Start Spreading the News: A Comparative Experiment on the Effects of Populist Communication on Political Engagement in Sixteen European Countries

Abstract

Although populist communication has become pervasive throughout Europe, many important questions on its political consequences remain unanswered. First, previous research has neglected the differential effects of populist communication on the Left and Right. Second, internationally comparative studies are missing. Finally, previous research mostly studied attitudinal outcomes, neglecting behavioral effects. To address these key issues, this paper draws on a unique, extensive, and comparative experiment in sixteen European countries (N = 15,412) to test the effects of populist communication on political engagement. The findings show that anti-elitist populism has the strongest mobilizing effects, and anti-immigrant messages have the strongest demobilizing effects. Moreover, national conditions such as the level of unemployment and the electoral success of the populist Left and Right condition the impact of populist communication. These findings provide important insights into the persuasiveness of populist messages spread throughout the European continent.

Keywords

populism populist communication internationally comparative research experimental research political engagement social identity framing

Populism is highly visible and electorally successful across Europe. To explain populism’s success, the media, in particular, have been regarded as an important supply-side factor (e.g., Mazzoleni 2008). Previous studies have shown that populist media cues have an impact on blame perceptions, political cynicism, and party preferences (e.g., Bos et al. 2013; Hameleers et al. 2017). It is, thus far, unclear, however, to what extent these cues prompt political engagement, and how the effects of populist communication differ across countries and contexts. Therefore, this research tests the behavioral effects of populist communication on the Left and Right by drawing on a comparative experiment in sixteen countries.

After a profound scholarly debate on how populism should be defined, a general consensus has been reached. Populism is conceived as a set of ideas emphasizing that society is separated by the “good” ordinary people versus “the corrupt” political elites (Jagers and Walgrave 2007; Mudde 2004). Populist political communication may not only identify a vertical outgroup as opposed to the political heartland, but, oftentimes, also draws a line between the pure people and specific horizontal outgroups. Such constructions of the divide between “us” and “them” have been defined as “complete” populism (Jagers and Walgrave 2007; also see Introduction). In this paper, we discern two types of complete populists: first, right-wing populists such as the PVV in the Netherlands or the FPÖ in Austria who perceive immigrants as posing a threat to the purity of the heartland; second, left-wing populists such as Podemos in Spain who blame capitalists and the extreme wealthy for the problems of common hardworking citizens.

In the past few years, a number of studies have attempted to disentangle the effect of populist communication on political attitudes (e.g., Bos et al. 2013; Hameleers et al. 2017; Matthes and Schmuck 2017). Although these studies have provided important foundational evidence of the effects of populist communication, three general shortcomings can be identified. First, they focus on only a subset of populist components, either focusing on right-wing exclusionist populism or populism’s core anti-elitist ideology (e.g., Bos et al. 2013; Matthes and Schmuck 2017). Second, these studies are mostly conducted in one single country (e.g., Hameleers et al. 2017). Because internationally comparative studies are missing, it is unclear how well these results travel to other countries with different real-life opportunity structures for populist communication to root (see Aalberg et al. 2017 for discussion). Finally, previous studies mainly focus on attitudinal consequences of populist exposure (e.g., Hameleers et al. 2017), neglecting behavioral outcomes (or intentions).

To move forward within this research field, we have conducted an extensive sixteen-country experiment (N = 15,530). In this experiment, the central components of empty, anti-elites, anti-outgroups on the Left and Right, complete left-wing and complete right-wing populism have been manipulated and contrasted with two control groups. As key dependent variable, we investigated how these populist messages prime political engagement. The general expectation was that populist messages, by means of social identity framing, motivate people to engage politically (Mols 2012; Van Zomeren et al. 2008). Specifically, populist rhetoric primes and augments social identity by pitting “blameless people” against the “culprit elites” and other outgroups. By enhancing feelings of injustice and deprivation while offering credible scapegoats, these social identity frames are assumed to be effective in priming political engagement (e.g., Van Zomeren et al. 2008).

In the context of a salient European issue, declining consumer purchasing power, we found that the effects of populist communication were contingent upon contextual opportunity structures in the various countries. The findings of this study offer important empirical evidence for the behavioral effects of different types of left-wing and right-wing populist messages in a multicountry setting. Taken together, this study is the first to provide comprehensive insights into how populist messages affect political engagement in a diversified European context.

