Belgium, The Netherlands, and also Germany drawing on aggregated measures for media attention and RRP party support. These findings are contradicting with the Van der Pas et al. (2011) study, which might be because of differences of the trajectory span. In contrast to changes over years, Van der Pas et al. (2011) focused on weekly changes within 1 year, also drawing on aggregated data. In sum, previous research is limited to evidence of aggregated changes and somewhat inconclusive. The Interaction of Perceived Ethnic Threat and Euroscepticism With Media Attention to RRP Parties Sympathies for RRP parties arise from some people’s perception that immigrants are threatening the interests of their in-group. They are expected to use their vote choice in favor of their policy preference for restrictive immigration legislation (Mughan & Paxton, 2006). Along this line of explanation, Eurosceptic voters are also expected to support RRP parties in rejection of European integration (Werts, Scheepers, & Lubbers, 2012). This motivation, that is, an attitudinal proximity of voters to their party, is by no means different to other parties (Van der Brug et al., 2000). Explanations for RRP party sympathy as nonideological protest find only limited support in the literature (Swyngedouw, 2001; van der Brug & Fennema, 2003, but see Schumacher & Rooduijn, 2013). To strengthen the postulated assumption of a genuine effect of media attention on sympathies for RRP parties, we turn to the interaction of media attention with perceived ethnic threat or Euroscepticism and their relation to sympathies for RRP parties. In fact, testing this interaction is important not only analytically, as such as the nature or strength of the relationship depends on perceived ethnic threat or Euroscepticism, respectively, but also for a more integrated understanding of how these concepts are linked. We assume that media attention does not only affect sympathies for RRP parties directly but also reinforces the effect of perceived ethnic threat and Euroscepticism on RRP party sympathies (Lubbers & Scheepers, 2001: 434). An increase of RRP parties and their message in the news likely heightens the voters’ perceptions that RRP parties’ issues are more relevant and that their perception of ethnic threat and skepticism about the European integration are sensible. This in turn may contribute to an urge to alleviate these concerns by sympathizing with RRP parties. Put differently, the RRP electorate, who perceives immigrants and the European Union as threatening, is likely to feel that they are right about their perception when RRP parties are frequently mentioned in the media. Furthermore, an increase in salience of RRP parties may trigger the link between RRP party sympathies and perceived ethnic threat, as well as Euroscepticism. An increase in awareness of those parties may simply contribute to the feeling that RRP parties are a good representation of people who perceive immigrants as threatening and are against the European Union. Following the priming argument, salience of RRP parties and their issues will especially resonate with individuals who hold Eurosceptic attitudes and perceive immigrants as threatening, as individuals will interpret new information based on their preexisting attitudes (Tesler, 2015). Hypotheses Our theoretical assumptions described above are tested with the following hypotheses: Hypothesis 1: The higher the media attention to RRP parties, the more people sympathize with RRP parties. Hypothesis 2: The higher the media attention to immigration, the more people sympathize with RRP parties. Hypothesis 3: The positive effect of media attention to RRP parties on RRP party sympathies increases, when media attention to immigration is more salient. Hypothesis 4: The positive effect of media attention to RRP parties on RRP party sympathies increases, when perceived ethnic threat is larger. Hypothesis 5: The higher the media attention to the European Union, the more people sympathize with RRP parties. Hypothesis 6: The positive effect of media attention to RRP parties on RRP party sympathies increases, when media attention to the European Union is more salient. Hypothesis 7: The positive effect of media attention to RRP parties on RRP party sympathies increases, when Euroscepticism is larger. Data and Methods Data To examine our theoretical assumptions, we need two different data sources. We combine individual-level multiwave panel data with longitudinal information from computer-assisted content analyses of newspaper articles. Regarding the individual-level data, we make use of the Longitudinal Internet Studies for the Social Sciences (LISS) administered by CentERdata (Tilburg University, The Netherlands). The LISS panel is a representative sample of Dutch citizens, drawn from the population based on a true probability sample (Scherpenzeel & Das, 2010).3 The data are available for six waves from 2008 to 2013. We limit our analyses to respondents who participated at least in two waves of the study period.4 As we focus on attitudes of majority group members, we excluded respondents with a migration background. The data structure follows a long format with Nobs. = 25,110 observations nested in Nind. = 6,184 individuals. Measures RRP party sympathy. We use two items to measure RRP party sympathy.5 For the first item, respondents were asked to indicate their sympathy for the PVV: “What do you think of the Party for Freedom?” The second item gauges the evaluation of the party leader by asking: “What do you think of Geert Wilders?” Response options for both items were given on an 11-point scale and ranged from 0 (very unsympathetic) to 10 (very sympathetic). We then averaged both items to a single index of respondents’ RRP party sympathy, with higher values indicating higher RRP party sympathy. Reliability of this measure was high (Cronbach’s alpha = .96). This measure has several advantages compared with discrete choice conceptualizations. For instance, it allows respondents to sympathize with more than one party, which is especially relevant in The Netherlands with many parties competing for votes. It measures even small differences in party preferences, a momentary measure of party ratings, and it resembles a construct, which is not only relevant during elections but rather captures the underlying party utility function. Conceptually, it refers to party preference, which is the central intervening variable between reasoning about the party and actual voting behavior (Pappi, 1996). Media attention. To assess media attention, we conducted computer-assisted content analyses of five major Dutch newspapers. Specifically, we compiled the number of articles published in De Telegraaf, AD/Algemeen Dagblad, De Volkskrant, NRC Handelsblad, and Trouw for RRP party related news. Both De Telegraaf and AD/Algemeen Dagblad are considered as more right-wing leaning, tabloid kind of newspapers. De Telegraaf is the largest newspaper in the country, with a circulation of half a million-printed copies. Nonetheless, the online version reaches a much larger audience. The other three papers are described as quality papers, with the Volkskrant more left-wing leaning, the NRC Handelsblad more social–liberal, and the Trouw with more attention to religiosity. To gauge the number of articles related to the RRP party, we electronically searched for the terms “PVV” or “Wilders.” To measure salience of immigration related news, we build on previous research and used a search string suggested and tested by Vliegenthart et al. (2012: 335). Finally, we assessed media attention to the European Union, by searching all five newspapers for articles about the “European Union” or the “EU.”6 We analyzed all articles published up to 8 weeks before the interview for each wave and each person. This period is long enough to let news reports influence the public’s agenda and helps to reduce the risk of simultaneity bias (Schlueter and Davidov, 2013).7 Moreover, measuring media attention before the interviews is in line with our assumption regarding the temporal order, that is, media attention precedes RRP party sympathies.8 To probe the validity of our data, we manually reviewed a small number of randomly chosen newspaper articles from the total sample of the newspaper articles we collected. Consistent with our search criteria, the small subset of articles turned out to cover indeed issues pertaining to the PVV/Wilders, immigration, or the European Union, respectively. To aid in interpretation of the unstandardized regression coefficients, we transformed the frequency scores to range from a minimum of 0 to a maximum of 1. Thus, a one-unit change in these variables equals the difference between the lowest and the highest possible score. The individual-level panel data were collected in 6 consecutive years. However, as it is common in survey research, not all interviews were conducted on the same date. In total, the LISS panel consists of 311 different interviews dates (see Supplementary Appendix Figure A). We therefore compiled 933 search strings for all media attention