The Effects of Populist Communication on Political Engagement

Populist political communication reduces complex and technocratic political issues to an all-encompassing binary divide in society: The “good” ordinary people are constructed in opposition to the “evil” and “corrupt” elites, potentially supplemented by the exclusion of outgroups on the Left and Right (e.g., Mudde 2004; also see Introduction and Online Appendix A for typology and definition of populist communication). It has been argued that this simplified discourse can be very persuasive (e.g., Rooduijn 2014). Indeed, experimental studies have demonstrated that populist messages activate or prime citizens’ political perceptions (e.g., Hameleers et al. 2017; Matthes and Schmuck 2017).

The processes of inclusion and exclusion central to populist communication can be interpreted in light of social identity theory. According to social identity theory, individuals can express and experience belonging to a variety of different selves (Tajfel 1978; Tajfel and Turner 1986). These self-concepts are triggered differentially depending on the social context. In populist discourse, the social context can be understood as a sense of perceived deprivation in a situation constructed as a crisis or ingroup threat (Elchardus and Spruyt 2016; Taggart 2000). In this crisis, the people are envisioned as a homogeneous entity whose will is no longer represented by the elites. The elites have failed to represent their “own” people, and have, instead, taken care of themselves and the “undeserving” superrich and/or migrants. Populist identity constructions, thus, emphasize a threat to the ingroup of the people, who are perceived as relatively worse off than other groups in society (Elchardus and Spruyt 2016).

As a response to this perception of ingroup deprivation, collective action can be triggered (e.g., Van Zomeren et al. 2008). This means that people are mobilized when they perceive that their ingroup is disadvantaged by an outgroup (e.g., Simon and Klandermans 2001). In this paper, we regard political engagement as a form of collective action. Political engagement can be defined as political acts initiated by citizens, acting on behalf of their ingroup (e.g., Bimber 2001). Political engagement may involve different acts, such as using the Internet to search for political information, expressing views on the functioning of the government, or signing an (online) petition (e.g., Bimber 2001). Citizens can, thus, engage politically on different levels, using different (social) media channels.

In this study, we regard political engagement as a consequence of exposure to populist messages. In line with the premises of social identity framing, political messages that emphasize ingroup threats and external causes should promote political engagement, for example, in the form of voicing opposition to the ruling elites (Bimber 2001). Political engagement can, thus, be regarded as a strategy to deal with deprivation and threat, for example, to renegotiate a severe power discrepancy between the ingroup and others (e.g., Tajfel 1978). Such a threat can, for example, be caused by the perception of being treated unfairly by the political elites (e.g., Van Zomeren et al. 2008). Importantly, a politicized identity has been found to result in more obligation to take action (e.g., Simon and Klandermans 2001) and, therefore, has stronger effects on political engagement than identification with the disadvantaged group in more general terms.

Because mobilization of the ingroup is not necessarily determined by the mere belonging to existing ingroups, the framing of identity is crucial for the promotion of political engagement (e.g., Polletta and Jasper 2001). Specifically, in populist communication, the ingroup of the people is framed as a politicized identity consisting of the silenced majority of ordinary citizens (Caiani and della Porta 2011). Populist communication does not only appeal to this ingroup, but brings “the people” into being by actively constructing this identity through communication (e.g., Laclau 2005; Moffitt and Tormey 2014). Anti-elites populism and complete populist communication explicitly emphasize that this ingroup is threatened: Either the elites or societal outgroups are depriving the people from what they deserve (e.g., Elchardus and Spruyt 2016). As a response to this ingroup threat, receivers of populist identity framing are expected to be mobilized to engage politically. In other words, populist communication may promote political engagement by priming the perception of a politicized and relatively deprived ingroup opposed to a threatening outgroup. Applied to the aforementioned typology of populist communication (also see Introduction), we first of all expect that “empty” populist communication has a positive effect on political engagement, so that

Hypothesis 1a (H1a): Messages with people centrality cues lead to more political engagement than messages that only include a factual description of the issue at hand.

Persuasive identity frames not only emphasize a severe sense of injustice and collective agency, they also mark the boundary between the victimized ingroup and responsible outgroups (e.g., Polletta and Jasper 2001). In line with this, research on collective action framing found that the cultivation of a salient ingroup threat (also termed as injustice framing) is not sufficient to engage the ingroup politically. Rather, it is crucial to emphasize that there is a target to blame for the people’s injustice (e.g., Gamson 1992). The motivation to engage politically may further be strengthened by efficacy beliefs: The deprived ingroup should also be offered treatments to remedy their deprivation.