measures, that is, we applied our three media attention measure (PVV/Wilders, immigration, and the European Union) to 311 different time intervals. The individual-level panel data were then merged with the results of our computer-assisted content analyses, that is, we assigned three frequency scores to each interview (date). For example, Respondent A answered the questionnaire on December 3, 2007 and Respondent B did so on March 4, 2008, with both respondents being interviewed during the first wave. Thus, for Respondent A, the results from the content analyses refer to the 8 weeks period before December 3, 2007, while for Respondent B the frequency scores refer to the 8 weeks period before March 4, 2008. Perceived ethnic threat and Euroscepticism. To operationalize perceived ethnic threat, we use three items. For the first two items, respondents were asked to evaluate the following statements on a five-point Likert type scale: “There are too many people of foreign origin or descent in The Netherlands” and “Some sectors of the economy can only continue to function because people of foreign origin or descent work there.” Answer options ranged from 1 (fully disagree) to 5 (fully agree). These items correspond to operationalizations of perceived ethnic threat in related research (Scheepers, Gijsberts, & Coenders, 2002; Schlueter, Meuleman, & Davidov, 2013; Semyonov, Raijman, & Gorodzeisky, 2006). Additionally, respondents were asked: “Where would you place yourself on a scale of 1 to 5, where 1 means that immigrants retain their own culture and 5 means that they should adapt entirely?” After inverting the second item, higher values of all measures reflect higher ethnic threat perception. To gauge the effect of Euroscepticism, we used respondents’ answers to the following question: “Where would you place yourself on a scale from 1 to 5, where 1 means that European unification should go further and 5 means that it has already gone too far?” Control variables. Previous research shows that men are more likely to sympathize with RRP parties (Givens, 2004). Hence, we control for gender using a dichotomous measure with 0 for female and 1 for male. Age is expected to have a negative influence (Coffé and Voorpostel, 2010). We capture this by including year of birth as a continuous measure. There is also evidence that low education increases RRP party sympathy (Betz, 1994). We measure education on a three-point scale with 1 for primary school, vmbo (intermediate secondary education), 2 for havo/vwo (higher secondary education/preparatory university education), mbo (intermediate vocational education), and 3 for hbo (higher vocational education), wo (university). Further, we control for the effect of economic developments with an individual measure of unemployment (Lancee and Pardos-Prado, 2013), measured by a dichotomous variable with 0 (not unemployed) and 1 (unemployed). To control for the potential confounding factor of media exposure, we include a dichotomous item, asking respondents: “Do you follow the news in a bought newspaper or one that you have a subscription to?” Answers were coded as 0 (no) and 1 (yes). Descriptive statics for all control variables are presented in Supplementary Table A. Method We use multilevel structural equation modeling (MLSEM) for longitudinal data to test our hypotheses (Song, Lee, & Hser, 2008). This approach offers several advantages. First, we can take the nested data structure into account, that is, observations nested within individuals, which yields to adjusted standard errors, and we are able to decompose the total variance of RRP party sympathy into within and between components. The within component refers to intraindividual variance and the between component is based on interindividual variance (Curran & Bauer, 2011). Second, our analytical approach does not compel balanced data, that is, individuals do not require having the same number of observations (Hox, 2002). Third, this technique allows us to measure perceived ethnic threat as a latent variable at both levels, adequately accounting for sampling and measurement error (Marsh et al., 2009). Moreover, simultaneously modeling relationships at two levels enables us to deconflate estimates to between- and within-individual effects (Zhang, Zyphur, & Preacher, 2009). All models are based on full information maximum likelihood estimates, available in Mplus 7 (Muthén & Muthén, 1998-2012). To assess the goodness of fit of our measurement model, we use multiple fit indices: χ2/df ratio, the comparative fit index (CFI), and the root mean square error of approximation (RMSEA), and the standardized root mean square residual (SRMR) (Bentler, 1990; Boomsma, 2000; Hu & Bentler, 1999; Marsh & Hocevar, 1985). We consider models with χ2/df < 5, RMSEA < .06, SRMR(within; between) <.06, and CFI > .95 to have a good fit to the data. Our structural models include random slopes, which imply that the variance of the dependent variable varies across observations, with no single covariance matrix for the overall model. Thus, for our structural models, the model fit indices described above are not available. We measure perceived ethnic threat with three items as one latent factor both at the within- and between-person level of analysis. To ensure that this factor of perceived ethnic threat is in fact the same construct measured on both levels, we need to establish metric cross-level invariance (Marsh et al., 2009). We use Satorra–Bentler scaled χ2 difference tests (Satorra & Bentler, 2001) for model comparison. In favor for interpretation of our structural models, we centered year of birth on its mean and following the suggestions of Enders and Tofighi (2007), we facilitate group mean centering for all other variables, except dichotomous measures. Results We begin presenting our results by describing the trends of media attention and RRP party sympathies in The Netherlands. We then briefly discuss the results of our measurement model for perceived ethnic threat, and continue with the results of our multilevel structural equation models. Media Attention and RRP Party Sympathies in The Netherlands Figure 1 depicts the average frequency of newspaper articles mentioning Geert Wilders or his PVV (dashed-dotted line), newspaper articles on immigration (dashed line), newspaper articles mentioning the European Union (dotted line), as well as the aggregated scores of RRP party sympathies (solid line) per year. The graph shows that sympathies for an RRP party largely follow the trend of media attention to Wilders and the PVV. Both lines start rather low and build up to their peak in 2011. All respondents were interviewed around the turn of the year, that is, the maxima at 2011 were about 2–3 months after the right-wing government formation of the VVD and the Christen-Democratisch Appèl (CDA, Christian Democratic Appel), propped up by the PVV. Figure 1 View largeDownload slide RRP party sympathies and media attention. Solid line depicts RRP party sympathy; dash-dotted line represents media attention to PVV/Wilders; dashed line represents media attention to immigration; dotted line indicates media attention to the European Union Figure 1 View largeDownload slide RRP party sympathies and media attention. Solid line depicts RRP party sympathy; dash-dotted line represents media attention to PVV/Wilders; dashed line represents media attention to immigration; dotted line indicates media attention to the European Union The formation of the government took about 4 months preceding the general elections in June 2010 with Geert Wilders’ PVV winning the third largest share of seats. After both trends rocketed in 2011, they significantly sank in 2012 and plummeted in 2013, right after the government resolved prematurely in mid-2012. In the general elections in September 2012, the PVV once more came in third; yet this time, VVD and the Partij van de Arbeid (PvdA, Labor Party) were able to form a government without the PVV’s support. Generally, media attention to the European Union and immigration is much more balanced throughout the years. Media attention to the European Union peaks in 2012, after Wilders’ anti-European Union campaign for the general elections. Nevertheless, media attention to immigration and the European Union follows a similar trend as RRP party sympathies. Collectively, these descriptive findings suggest that changes of media attention to RRP parties clearly relate to changes in sympathies for such parties. Further, this association appears much more pronounced than the nexus between media attention to immigration or the European Union with RRP party sympathy. However, these conclusions are based on aggregated data only. To achieve a better understanding if and to what extent media attention shapes RRP sympathy within persons over time, we now turn to the results from hypothesis testing using multilevel structural equation modeling. Measurement Model As mentioned earlier, we measure perceived ethnic threat with a latent variable on two levels, that is, within and between respondents. The results show that our measurement model fits the data really