These theoretical premises can be extrapolated to the conceptualization of fuller types of populist communication (Jagers and Walgrave 2007). Anti-elite populist cues do not only cultivate feelings of injustice or ingroup deprivation, but also offer a credible and visible external scapegoat. These populist cues further emphasize that only if the culpable other is removed, the ingroup’s crisis may be averted. This connects to the literature that argues how the framing of efficacy, responsibility, and injustice may result in the strongest motivation to take action (e.g., Gamson 1992). Against this backdrop, we hypothesize that

Hypothesis 1b (H1b): Messages with antipolitical elite populist cues lead to more political engagement than neutrally framed messages without populist cues.

In complete populism, the ingroup threat is framed as even more encompassing than in anti-elites populism: The elites are not only failing to represent the people, they are also held responsible for prioritizing societal outgroups that deprive the people (Elchardus and Spruyt 2016). Specifically, left-wing and right-wing complete populism highlight an even stronger sense of urgency for citizens to engage and spread the word about the failing elites and the “dangerous” societal outgroups. In line with the premises of collective action framing, ingroup deprivation is paired with the presentation of multiple scapegoats that threaten the people from above (the elites) and within (the others). The perception of deprivation is consequentially augmented in “complete” populism: The elites are blamed for failing to represent the people, and societal outgroups are, in addition, seen as fierce competition for the people’s resources. Taken together, the social identities in complete populism are more politicized, as they resonate with specific left-wing (opposing the superrich) and right-wing (opposing immigrants) frames (Mols 2012). Complete populism, thus, identifies different scapegoats that can be held accountable for the injustice facing the people’s ingroup (Gamson 1992). These populist identity frames strengthen identification with a collective “movement identity” that is connected by a shared commitment to exclude the culpable other (Klandermans 1977). Against this backdrop, we hypothesize that

Hypothesis 2 (H2): Complete left-wing and complete right-wing populism result in more political engagement than “merely” anti-elitist populism.

We, thus, expect that messages that combine anti-elitist populist cues with left-wing and right-wing outgroup cues result in higher levels of political engagement than messages with only anti-elitist populist cues. These populist cues are contrasted with people-centrality messages that do not contain blame attribution to the elites and/or other outgroups.

Contextual Differences and Opportunity Structures for Populist Communication

The comparative setup of this experiment allows us to assess how real-life external supply-side conditions may impact the effects of populist communication (Stanyer et al. 2017). Indeed, it has been argued that populist communication is highly chameleonic, adjusting itself to different constructions of a threat to the people (Mazzoleni 2008). This implies that the stickiness of populist communication depends on the discursive opportunity structures provided by the context (Aalberg et al. 2017). In other words, the persuasiveness of populist political communication in Europe may depend on the extent to which it resonates with external supply-side factors (Stanyer et al. 2017).

We assume that a key contextual factor for the countries investigated is the European economic crisis, resulting in austerity measures introduced by various national governments. This may, in particular, be an important factor to explain the persuasiveness of populism in Southern European countries. Indeed, Spain, Portugal, Italy, Greece, and Romania have, in particular, suffered from the consequences of harsh recessions and rising levels of unemployment. In line with the premises of injustice and collective action framing, populist cues that blame outgroups resonate stronger in the hard-hit countries (Gamson 1992). In the aftermath of the recession, ingroup deprivation and outgroup blame should be seen as most credible in countries that have actually lost out relatively more than other countries. Therefore, we hypothesize,

Hypothesis 3a (H3a): The higher the level of unemployment in a country, the more populist cues lead to political engagement.

Another important supply-side factor related to the attractiveness of right-wing populism is the rise of immigration within and outside Europe (Stanyer et al. 2017). Western and Nordic European countries that have witnessed a stronger influx of refugees and immigrants may perceive the outgroup construction of immigrants in right-wing populist communication as more real and severe. Hence, we argue that immigrants pose a credible scapegoat for national issues when they are present in great numbers. Right-wing populist communication is, therefore, expected to resonate most saliently in European countries with higher levels of immigration. In these countries, the discursive opportunity structure is consolidated by stronger fears of relative deprivation due to the influx of people from abroad. Against this backdrop, we hypothesize,

Hypothesis 3b (H3b): The higher the level of immigration in a country, the more right-wing populist cues lead to political engagement.

Finally, we want to know whether individual effects of populist messages are also affected by a country’s general political climate and discourse structure. It could be assumed that the frequency in which citizens are confronted with populist arguments may foster their effects because citizens have gotten used to them and, thus, perceive them as more legitimate. Alternatively, it could be argued that citizens already know all populist arguments in countries where populism is successful, which could result in smaller effects of yet another populist appeal. We, therefore, ask whether the public visibility of populist messages in a country has an impact upon the effects of populist messages (Research Question 1; RQ1) and use the success of right-wing and left-wing populist parties in the different countries as a proxy for the public prevalence of populist messages.

Method

Experimental Design

In all sixteen countries, the design of the experiment was identical. It was developed by a research team in which all countries were represented by nationals of that country. The setup was a three × two between-subjects experiment with two control groups in which we investigated the differential impact of a focus on the national ingroup and of blame attributed to horizontal and vertical outgroups (see Table 1). The topic was held constant in all conditions and concerned a prediction of a future decrease of purchasing power for the respective citizens of the sixteen countries. The source of the populist messages was also held constant: A representative of a fictional foundation explained the reasons and responsibilities for the predicted development.

Table 1.

Overview of the Experimental Design.

	Blame on Outgroup	Blame on Political Elite
	Blame on Outgroup	No	Yes
People centrism
Yes(populism)	No	(1) emptypopulism	(2) anti politicalelite populism
	On immigrants	(3) right-wing exclusionist populism	(4) right-wingcomplete populism
	On the rich	(5) left-wing exclusionist populism	(6) left-wingcomplete populism
No(no populism)	No	(7) control 1:factual story	(8) control 2:antipolitical elite

The stimuli and questionnaire were extensively pretested using convenience samples in two countries, which were selected based on diversity criteria regarding right-wing versus left-wing populism and Western versus Southern Europe. For this reason, the pretest was conducted in Germany (N = 264) and Greece (N = 1,565). Based on the outcomes of the pilot studies, the stimuli and questionnaire were further improved to increase their credibility irrespective of contextual differences between countries.

Sample

This experiment is based on a diverse sample of citizens in sixteen countries. The countries were chosen because they represent a wide range of European contexts that differ in their political, economic, and social situations as well as in their history and electoral success of populism on the Left and Right. Countries (N_Total = 17,597) included were Austria (N = 1,138), France (N = 1,192), Germany (N = 991), Greece (N = 1,116), Ireland (N = 951), Israel (N = 1,016), Italy (N = 1,056), the Netherlands (N = 934), Poland (N = 1,368), Portugal (N = 1,048), Spain (N = 1,010), Sweden (N = 1,063), Switzerland (N = 1,134), United Kingdom (N = 1,103), Norway (N = 1,009), and Romania (N = 1,468). The data were collected in the first months of 2017 by two international research organizations, which were thoroughly instructed to apply the same procedures regarding recruiting, sampling, presentation of the survey, and data collection. Respondents were participants of online panels. National quota were applied for gender, age, and education based on official national data (e.g., micro census). In most cases, differences between quota specifications and the quota realized were small (see Online Appendix B). Before the analysis, 2,185 inattentive respondents were removed to guarantee the quality of our data (see Online Appendix C), leading to a total of 15,412 respondents.¹ The complete data set constitutes a diverse sample of European citizens with regard to age (M = 45.91, SD = 15.19), gender (M = 0.50, SD = 0.50), education² (M = 2.25, SD = 0.70), political interest³ (M = 4.67, SD = 1.69), and ideology⁴ (M = 5.05, SD = 2.55; see Online Appendix B for an overview per country).

Procedure

All sixteen experiments were conducted online. After giving their informed consent, participants filled out a pretest consisting of demographics, moderator variables, and control variables. In the next step, participants were randomly assigned to one of the eight conditions and read an online news item, which was visible for at least twenty seconds. A randomization check shows that the eight conditions do differ significantly with regard to age (F_{7, 15243} = 2.10, p = .04), but not with regard to gender (F_{7, 15392} = 0.15, p =.99), education (F_{7, 15351} = 1.33, p = .23), political interest (F_{7, 15397} = 1.70, p = .10), and ideology (F_{7, 13894} = 1.14, p = .34). Finally, participants had to complete a posttest survey measuring the dependent variables and manipulation checks. Once completed, participants were debriefed and thanked for their cooperation. Participants received a financial incentive from the panel agencies.