well with χ2 = 5.896, df = 2, p-value = .052, χ2/df = 2.948, RMSEA = .008, CFI = .999, SRMRwithinperson = .007, and SRMRbetweenperson = .003. Further, comparing a model with all factor loadings within and between persons restricted to be equal did not fit worse to the data than the initial model with all factor loadings estimated freely (Δχ2 = 5.896; Δdf = 2; n.s.). This establishes metric measurement invariance, which ensures that the factor loadings reflect the same meaning within and between persons. Interestingly, we find that most variance in respondents’ perceived ethnic threat is situated at the between-person level (ICC = .934). This means that only about 6.5% of all variation of perceived ethnic threat is attributable to differences within persons over time. Nevertheless, because the results show significant within-person variance, we proceed with a decomposed measure of perceived ethnic threat. Structural Model We now present the results of our hypotheses tests. An initial model (not presented) without explanatory variables shows that 76.3% of the total variance of RRP party sympathy is located at the individual level. In other words, 23.7% of differences in sympathies for RRP parties are because of differences over time within individuals. In Table 1, we present the results of our structural model testing Hypotheses 1–4. Model 1a depicts the results for the test of Hypotheses 1 and 2. Substantially, we find a significant positive effect of media attention to Wilders/PVV, as well as media attention to immigration on RRP party sympathies. With this, we can reaffirm our bivariate observations in respect to Hypotheses 1 and 2. While standardized effects are not available for random slope models, we reran Model 1a as a random intercept model to compare the standardized effects of media attention with Wilders/PVV and immigration. The data reveal that media attention to Wilders/PVV exerts a considerably stronger influence (β = 0.165) on respondents’ RRP party sympathy than media attention to immigration (β = 0.054). In additional analyses, we separately included the effects of media attention to either Wilders/PVV or immigration to evaluate the proportion of explained variance. We find that media attention to Wilders/PVV accounts for 11.7% of the within-person variance of RRP party sympathy, and media attention to immigration only accounts for 2.3% of the variance. Table 1 Multilevel Structural Equation Models: RRP Party Sympathy Regressed on Media Attention to PVV/Wilders and Media Attention to Immigration Variable Model 1a Model 1b Model 1c b SE b SE b SE Within person     PVV/Wilders media 0.806 0.035*** 0.299 0.110** 0.78 0.040***     Immigration media 0.294 0.032*** 0.199 0.076** 0.4 0.059***     Unemployed 0.028 0.098 0.036 0.098 0.037 0.098     Media exposure −0.032 0.038 −0.036 0.038 −0.037 0.038     Euroscepticism 0.038 0.014** 0.038 0.014** 0.038 0.014*     Perceived threat 1.363 0.165*** 1.362 0.166*** 1.343 0.165***     PVV/Wilders media × Immigration media – 0.928 0.205*** − Between person     Intercept 2.064 0.134*** 2.131 0.159*** 2.268 0.154***     Male 0.371 0.050*** 0.362 0.050*** 0.365 0.050***     Year of birth 0.018 0.002*** 0.018 0.002*** 0.018 0.002***     Low education – – –     Med education −0.114 0.062 −0.115 0.062 −0.118 0.062     High education −0.232 0.064*** −0.245 0.064*** −0.246 0.065***     Unemployed −0.059 0.337 −0.049 0.338 −0.043 0.338     Media exposure −0.165 0.076* −0.177 0.076* −0.177 0.077*     Euroscepticism 0.294 0.032*** 0.236 0.038*** 0.239 0.038***     Perceived threat 2.277 0.056*** 2.308 0.058*** 2.333 0.057***     PVV/Wilders media × perceived threat – – 0.419 0.052*** Variance within 1.413 0.035*** 1.408 0.035*** 1.418 0.035*** Variance between 2.783 0.068*** 2.863 0.129*** 2.764 0.068*** Variance slope PVV/Wilders 1.091 0.136*** 1.055 0.316** 1.018 0.136*** Variance slope Immigration 0.446 0.299 0.492 0.317 0.317 0.295 Variance slope PVV/Wilders × Immigration – 0.158 0.92 – Cov (slope PVV/Wilders, intercept) 0.33 0.47 0.26 Cov (slope Immigration, intercept) 0.04 0.1 0.04 Cov (slope PVV/Wilders × Immigration, intercept) – −0.291 − Log likelihood −156,695.152 −156,643.606 −156,663.151 AIC 313,456.304 313,361.212 313,396.302 BIC 313,724.628 313,662.06 313,680.888 Variable Model 1a Model 1b Model 1c b SE b SE b SE Within person     PVV/Wilders media 0.806 0.035*** 0.299 0.110** 0.78 0.040***     Immigration media 0.294 0.032*** 0.199 0.076** 0.4 0.059***     Unemployed 0.028 0.098 0.036 0.098 0.037 0.098     Media exposure −0.032 0.038 −0.036 0.038 −0.037 0.038     Euroscepticism 0.038 0.014** 0.038 0.014** 0.038 0.014*     Perceived threat 1.363 0.165*** 1.362 0.166*** 1.343 0.165***     PVV/Wilders media × Immigration media – 0.928 0.205*** − Between person     Intercept 2.064 0.134*** 2.131 0.159*** 2.268 0.154***     Male 0.371 0.050*** 0.362 0.050*** 0.365 0.050***     Year of birth 0.018 0.002*** 0.018 0.002*** 0.018 0.002***     Low education – – –     Med education −0.114 0.062 −0.115 0.062 −0.118 0.062     High education −0.232 0.064*** −0.245 0.064*** −0.246 0.065***     Unemployed −0.059 0.337 −0.049 0.338 −0.043 0.338     Media exposure −0.165 0.076* −0.177 0.076* −0.177 0.077*     Euroscepticism 0.294 0.032*** 0.236 0.038*** 0.239 0.038***     Perceived threat 2.277 0.056*** 2.308 0.058*** 2.333 0.057***     PVV/Wilders media × perceived threat – – 0.419 0.052*** Variance within 1.413 0.035*** 1.408 0.035*** 1.418 0.035*** Variance between 2.783 0.068*** 2.863 0.129*** 2.764 0.068*** Variance slope PVV/Wilders 1.091 0.136*** 1.055 0.316** 1.018 0.136*** Variance slope Immigration 0.446 0.299 0.492 0.317 0.317 0.295 Variance slope PVV/Wilders × Immigration – 0.158 0.92 – Cov (slope PVV/Wilders, intercept) 0.33 0.47 0.26 Cov (slope Immigration, intercept) 0.04 0.1 0.04 Cov (slope PVV/Wilders × Immigration, intercept) – −0.291 − Log likelihood −156,695.152 −156,643.606 −156,663.151 AIC 313,456.304 313,361.212 313,396.302 BIC 313,724.628 313,662.06 313,680.888 Note: *p < .05; **p < .01; ***p < .001; PVV/Wilders media = media attention to the PVV/Wilders; Immigration media = media attention to immigration. Table 1 Multilevel Structural Equation Models: RRP Party Sympathy Regressed on Media Attention to PVV/Wilders and Media Attention to Immigration Variable Model 1a Model 1b Model 1c b SE b SE b SE Within person     PVV/Wilders media 0.806 0.035*** 0.299 0.110** 0.78 0.040***     Immigration media 0.294 0.032*** 0.199 0.076** 0.4 0.059***     Unemployed 0.028 0.098 0.036 0.098 0.037 0.098     Media exposure −0.032 0.038 −0.036 0.038 −0.037 0.038     Euroscepticism 0.038 0.014** 0.038 0.014** 0.038 0.014*     Perceived threat 1.363 0.165*** 1.362 0.166*** 1.343 0.165***     PVV/Wilders media × Immigration media – 0.928 0.205*** − Between person     Intercept 2.064 0.134*** 2.131 0.159*** 2.268 0.154***     Male 0.371 0.050*** 0.362 0.050*** 0.365 0.050***     Year of birth 0.018 0.002*** 0.018 0.002*** 0.018 0.002***     Low education – – –     Med education −0.114 0.062 −0.115 0.062 −0.118 0.062     High education −0.232 0.064*** −0.245 0.064*** −0.246 0.065***     Unemployed −0.059 0.337 −0.049 0.338 −0.043 0.338     Media exposure −0.165 0.076* −0.177 0.076* −0.177 0.077*     Euroscepticism 0.294 0.032*** 0.236 0.038*** 0.239 0.038***     Perceived threat 2.277 0.056*** 2.308 0.058*** 2.333 0.057***     PVV/Wilders media × perceived threat – – 0.419 0.052*** Variance within 1.413 0.035*** 1.408 0.035*** 1.418 0.035*** Variance between 2.783 0.068*** 2.863 0.129*** 2.764 0.068*** Variance slope PVV/Wilders 1.091 0.136*** 1.055 0.316** 1.018 0.136*** Variance slope Immigration 0.446 0.299 0.492 0.317 0.317 0.295 Variance slope PVV/Wilders × Immigration – 0.158 0.92 – Cov (slope PVV/Wilders, intercept) 0.33 0.47 0.26 Cov (slope Immigration, intercept) 0.04 0.1 0.04 Cov (slope PVV/Wilders × Immigration, intercept) – −0.291 − Log likelihood −156,695.152 −156,643.606 −156,663.151 AIC 313,456.304 313,361.212 313,396.302 BIC 313,724.628 313,662.06 313,680.888 Variable Model 1a Model 1b Model 1c b SE b SE b SE Within person     PVV/Wilders media 0.806 0.035*** 0.299 0.110** 0.78 0.040***     Immigration media 0.294 0.032*** 0.199 0.076** 0.4 0.059***     Unemployed 0.028 0.098 0.036 0.098 0.037 0.098     Media exposure −0.032 0.038 −0.036 0.038 −0.037 0.038     Euroscepticism 0.038 0.014** 0.038 0.014** 0.038 0.014*     Perceived threat 1.363 0.165*** 1.362 0.166*** 1.343 0.165***     PVV/Wilders media × Immigration media – 0.928 0.205*** − Between person     Intercept 2.064 0.134*** 2.131 0.159*** 2.268 0.154***     Male 0.371 0.050*** 0.362 0.050*** 0.365 0.050***     Year of birth 0.018 0.002*** 0.018 0.002*** 0.018 0.002***     Low education – – –     Med education −0.114 0.062 −0.115 0.062 −0.118 0.062     High education −0.232 0.064*** −0.245 0.064*** −0.246 0.065***     Unemployed −0.059 0.337 −0.049 0.338 −0.043 0.338     Media exposure −0.165 0.076* −0.177 0.076* −0.177 0.077*     Euroscepticism 0.294 0.032*** 0.236 0.038*** 0.239 0.038***     Perceived threat 2.277 0.056*** 2.308 0.058*** 2.333 0.057***     PVV/Wilders media × perceived threat – – 0.419 0.052*** Variance within 1.413 0.035*** 1.408 0.035*** 1.418 0.035*** Variance between 2.783 0.068*** 2.863 0.129*** 2.764 0.068*** Variance slope PVV/Wilders 1.091 0.136*** 1.055 0.316** 1.018 0.136*** Variance slope Immigration 0.446 0.299 0.492 0.317 0.317 0.295 Variance slope PVV/Wilders × Immigration – 0.158 0.92 – Cov (slope PVV/Wilders, intercept) 0.33 0.47 0.26 Cov (slope Immigration, intercept) 0.04 0.1 0.04 Cov (slope PVV/Wilders × Immigration, intercept) – −0.291 − Log likelihood −156,695.152 −156,643.606 −156,663.151 AIC 313,456.304 313,361.212 313,396.302 BIC 313,724.628 313,662.06 313,680.888 Note: *p < .05; **p < .01; ***p < .001; PVV/Wilders media = media attention to the PVV/Wilders; Immigration media = media attention to immigration. We include an interaction term of media attention to Wilders/PVV and immigration in Model 1b, to test our expectation regarding their joint effect on RRP party sympathies (Hypothesis 3). The results reveal a significant interaction effect. Figure 2 presents the marginal effects of media attention to immigration by media attention to Wilders/PVV. The results show that the positive effect of media attention to immigration is only significant when media attention to Wilders/PVV is at least at an average level. The effect increases with the level of media attention to Wilders/PVV. These findings support Hypothesis 3. Including the interaction between media attention to Wilders/PVV and immigration decreases the within-person variance of RRP party sympathy by only 0.3%. Figure 2 View largeDownload slide Marginal effect of media attention to immigration on RRP party sympathy by media attention to PVV/Wilders (Model 1b). Note: Dashed lines are 95% confidence intervals Figure 2 View largeDownload slide Marginal effect of media attention to immigration on RRP party sympathy by media attention to PVV/Wilders (Model 1b). Note: Dashed lines are 95% confidence intervals In Models 1a and 1b, we find that the effect of media attention to Wilders/PVV significantly varies across individuals. Thus, we proceed by testing Hypothesis 4, the interaction of media attention to Wilders/PVV with perceived ethnic threat on RRP party sympathies. Thereby, we are testing if the variability of media attention’s effect is depending on an individual level of ethnic threat perception, that is, the between-person level variable. Notably, neither the effect of media attention to immigration nor the interaction between both salience measures shows significant slope variance. Model 1c reports the interaction effect between media attention to Wilders/PVV and perceived ethnic threat on RRP party sympathies. The results show a positive and significant interaction. To interpret the underlying effect, we turn to the marginal effects plot presented in Figure 3. The plot shows that the positive effect of perceived ethnic threat is larger, when media attention to Wilders/PVV increases, as we postulated in Hypothesis 4. The interaction effect between media attention to Wilders/PVV and perceived ethnic threat explains 6.7% of the slope variation of the effect of media attention to Wilders/PVV. In other words, 6.7% of the differences of the effect of media attention to Wilders/PVV on RRP party sympathies are explained by perceived ethnic threat. Because conventional measures of variance explained are considered less suitable to gauge the relevance of cross-level interactions, we consider the differences of BIC between a model without versus a model with the cross-level interaction. Comparing Model 1a with Model 1c, we find a rather large difference of ΔBIC = 43.74, which is strong evidence that Model 1c is the preferred model (Raftery, 1995). Figure 3 View largeDownload slide Marginal effect of perceived ethnic threat on RRP party sympathy by media attention to PVV/Wilders (Model 1c). Note: Dashed lines are 95% confidence intervals Figure 3 View largeDownload slide Marginal effect of perceived ethnic threat on RRP party sympathy by media attention to PVV/Wilders (Model 1c). Note: Dashed lines are 95% confidence intervals Table 2 presents the results for our tests of Hypotheses 5–7. In Model 2a, we find that media attention to the European Union has a significant positive effect on RRP party sympathy. As in Table 1, we find a significant positive effect of media attention to Wilders/PVV on RRP party sympathy. The explanatory contribution of media attention to the European Union is only marginal. To compare the strength of the effects of media attention with Wilders/PVV and the European Union, we reran Model 2a with fixed slopes. We find that the standardized effects of media attention to the European Union on respondents’ RRP party sympathy (β = 0.019) are much weaker than the effect of media attention to Wilders/PVV (β = 0.185). Including the effect of media attention to the European Union into the model explains only 0.9% of the within individual variance of RRP party sympathy. Turning to the interaction effect between media attention to the European Union and media attention to Wilders/PVV, tested in Model 2b, we find a significant negative effect. The findings of the interaction effect are presented in Figure 4. Contrary to our expectations of Hypothesis 6, the marginal effects plot shows that the effect of media attention to the European Union slightly decreases, when media attention to Wilders/PVV increases. The interaction effect between media attention to Wilders/PVV and the European Union decreases the variance of RRP party sympathy by only 0.1%. We find no significant slope variance for media attention to the European Union and thus proceed with the test of the interaction effect between media attention to Wilders/PVV and Euroscepticism (Hypothesis 7). The results of Model 2c show a significant positive interaction between Euroscepticism and media attention to Wilders and his PVV. Figure 5 displays the marginal effects of Euroscepticism by media attention to Wilders/PVV. We find that the positive effect of Euroscepticism is larger, when media attention to Wilders/PVV increases. The explanatory power of the interaction effect between media attention to Wilders/PVV and Euroscepticism is only limited. Including the moderating effect decreases the slope variance of the effect of media attention to Wilders/PVV on RRP party sympathy by only 3.0%. However, a BIC difference of ΔBIC = 11.92 between Model 2a and Model 2c suggests that including the interaction substantively increases the model fit. Table 2 Multilevel Structural Equation Models: RRP Party Sympathy Regressed on Media Attention to PVV/Wilders and Media Attention to the European Union Variable Model 2a Model 2b Model 2c b SE b SE b SE Within person     PVV/Wilders media 0.866 0.035*** 1.112 0.105*** 0.201 0.14     EU media 0.171 0.053** 0.348 0.100** 0.169 0.053**     Unemployed 0.013 0.097 0.019 0.097 0.012 0.097     Media exposure −0.023 0.038 −0.029 0.038 −0.022 0.038     Euroscepticism 0.036 0.014* 0.039 0.014* 0.034 0.014*     Perceived threat 1.392 0.166*** 1.38 0.166*** 1.398 0.166***     PVV/Wilders media × EU media – −0.427 0.180* – Between person     Intercept 2.272 0.154*** 2.385 0.159*** 2.232 0.154***     Male 0.364 0.050*** 0.364 0.050*** 0.364 0.050***     Year of birth 0.018 0.002*** 0.018 0.002*** 0.018 0.002***     Low education – – –     Med education −0.115 0.062 −0.119 0.062 −0.115 0.062     High education −0.243 0.064*** −0.249 0.064*** −0.242 0.064***     Unemployed −0.037 0.337 −0.053 0.338 −0.036 0.338     Media exposure −0.179 0.076* −0.169 0.076* −0.179 0.077*     Euroscepticism 0.238 0.038*** 0.23 0.038*** 0.25 0.038***     Perceived threat 2.312 0.057*** 2.316 0.057*** 2.312 0.057*** Media attention × Euroscepticism – – 0.192 0.041*** Variance within 1.433 0.033*** 1.431 0.033*** 1.433 0.033*** Variance between 2.766 0.068*** 2.929 0.135*** 2.765 0.068*** Variance slope PVV/Wilders 1.066 0.135*** 1.008 0.256*** 1.034 0.133*** Variance slope EU 0.064 0.476 0.039 0.504 0.063 0.474 Variance slope PVV/Wilders × EU – 0.149 0.522 – Cov (slope PVV/Wilders, intercept) 0.36 0.57 0.35 Cov (slope EU, intercept) −0.11 −0.01 −0.11 Cov (slope PVV/Wilders × EU, intercept) – −0.39 – Log likelihood −156,790.602 −156,740.085 −156,779.57 AIC 313,649.203 313,554.171 313,629.141 BIC 313,925.658 313,855.018 313,913.726 Variable Model 2a Model 2b Model 2c b SE b SE b SE Within person     PVV/Wilders media 0.866 0.035*** 1.112 0.105*** 0.201 0.14     EU media 0.171 0.053** 0.348 0.100** 0.169 0.053**     Unemployed 0.013 0.097 0.019 0.097 0.012 0.097     Media exposure −0.023 0.038 −0.029 0.038 −0.022 0.038     Euroscepticism 0.036 0.014* 0.039 0.014* 0.034 0.014*     Perceived threat 1.392 0.166*** 1.38 0.166*** 1.398 0.166***     PVV/Wilders media × EU media – −0.427 0.180* – Between person     Intercept 2.272 0.154*** 2.385 0.159*** 2.232 0.154***     Male 0.364 0.050*** 0.364 0.050*** 0.364 0.050***     Year of birth 0.018 0.002*** 0.018 0.002*** 0.018 0.002***     Low education – – –     Med education −0.115 0.062 −0.119 0.062 −0.115 0.062     High education −0.243 0.064*** −0.249 0.064*** −0.242 