Stimuli

The questionnaire and stimulus materials were first developed in an English mother version. These templates were translated by native speakers in all countries. Inconsistencies and cultural-sensitive translations and meanings were exhaustively discussed with all country members until complete agreement was achieved. The basic stimulus material in all countries and conditions consisted of a news item on an online fictional news outlet called “news.” The layout of the stimuli was based on Euronews, as this is an equally familiar template in all countries. The post consisted of an image with a wallet and a hand, signifying the topic of purchasing power (see Online Appendix D for all stimuli). This topic was chosen as it was found to be realistic in countries with varying levels of economic recovery. Next to this, the issue can credibly be framed in light of an opposition between the deprived ingroup and culprit outgroups as a threat—both on the left and right wing. In all conditions, a fictional foundation called FutureNow was the source of the (populist) message. We aimed for equivalence between all conditions. This means that all treatment conditions were based on exactly the same storyline, and that we only varied the specific populist interpretation of the issue between conditions. This, for example, means that the factual control and the anti-elites populist condition only differed to the extent that the populist condition framed the issue as being caused by the corrupt elites that threaten the ingroup of the ordinary people.

Specifically, in the six treatment conditions, the typology of populist communication as outlined in the theoretical framework was manipulated (also see Introduction). In these conditions, the national ingroup was framed as a victim of the situation (condition 1) (see Table 1). The anti-elitist populist condition (condition 2) explicitly blamed the self-interested politicians for depriving the ordinary people. In the right-wing exclusionist condition (condition 3), immigrants were blamed for taking away resources from the hardworking native people. The left-wing exclusionist condition (condition 4) shifted blame to the extreme rich, who were depicted as self-serving, corrupt, and egoistic. The right-wing complete condition (condition 5) emphasized the vertical opposition between the good ordinary citizens and the corrupt elites and added an exclusionist component by stressing that immigrants were also responsible for the people’s crisis. In addition, the elites were held responsible for only representing the needs of migrants, instead of their own people. Finally, in the complete left-wing populist condition (condition 6), the elites and extreme rich were blamed for taking away the people’s resources and only looking after themselves. Again, the elites were accused of prioritizing the needs of the outgroup of the extreme rich instead of the hardworking ordinary people. Two control conditions were added. The first control condition (condition 7) entailed a neutrally framed article on declining purchasing power, focusing on the facts of the development only without people centrality cues and without blaming any group as responsible. The second control condition (condition 8) added only the political elite as a vertical outgroup, blaming them for the expected development.

Measures

Political engagement

Our dependent variable—political engagement—was measured with three items, tapping the willingness of the respondent to (1) share the news article on social network sites, (2) talk to a friend about the article, and (3) sign an online petition to support the nongovernmental organization (NGO) mentioned in the article (7-point scale, running from 1 = very unwilling to 7 = very willing). The items are based on the conceptualization of political engagement in information settings (e.g., Bennett 2008; Bimber 2001). They intend to measure different behavioral intentions for political acts related to the stimuli, in both offline and online contexts. Informed by theory on social identity framing, they measure the extent to which members of the “ordinary people” engage politically with the article on behalf of their ingroup. In line with collective action framing, these items measure the extent to which citizens become politically activated after exposure to messages that emphasize ingroup deprivation and outgroup responsibility (Gamson 1992). A principal component analysis showed that all three items load on the same factor, with loadings varying from .83 to .88. We consequentially computed a mean scale of political engagement (Cronbach’s α = .83, M = 3.88, SD = 1.74).

Contextual variables

Unemployment rates per country were retrieved from the ILOSTAT database of the International Labour Organization (World Bank 2016) and immigration rates from Eurostat (2015). The size of populist parties on the Left and Right was established on the basis of the vote share in the last national elections in each country.⁵ An overview can be found in Online Appendix E. The contextual variables were centered around their mean to aid interpretation of the cross-level interaction terms.

Manipulation Checks

After being exposed to the stimulus material and the posttest measures, participants were subject to five manipulation checks. F tests indicate that the populism conditions significantly differ from the control groups with regard to the extent the story described (1) the people of the country as hardworking, F(1, 15157) = 1070.27, p = .00; (2) a situation in which the national citizens will be affected by the economic developments described, F(1, 15192) = 58.98, p = .00; and (3) a threat to the well-being of the people, F(1, 15182) = 122.97, p = .00. In addition, the anti-elitist conditions differ significantly from the other conditions in the extent to which they ascribe responsibility for the predicted decline of purchasing power to politicians, F(1, 15171) = 2182.37, p = 0.00; the anti-immigrant conditions differ significantly from the other conditions in the extent to which they ascribe responsibility for the decline of purchasing power to immigrants, F(1, 15162) = 5079.79, p = 0.00; and the antirich conditions differ significantly from the other conditions in the extent to which they ascribe responsibility to the rich, F(1, 15168) = 2782.33, p = .00.