0.064***     Unemployed −0.037 0.337 −0.053 0.338 −0.036 0.338     Media exposure −0.179 0.076* −0.169 0.076* −0.179 0.077*     Euroscepticism 0.238 0.038*** 0.23 0.038*** 0.25 0.038***     Perceived threat 2.312 0.057*** 2.316 0.057*** 2.312 0.057*** Media attention × Euroscepticism – – 0.192 0.041*** Variance within 1.433 0.033*** 1.431 0.033*** 1.433 0.033*** Variance between 2.766 0.068*** 2.929 0.135*** 2.765 0.068*** Variance slope PVV/Wilders 1.066 0.135*** 1.008 0.256*** 1.034 0.133*** Variance slope EU 0.064 0.476 0.039 0.504 0.063 0.474 Variance slope PVV/Wilders × EU – 0.149 0.522 – Cov (slope PVV/Wilders, intercept) 0.36 0.57 0.35 Cov (slope EU, intercept) −0.11 −0.01 −0.11 Cov (slope PVV/Wilders × EU, intercept) – −0.39 – Log likelihood −156,790.602 −156,740.085 −156,779.57 AIC 313,649.203 313,554.171 313,629.141 BIC 313,925.658 313,855.018 313,913.726 Note: *p < .05; **p < .01; ***p < .001; PVV/Wilders media = media attention to the PVV/Wilders; EU media = media attention to the European Union. Table 2 Multilevel Structural Equation Models: RRP Party Sympathy Regressed on Media Attention to PVV/Wilders and Media Attention to the European Union Variable Model 2a Model 2b Model 2c b SE b SE b SE Within person     PVV/Wilders media 0.866 0.035*** 1.112 0.105*** 0.201 0.14     EU media 0.171 0.053** 0.348 0.100** 0.169 0.053**     Unemployed 0.013 0.097 0.019 0.097 0.012 0.097     Media exposure −0.023 0.038 −0.029 0.038 −0.022 0.038     Euroscepticism 0.036 0.014* 0.039 0.014* 0.034 0.014*     Perceived threat 1.392 0.166*** 1.38 0.166*** 1.398 0.166***     PVV/Wilders media × EU media – −0.427 0.180* – Between person     Intercept 2.272 0.154*** 2.385 0.159*** 2.232 0.154***     Male 0.364 0.050*** 0.364 0.050*** 0.364 0.050***     Year of birth 0.018 0.002*** 0.018 0.002*** 0.018 0.002***     Low education – – –     Med education −0.115 0.062 −0.119 0.062 −0.115 0.062     High education −0.243 0.064*** −0.249 0.064*** −0.242 0.064***     Unemployed −0.037 0.337 −0.053 0.338 −0.036 0.338     Media exposure −0.179 0.076* −0.169 0.076* −0.179 0.077*     Euroscepticism 0.238 0.038*** 0.23 0.038*** 0.25 0.038***     Perceived threat 2.312 0.057*** 2.316 0.057*** 2.312 0.057*** Media attention × Euroscepticism – – 0.192 0.041*** Variance within 1.433 0.033*** 1.431 0.033*** 1.433 0.033*** Variance between 2.766 0.068*** 2.929 0.135*** 2.765 0.068*** Variance slope PVV/Wilders 1.066 0.135*** 1.008 0.256*** 1.034 0.133*** Variance slope EU 0.064 0.476 0.039 0.504 0.063 0.474 Variance slope PVV/Wilders × EU – 0.149 0.522 – Cov (slope PVV/Wilders, intercept) 0.36 0.57 0.35 Cov (slope EU, intercept) −0.11 −0.01 −0.11 Cov (slope PVV/Wilders × EU, intercept) – −0.39 – Log likelihood −156,790.602 −156,740.085 −156,779.57 AIC 313,649.203 313,554.171 313,629.141 BIC 313,925.658 313,855.018 313,913.726 Variable Model 2a Model 2b Model 2c b SE b SE b SE Within person     PVV/Wilders media 0.866 0.035*** 1.112 0.105*** 0.201 0.14     EU media 0.171 0.053** 0.348 0.100** 0.169 0.053**     Unemployed 0.013 0.097 0.019 0.097 0.012 0.097     Media exposure −0.023 0.038 −0.029 0.038 −0.022 0.038     Euroscepticism 0.036 0.014* 0.039 0.014* 0.034 0.014*     Perceived threat 1.392 0.166*** 1.38 0.166*** 1.398 0.166***     PVV/Wilders media × EU media – −0.427 0.180* – Between person     Intercept 2.272 0.154*** 2.385 0.159*** 2.232 0.154***     Male 0.364 0.050*** 0.364 0.050*** 0.364 0.050***     Year of birth 0.018 0.002*** 0.018 0.002*** 0.018 0.002***     Low education – – –     Med education −0.115 0.062 −0.119 0.062 −0.115 0.062     High education −0.243 0.064*** −0.249 0.064*** −0.242 0.064***     Unemployed −0.037 0.337 −0.053 0.338 −0.036 0.338     Media exposure −0.179 0.076* −0.169 0.076* −0.179 0.077*     Euroscepticism 0.238 0.038*** 0.23 0.038*** 0.25 0.038***     Perceived threat 2.312 0.057*** 2.316 0.057*** 2.312 0.057*** Media attention × Euroscepticism – – 0.192 0.041*** Variance within 1.433 0.033*** 1.431 0.033*** 1.433 0.033*** Variance between 2.766 0.068*** 2.929 0.135*** 2.765 0.068*** Variance slope PVV/Wilders 1.066 0.135*** 1.008 0.256*** 1.034 0.133*** Variance slope EU 0.064 0.476 0.039 0.504 0.063 0.474 Variance slope PVV/Wilders × EU – 0.149 0.522 – Cov (slope PVV/Wilders, intercept) 0.36 0.57 0.35 Cov (slope EU, intercept) −0.11 −0.01 −0.11 Cov (slope PVV/Wilders × EU, intercept) – −0.39 – Log likelihood −156,790.602 −156,740.085 −156,779.57 AIC 313,649.203 313,554.171 313,629.141 BIC 313,925.658 313,855.018 313,913.726 Note: *p < .05; **p < .01; ***p < .001; PVV/Wilders media = media attention to the PVV/Wilders; EU media = media attention to the European Union. Figure 4 View largeDownload slide Marginal effect of media attention to the European Union on RRP party sympathy by media attention to PVV/Wilders (Model 2b). Note: Dashed lines are 95% confidence intervals Figure 4 View largeDownload slide Marginal effect of media attention to the European Union on RRP party sympathy by media attention to PVV/Wilders (Model 2b). Note: Dashed lines are 95% confidence intervals Figure 5 View largeDownload slide Marginal effect of Euroscepticism on RRP party sympathy by media attention to PVV/Wilders (Model 2c). Note: Dashed lines are 95% confidence intervals Figure 5 View largeDownload slide Marginal effect of Euroscepticism on RRP party sympathy by media attention to PVV/Wilders (Model 2c). Note: Dashed lines are 95% confidence intervals To account for potential confounding factors, we included a set of control variables in all models presented above. The results show that the effects are generally in line with previous research. We find that men have higher levels of RRP party sympathies, as well as younger respondents. Moreover, we can show that education decreases RRP party sympathies. We do not find an effect of unemployment.9 Nonetheless, accounting for these factors does not alter our findings.10 An alternative conceptualization of media attention to RRP parties and issues that are (perceived as) relevant to those parties is the co-occurrence of the RRP party and the European Union or immigration in the same article. We conducted additional analyses for the congruent media attention to RRP parties and to the European Union or to immigration, respectively (Supplementary Appendix Tables B and C). The findings support the results presented here. The main effects and interactions are positive and statistically significant. The effect of congruent media attention to RRP parties and immigration, as well as the interaction with perceived ethnic threat, is considerably larger than the effect of congruent media attention to RRP parties and the European Union, and the interaction with Euroscepticism. The correlation between the explanatory variables and the dependent variable might result from some unobserved, time trending factor (Wooldridge, 2009: 363). Thus, we reran our analysis controlling for survey year measured with a continuous variable ranging from 0 (2008) to 5 (2013). Furthermore, the relationship might be linked to the perceived party size or actual immigration inflow. We therefore reran our analyses including polling results11 and migration rates of nonwestern immigrants. The immigration rates between 2008 and 2013 follow a strict linear trend and highly correlate with survey years. In addition, we reran all full models including a lagged dependent variable, to control for the autoregressive effect of RRP party sympathies. The effect of media attention to the European Union turns out to be not significant after including party sizes, survey year, and a lagged dependent variable. The media attention to the European Union has only limited variance over time and might be linked to an unobserved time trending factor. Nonetheless, the results are generally in line with those presented in the paper. All robustness checks are presented in Supplementary Appendix Table D. To further focus on the within-person variation, we also estimate a fixed-effects panel regression model. The results are shown in Supplementary Appendix Table E. We find no substantial differences to our findings based on the multilevel structural equation model presented in Table 1 and Table 2. To test if the effect of media attention depends on whether respondents have read the newspaper, we included an interaction with media exposure (not shown). However, we find no significant interaction between exposure and media attention, which further supports the notion that issues frequently mentioned in the news are also salient in the public debate (Yang and Stone, 2003). Discussion The goal of this research was to examine individual changes of RRP party sympathies and how they relate to media attention. In particular, we first analyzed to what extent media attention relates to changes in RRP party sympathies. Second, we assessed if, and if so, to what extent individual attitudes interact with the effects of media attention on RRP party sympathies. To investigate our theoretical assumptions, we used a multilevel structural equation approach drawing on six annual waves of a representative Dutch panel survey from 2008 to 2013. To gauge the effects of media attention, we combined the panel data with results from a computer-assisted content analysis of media attention to Wilders/PVV, immigration, and the European Union. Our results show that media attention is positively related to individual sympathies for RRP parties. We find that while the explanatory contribution of media attention to immigration or to the European Union is only marginal, the proportion of explained variance by media attention to Wilders/PVV is rather substantial. Further analyses reveal that perceived ethnic threat and Euroscepticism moderate the effect of media attention on individual RRP party sympathies. With these findings, we show that visibility of RRP parties does not only mobilize the potential RRP electorate but interacts with attitudinal predictors of RRP party sympathy. In particular, among people who perceive stronger ethnic threats, more media attention to the RRP party is associated with more sympathy for this party. Similarly, Euroscepticism strengthens the effect of media attention to the RRP party on sympathy for the PVV. It is critical to note that ethnic threat explains significantly more variance of the effect of media attention to PVV/Wilders on RRP party sympathy compared to Euroscepticism. These findings extend previous research in several significant ways. First, our findings expand on earlier research on the effect of media attention on changes in RRP party sympathies. These findings are particularly important, as antecedents for within-individual changes of RRP party sympathies is largely underresearched. Second, the interaction of perceived ethnic threat and Euroscepticism with media attention is crucial to shed light on the puzzling longitudinal relationship of rather stable attitudes and volatile RRP party sympathies. Third, this study contributes to an evolving line of research, which shows that populism is increasingly part of newspaper reports and the public debate (Rooduijn, 2014). In sum, we extend our theoretical knowledge of the underlying mechanism at work, especially since group threat theory postulates a dynamic relationship. The interpretations of our findings come with certain limitations. We only draw inference on the effects of media attention in the sense of general salience. We refrain from analyses of the news reports’ valence. On the one hand, we believe that it is first and most importantly the visibility of topics that affects RRP sympathies, rather than the tone of news content. On the other hand, we have no information on the news a respondent consumed. Furthermore, one can question the generalizability of our findings beyond the Dutch context. Nevertheless, previous research shows similar effects of media attention on RRP party sympathies across different RRP parties and contexts (Vliegenthart et al., 2012). That said, we encourage future research to investigate to what extent the longitudinal mechanism presented here is potentially moderated by cross-country differences. Relatedly, we believe that future research might identify potentially moderating conditions by integrating the findings presented here into a cross-party comparison. To conclude, the present study provides insights on the longitudinal explanation underlying sympathies for RRP parties. Our study shows that media attention affects individual changes of RRP party sympathies. We substantiate this effect with evidence that the relationship between media attention and RRP party sympathies positively depends on individual perceived ethnic threat and Euroscepticism. Supplementary Data Supplementary Data are available at IJPOR online. Acknowledgment The LISS panel data were collected by CentERdata (Tilburg University, The Netherlands) through its MESS project funded by The Netherlands Organization for Scientific Research. Footnotes 1In this study, we use the term radical RRP parties. This term goes back to the seminal work of Betz (1993) but is also repeatedly used in more recent work especially for the party we study (Akkerman et al. 2016; Berning and Schlueter, 2016). We note that the classification and labeling of this party family were and still are subjects to rigorous scientific debate (Eger and Valdez, 2015; Mudde, 1996). 2In contrast to the relation mentioned above, one could at least argue on the reverse causal direction, that is, an increase in sympathies for RRP parties will lead to an increase in media attention. In the present study, we refrain from causal interpretation; yet, using a conceptualization consistent with related research (Schlueter and Davidov, 2013; Vliegenthart et al., 2012). Moreover, Van der Pas et al. (2011) have shown that increased support for the PVV was related to less visibility of the party in the media a week later. 3The LISS panel was conducted via Internet. Recent research shows that in the case of LISS, the interview mode had no effect on data quality (Revilla and Saris, 2013). All respondents without access to the Internet were provided with a PC and adequate devices. Further information in the LISS panel can be found at www.lissdata.nl. 4Of all respondents, 50.3% participated in all six waves, 16.5% in five waves, 16.2% only in four waves, 8.3% in three waves, and 8.7% participated only in two waves. We reran our analyses limited to individuals who were interviewed in all consecutive survey rounds and found that the results were substantively identical. 5We replicated all analyses with only single items for RRP sympathies, that is, sympathy for the leader and sympathy for the party, and found no substantial differences in the results, because of high correlation between both items (r = .92). 6Specifically, to measure media attention to PVV/Wilders, we searched for “DATE(<[Date of interview]) AND DATE(>=[Date of interview - 8 weeks]) AND (PVV OR Wilders).” To measure articles related to immigration, we searched for “DATE(<[Date of interview]) AND DATE(>=[Date of interview - 8 weeks]) AND (discrim OR (haat w/5 aanzet) OR inburgering* OR (scholing or (cursus* OR les* OR onderwijs) w/10 (immi* OR alloch* OR asiel* OR buitenl*)) OR taalcur* OR taalles* OR taalonderw* OR gezinsherenig* OR schijnhuw* OR nephuw* OR uithuw* OR immig* OR alloch* OR multicult* OR (verpaup* AND (buurt* or wijk*)) OR moslim* OR islam* OR asiel* OR uitgeproc* OR verblijfs* OR (grondwet w/10 artikel 7) OR importbruid OR (bruid* AND buitenland) OR (inkomenseis w/20 trouw*) OR pluriform* OR asielzoeker* OR vluchteling* OR (generaal pardon) OR pardonregeling)” and to measure media attention to the European Union, we searched for: “DATE(<[Date of interview]) AND DATE(>=[Date of interview - 8 weeks]) AND (europese pre/1 unie) OR EU.” 7The different media attention measures correlate with each other. PVV/Wilders media with Immigration mediar = .42; PVV/Wilders media with EU mediar =.32; EU media with Immigration mediar = .56. 8In subsequent analyses, we used a subset of interview dates and matched it to media attention scores based on articles published 6 weeks and 10 weeks before the interview. The shorter and longer intervals did not alter our results. 9We also tested an interaction of media attention to Wilders/PVV and perceived ethnic threat at the within-person level. Controlling for the within-level interaction led to estimation problems because of only small variance of perceived ethnic threat at the lower level. To reduce model complexity given the rather small number of time points, we decided to only allow the slopes of media attention to vary across individuals. Subsequential analyses showed that setting slopes of other within measures at random did not alter our results. 10Previous research further introduced political dissatisfaction as a motivation to prefer RRP parties (Mudde, 2007; but Zhirkov, 2013). Aside the merits of this factor as an additional motivation to prefer an RRP party, the longitudinal argument of political satisfaction to predict sympathies for a RRP party who joined the government is theoretically inconclusive. Nevertheless, to test the robustness of our findings, we reran our analyses including the potential confounding factor. The results were substantially concordant. 11We use the polling results from “de politieke barometer” conducted by Ipsos Synovate. References Aaldering L. , Vliegenthart R. ( 2016 ). Political leaders and the media: Can we measure political leadership images in newspapers using computer-assisted content analysis? Quality and Quantity , 50 , 1871 – 1905 .. doi:10.1007/s11135-015-0242-9 Google Scholar CrossRef Search ADS PubMed Akkerman T. , de Lange S. L. , Rooduijn M. ( 2016 ). Radical right-wing populist parties in Western Europe. Into the mainstream ? Oxford : Routledge . Arzheimer K. ( 2008 ). Protest, neo-liberalism or anti-immigrant sentiment: What motivates the voters of the extreme right in Western Europe? Comparative Governance and Politics , 2 , 173 – 197 .. doi:10.1007/s12286-008-0011-4 Arzheimer K. ( 2012 ). Electoral sociology: Who votes for the extreme right and why - and when? In Backes U. , Morrow D. (Eds.), The extreme right in Europe. Current trends and perspectives (pp. 35 – 50 ). Göttingen : Vendenhoeck & Ruprecht . Beck P. A. , Dalton R. J. , Greene S. , Huckfeldt R. ( 2002 ). The social calculus of voting: Interpersonal, media, and organizational influence on presidential choice . American Political Science Review , 96 ( 1 ), 37 – 73 . Google Scholar CrossRef Search ADS Bentler P. M. ( 1990 ). Comparative fit indexes in structural models . Psychological Bulletin , 107 , 238 – 246 . doi:10.1037/0033-2909.107.2.238 Google Scholar CrossRef Search ADS PubMed Berning C. C. ( 2016 ). Contextual perceived group threat and radical right-wing populist party preferences: Evidence from Switzerland . Research and Politics , 3 ( 1 ). Retrieved from https://doi.org/10.1177/2053168016635670 Berning C. C. , Schlueter E. ( 2016 ). The dynamics of radical right-wing populist party preferences and perceived group threat: A comparative panel analysis of three competing hypotheses in The Netherlands and Germany . Social Science Research , 55 , 83 – 93 . doi: 10.1016/j.ssresearch.2015.09.003 Google Scholar CrossRef Search ADS PubMed Betz H. G. ( 1993 ). The New Politics of Resentment. Radical Right-Wing Populist Parties in Western Europe . Comparative Politics , 25 , 413 – 427 . Google Scholar CrossRef Search ADS Betz H. G. ( 1994 ). Radical right-wing populism in Western Europe . London : Macmillan Basingstoke . Google Scholar CrossRef Search ADS Blalock H. M. ( 1967 ). Toward a theory of minority-group relations . New York, NY : Wiley . Boomgaarden H. G. , Vliegenthart R. ( 2007 ). Explaining the rise of anti-immigrant parties: The role of news media content . Electoral Studies , 26 , 404 – 417 . doi:10.1016/j.electstud.2006.10.018 Google Scholar CrossRef Search ADS Boomsma A. ( 2000 ). Reporting analyses of covariance structures . Structural Equation Modeling , 7 , 461 – 483 . doi:10.1207/S15328007SEM0703_6 Google Scholar CrossRef Search ADS Coffé H. , Voorpostel M. ( 2010 ). Young people, parents and radical right voting. The case of the Swiss people’s party . Electoral Studies , 29 , 435 – 443 . doi:10.1016/j.electstud.2010.03.015 Google Scholar CrossRef Search ADS Cohen B. C. ( 1963 ). The press and foreign policy . Princeton, NJ : Princeton University Press . Curran P. J. , Bauer D. J. ( 2011 ). The disaggregation of within-person and between-person effects in longitudinal models of change . Annual Review of Psychology , 62 ( 1 ), 583 – 619 . doi:10.1146/annurev.psych.093008.100356 Google Scholar CrossRef Search ADS PubMed Cutts D. C. , Ford R. , Goodwin M. J. ( 2011 ). Anti-immigrant, politically disaffected or still racist after all? Examining the attitudinal drivers of extreme right support in Britain in the 2009 European elections . European Journal of Political Research , 50 , 418 – 440 . doi: 10.1111/j.1475-6765.2010.01936.x Google Scholar CrossRef Search ADS Dalton R. J. ( 2014 ). Citizen politics: Public opinion and political parties in advanced industrial democracies ( 6th ed. ). Thousand Oaks, CA : CQ Press . Eger M. A. , Valdez S. ( 2015 ). Neo-nationalism in Western Europe . European Sociological Review , 31 ( 1 ), 115 – 130 . doi: 10.1093/esr/jcu087 Google Scholar CrossRef Search ADS Elchardus M. , Spruyt B. ( 2010 ). Does Higher Education Influence the Attitudes with Regard to the Extreme Right? European Journal of Social Sciences , 18 , 181 – 195 . Elchardus M. , Spruyt B. ( 2012 ). The contemporary contradictions of egalitarianism: An empirical analysis of the relationship between the old and new left/right alignments . European Political Sciences Review , 4 , 217 – 239 . doi:10.1017/S1755773911000178 Google Scholar CrossRef Search ADS Enders C. K. , Tofighi D. ( 2007 ). Centering predictor variables in cross-sectional multilevel models: a new look at an old issue . Psychological Methods , 12 , 121 – 138 . doi:10.1037/1082-989X.12.2.121 Google Scholar CrossRef Search ADS PubMed Gerber A. S. , Karlan D. , Bergan D. ( 2009 ). Does the media matter? A field experiment measuring the effect of newspapers on voting behavior and political opinions . American Economic Journal: Applied Economics , 1 , 35 – 52 . Google Scholar CrossRef Search ADS Givens T. E. ( 2004 ). The radical right gender gap . Comparative Political Studies , 37 ( 1 ), 30 – 54 . doi:10.1177/0010414003260124 Google Scholar CrossRef Search ADS Hopmann D. N. , Vliegenthart R. , De Vreese C. , Albæk E. ( 2010 ). Effects of election news coverage: How visibility and tone influence party choice . Political Communication , 27 , 389 – 405 . doi: 10.1080/10584609.2010.516798 Google Scholar CrossRef Search ADS Hox J. ( 2002 ). Multilevel analysis: Techniques and applications . Mahwah, NJ : Lawrence Erlbaum Associates . Hu L. , Bentler P. M. ( 1999 ). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives . Structural Equation Modeling , 6 ( 1 ), 1 – 55 . doi:10.1080/10705519909540118 Google Scholar CrossRef Search ADS Immerzeel T. , Lubbers M. , Coffé H. ( 2015 ). Competing with the radical right: Distances between the European radical right and other parties on typical radical right issues . Party Politics , 22 , 823 – 834 . Published online before print January 15, 2015. doi:10.1177/1354068814567975 Google Scholar CrossRef Search ADS Ivarsflaten E. ( 2008 ). What unites right-wing populists in Western Europe? Re-examining grievance mobilization models in seven successful cases . Comparative Political Studies , 41 ( 1 ), 3 – 23 . doi:10.1177/0010414006294168 Google Scholar CrossRef Search ADS Iyengar S. , Kinder D. R. ( 1987 ). News that matters: Television and American opinion . Chicago : University of Chicago Press . Kleinnijenhuis J. , van Hoof A. M. , Oegema D. , de Ridder J. ( 2007 ). A test of rivaling approaches to explain news effects: News on issue positions of parties, real-world developments, support and criticism, and success and failure . Journal of Communication , 57 , 366 – 384 . doi: 10.1111/j.1460-2466.2007.00347.x Google Scholar CrossRef Search ADS Koopmans R. , Muis J. ( 2009 ). The rise of right-wing populist Pim Fortuyn in The Netherlands: A discursive opportunity approach . European Journal of Political Research , 48 , 642 – 664 . doi: 10.1111/j.1475-6765.2009.00846.x Google Scholar CrossRef Search ADS Lancee B. , Pardos-Prado S. ( 2013 ). Group conflict theory in a longitudinal perspective: Analyzing the dynamic side of ethnic competition . International Migration Review , 47 ( 1 ), 106 – 131 . doi: 10.1111/imre.12015 Google Scholar CrossRef Search ADS Lubbers M. , Gijsberts M. , Scheepers P. ( 2002 ). Extreme right-wing voting in Western Europe . European Journal of Political Research , 41 , 345 – 378 . doi: 10.1111/1475-6765.00015 Google Scholar CrossRef Search ADS Lubbers M. , Scheepers P. ( 2001 ). Explaining the trend in extreme right-wing voting. Germany 1989-1998 . European Sociological Review , 17 , 431 – 449 . doi:10.1093/esr/17.4.431 Google Scholar CrossRef Search ADS Lucassen G. , Lubbers M. ( 2012 ). Who fears what? Explaining far-right-wing preference in Europe by distinguishing perceived cultural and economic ethnic threats . Comparative Political Studies , 45 , 547 – 574 . doi: 10.1177/0010414011427851 Google Scholar CrossRef Search ADS Marsh H. W. , Hocevar D. ( 1985 ). Application of confirmatory factor analysis to the study of self-concept: First- and higher order factor models and their invariance across groups . Psychological Bulletin , 97 , 562 – 582 . doi: 10.1037/0033-2909.97.3.562 Google Scholar CrossRef Search ADS Marsh H. W. , Lüdtke O. , Robitzsch A. , Trautwein U. , Asparouhov T. , Muthén B., , Nagengast B. ( 2009 ). Doubly-latent models of school contextual effects: Integrating multilevel and structural equation approaches to control measurement and sampling error . Multivariate Behavioral Research , 44 , 764 – 802 . doi:10.1080/00273170903333665 Google Scholar CrossRef Search ADS PubMed McCombs M. E. , Shaw D. L. ( 1972 ). The agenda-setting function of mass media . Public Opinion Quarterly , 36 , 176 – 187 . doi: 10.1086/267990 Google Scholar CrossRef Search ADS Mols F. , Jetten J. ( 2016 ). Explaining the appeal of populist right-wing parties in times of economic prosperity . Political Psychology , 37 , 275 – 292 . doi: 10.1111/pops.12258 Google Scholar CrossRef Search ADS Mudde C. ( 1996 ). The war of words defining the extreme right party family . West European Politics , 19 , 225 – 248 . Google Scholar CrossRef Search ADS Mudde C. ( 2007 ). Populist radical right parties in Europe . Cambridge : Cambridge University Press . Google Scholar CrossRef Search ADS Mughan A. , Paxton P. ( 2006 ). Anti-immigrant sentiment, policy preferences and populist party voting in Australia . British Journal of Political Science , 36 , 341 – 358 . doi:10.1017/S0007123406000184 Google Scholar CrossRef Search ADS Muthén L. K. , Muthén B. O. ( 1998 -2012). Mplus user’s guide ( 7th ed. ). Los Angeles, CA : Muthén & Muthén . Norris P. ( 2005 ). Radical right. Voters and parties in the electoral market . Cambridge; New York, NY : Cambridge University Press . Google Scholar CrossRef Search ADS Oegema D. , Kleinnijenhuis J. ( 2000 ). Personalization in political television news: A 13-wave survey study to assess effects of text and footage . Communications , 25 ( 1 ), 43 – 60 . doi: 10.1515/comm.2000.25.1.43 Google Scholar CrossRef Search ADS Pappi F. U. ( 1996 ). Political behavior: Reasoning voters and multi-party systems. In Goodin R. E. , Klingemann H. D. (Eds.), A new handbook of political science (pp. 255 – 275 ). Oxford : Oxford University Press . Poznyak D. , Abts K. , Swyngedouw M. ( 2011 ). The dynamics of the extreme right support: A growth curve model of the populist vote in Flanders-Belgium in 1987-2007 . Electoral Studies , 30 , 672 – 688 . doi: 10.1016/j.electstud.2011.06.011 Google Scholar CrossRef Search ADS Raftery A. ( 1995 ). Bayesian model selection in social research . Sociological Methodology , 25 , 111 – 163 . doi:10.2307/271063 Google Scholar CrossRef Search ADS Revilla M. A. , Saris W. E. ( 2013 ). A comparison of the quality of questions in a face-to-face and a web survey . International Journal of Public Opinion Research , 25 , 242 – 253 . doi:10.1093/ijpor/eds007 Google Scholar CrossRef Search ADS 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 , 411 – 424 . doi:10.1093/esr/jcn028 Google Scholar CrossRef Search ADS Rooduijn M. ( 2014 ). The mesmerising message: The diffusion of populism in public debates in Western European media . Political Studies , 62 , 726 – 744 . doi: 10.1111/1467-9248.12074 Google Scholar CrossRef Search ADS Rydgren J. ( 2005 ). Is extreme right-wing populism contagious? Explaining the emergence of a new party family . European Journal of Political Research , 44 , 413 – 437 . doi:10.1111/j.1475-6765.2005.00233.x Google Scholar CrossRef Search ADS Rydgren J. ( 2007 ). The sociology of the radical right . Annual Review of Sociology , 33 ( 1 ), 241 – 262 . doi: 10.1146/annurev.soc.33.040406.131752 Google Scholar CrossRef Search ADS Rydgren J. ( 2008 ). Immigration sceptics, xenophobes or racists? Radical right-wing voting in six West European countries . European Journal of Political Research , 47 , 737 – 765 . doi:10.1111/j.1475-6765.2008.00784.x Google Scholar CrossRef Search ADS Satorra A. , Bentler P. M. ( 2001 ). A scaled difference chi-square test statistic for moment structure analysis . Psychometrika , 66 , 507 – 514 . doi: 10.1007/BF02296192 Google Scholar CrossRef Search ADS Schafraad P. , Wester F. , Scheepers P. ( 2013 ). Media attention to the far-right in three Dutch newspapers 1986-2004. Characteristics and trends . Observatorio , 7 , 1 – 31 . Scheepers P. , Gijsberts M. , Coenders M. ( 2002 ). Ethnic exclusionism in European countries: public opposition to civil rights for legal migrants as a response to perceived ethnic threat . European Sociological Review , 18 ( 1 ), 17 – 34 . doi:10.1093/esr/18.1.17 Google Scholar CrossRef Search ADS Scherpenzeel A. C. , Das M. ( 2010 ). True longitudinal and probability-based Internet panels: Evidence from the Netherlands. In: M. Das, P. Ester, & L. Kaczmirek (Eds.), Social and Behavioral Research and the Internet: Advances in Applied Methods and Research Strategies. New York: Taylor & Francis, Routledge. Schlueter E. , Davidov E. ( 2013 ). Contextual sources of perceived group threat: Negative immigration-related news reports, immigrant group size and their interaction, Spain 1996–2007 . European Sociological Review , 29 , 179 – 191 . doi:10.1093/esr/jcr054 Google Scholar CrossRef Search ADS Schlueter E. , Meuleman B. , Davidov E. ( 2013 ). Immigrant integration policies and perceived group threat: A multilevel study of 27 Western and Eastern European countries . Social Science Research , 42 , 670 – 682 . doi:10.1016/j.ssresearch.2012.12.001 Google Scholar CrossRef Search ADS PubMed 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 . doi:10.1016/j.electstud.2012.11.003 Google Scholar CrossRef Search ADS Semetko H. , Schoenbach K. ( 1994 ). Germany’s “unity election”—Voters and the media . Cresskill, NJ : Hampton Press . Semyonov M. , Raijman R. , Gorodzeisky A. ( 2006 ). The rise of anti-foreigner sentiment in European societies, 1988-2000 . American Sociological Review , 71 , 426 – 449 . doi:10.1177/000312240607100304 Google Scholar CrossRef Search ADS Sheets P. , Bos L. , Boomgaarden H. G. ( 2016 ). Media cues and citizen support for right-wing populist parties . International Journal of Public Opinion Research , 28 , 307 – 330 . doi:10.1093/ijpor/edv014 Google Scholar CrossRef Search ADS Sherif M. , Harvey O. J. , White J. B. , Hood W. R. , Sherif C. W. ( 1961 ). Intergroup conflict and cooperation. The robbers cave experiment [1954] . Norman : University of Oklahoma . Song X. Y. , Lee S. Y. , Hser Y. I. ( 2008 ). A two-level structural equation model approach for analyzing multivariate longitudinal responses . Statistics in Medicine , 27 , 3017 – 3041 . doi: 10.1002/sim.3266 Google Scholar CrossRef Search ADS PubMed Swyngedouw M. ( 2001 ). The subjective cognitive and affective map of extreme right voters: using open-ended questions in exit polls . Electoral Studies , 20 , 217 – 241 . doi:10.1016/S0261-3794(00)00010-X Google Scholar CrossRef Search ADS Tesler M. ( 2015 ). Priming predispositions and changing policy positions: An account of when mass opinion is primed or changed . American Journal of Political Science , 59 , 806 – 824 . doi: 10.1111/ajps.12157 Google Scholar CrossRef Search ADS Van der Brug W. , Fennema M. ( 2003 ). Protest or mainstream? How the European anti-immigrant parties developed into two separate groups by 1999 . European Journal of Political Research , 42 ( 1 ), 55 – 76 . doi: 10.1111/1475-6765.00074 Google Scholar CrossRef Search ADS Van der Brug W. , Fennema M. ( 2007 ). Causes of voting for the radical right . International Journal of Public Opinion Research , 19 , 474 – 487 . doi:10.1093/ijpor/edm031 Google Scholar CrossRef Search ADS 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 .. doi:10.1111/1475-6765.00505 Van der Brug W. , Semetko H. A. , Valkenburg P. M. ( 2007 ). Media priming in a multi-party context: A controlled naturalistic study in political communication . Political Behavior , 29 , 115 – 141 . doi:10.1007/s11109-006-9020-7 Google Scholar CrossRef Search ADS Van der Meer T. W. G. , van Elsas E. , Lubbe R. , van der Brug W. ( 2013 ). Are volatile voters erratic, whimsical or seriously picky? A panel study of 58 waves into the nature of electoral volatility (The Netherlands 2006–2010) . Party Politics , 21 ( 1 ), 100 – 114 . doi:10.1177/1354068812472570 Google Scholar CrossRef Search ADS Van der Pas D. , de Vries C. , van der Brug W. ( 2011 ). A leader without a party: Exploring the relationship between Geert Wilders’ leadership performance in the media and his electoral success . Party Politics , 19 , 458 – 476 . doi: 10.1177/1354068811407579 Vliegenthart R. , Boomgaarden H. G. , Van Spanje J. ( 2012 ). Anti-immigrant party support and media visibility: A cross-party, over-time perspective . Journal of Elections, Public Opinion and Parties , 22 , 315 – 358 . doi:10.1080/17457289.2012.693933 Google Scholar CrossRef Search ADS Vossen K. ( 2010 ). Populism in The Netherlands after Fortuyn: Rita Verdonk and Geert Wilders compared . Perspectives on European Politics and Society , 11 ( 1 ), 22 – 38 . doi:10.1080/15705850903553521 Google Scholar CrossRef Search ADS Vossen K. ( 2011 ). Classifying Wilders: The ideological development of Geert Wilders and his party for freedom . Politics , 31 , 179 – 189 . doi:10.1111/j.1467-9256.2011.01417.x Google Scholar CrossRef Search ADS Walgrave S. , de Swert K. ( 2004 ). The making of the (issues of the) Vlaams Blok . Political Communication , 21 , 479 – 500 . doi:10.1080/10584600490522743 Google Scholar CrossRef Search ADS Werts H. , Scheepers P. , Lubbers M. ( 2012 ). Euro-scepticism and radical right-wing voting in Europe, 2002–2008: Social cleavages, socio-political attitudes and contextual characteristics determining voting for the radical right . European Union Politics , 14 , 183 – 205 . doi: 10.1177/1465116512469287 Google Scholar CrossRef Search ADS Wooldridge J. M. ( 2009 ). Introductory econometrics: a modern approach ( 4th ed .). Independence : South-Western . Yang J. , Stone G. ( 2003 ). The powerful role of interpersonal communication in agenda setting . Mass Communication and Society , 6 ( 1 ), 57 – 74 . doi:10.1207/S15327825MCS0601_5 Google Scholar CrossRef Search ADS Zhang Z. , Zyphur M. J. , Preacher K. J. ( 2009 ). Testing multilevel mediation using hierarchical linear models: Problems and solutions . Organizational Research Methods , 12 , 695 – 719 . doi: 10.1177/1094428108327450 Google Scholar CrossRef Search ADS Zhirkov K. ( 2013 ). Nativist but not alienated: A comparative perspective on the radical right vote in Western Europe . Party Politics , 20 , 286 – 296 . doi: 10.1177/1354068813511379 Google Scholar CrossRef Search ADS © The Author(s) 2018. Published by Oxford University Press on behalf of The World Association for Public Opinion Research. All rights reserved.

Journal

International Journal of Public Opinion ResearchOxford University Press

Published: Feb 14, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off