Analyses

Because our data set consists of samples in sixteen different countries, it has a hierarchical structure: Observations are nested within countries. To test our hypotheses in all country samples simultaneously and to control for the dependency of the observations, we ran multilevel (mixed-effects) models in Stata. This allows us to test the impact of explanatory variables at the level of the individual respondent as well as at the country level, on the response variable measured at the lowest level (Hox et al. 2010). An estimation of the empty model (with controls) shows that this model significantly outperforms a multiple regression model: Likelihood Ratio (LR) test = 1543.98, p < .001. The intraclass correlation coefficient is 0.096, which shows that more than 9 percent of the variability in political engagement is due to the country level. Yet, within-country differences are substantially larger than between-country differences.

Results

We, first of all, investigated whether messages with people centrality cues or anti-elitist cues increase political engagement compared with messages without such cues⁶ (full models in Online Appendix Tables). The first panel of Figure 1 shows that both aspects of populist rhetoric separately have no mobilizing effects. This means that H1a is not supported. People centrality cues on their own do not prime political engagement.

Figure 1.

The impact of people centrality and anti-elite cues on political engagement.

In the next step, we tested whether antipolitical elite populist messages (messages with people centrality and anti-elitist cues combined) have a stronger impact on populist engagement than messages without such populist cues (H1b). To do so, we estimated the interaction between people centrality and anti-elite populist cues. An LR test shows that this interaction significantly improves the model (see Online Appendix Table A1). In line with this, the second panel of Figure 1 illustrates that antipolitical elite populist messages that combine people centrality and anti-elite cues have a stronger mobilizing impact compared with messages without (any of) these cues. In addition, the interaction plots in Figure 2 show that these messages are more mobilizing than messages including only people centrality cues. H1b, therefore, finds support in these data.

Figure 2.

The impact of anti-elite cue on political engagement conditioned by people centrality cue presence.

To test our second hypothesis, we investigated the impact of messages including references to the ingroup combined with blame attribution to vertical and horizontal outgroups (see Online Appendix Table A2). We compare these complete populist messages with a message targeted only at the ingroup (i.e., people centrality cues).⁷ Figure 3 shows the results, and they are quite clear-cut. Only messages contrasting the ingroup to the immigrant outgroup have an impact, and it is a negative one. Combining an anti-elitist cue with an anti-immigrant cue (the first panel) or an antirich cue (the second panel) does not improve the model (also indicated by the LR tests). In other words, H2 cannot be supported: Complete left-wing and complete right-wing populism do not result in more political engagement than “mere” anti-elitist populism.

Figure 3.

The impact of right-wing and left-wing complete populism on political engagement.

To test the third hypothesis, level 2 (country-level) variables were added to the model. We, first of all, investigated whether the level of unemployment of a country affects the extent to which populist cues lead to political engagement (H3a; Online Appendix Table A3). Adding interaction terms between message cues and unemployment levels improves the model considerably. While the fixed regression coefficients indicate that only the impact of an anti-immigrant message is moderated by the nation’s level of unemployment, results are different when we plot the moderated effects in Figure 4. There, it is shown that the impact of the anti-elite cue is only positive and significant in countries with an (above) average level of unemployment, and that the negative impact of the anti-immigrant cue increases when unemployment levels rise. The impact of complete right-wing or left-wing messages is not specified by levels of unemployment (as indicated by the nonsignificant LR test). Thus, H3a is only supported with regard to anti-elite populist messages.

Figure 4.

The moderating impact of unemployment level.

The moderating impact of the level of immigration is subsequently tested in the models in Online Appendix Table A4. Our analyses clearly show that the level of immigration of a country does not moderate the impact of populist cues on political engagement. This does not provide support for H3b. Based on these findings, it can be concluded that only the level of unemployment moderates the impact of populist cues on political engagement. Specifically, the higher the level of unemployment in a country, the more an anti-immigrant cue decreases, and an anti-elite cue enhances political engagement.

Finally, we tested whether the impact of populist cues on political engagement depended on the success of right- and left-wing populist actors in the country (RQ1). The results are shown in Online Appendix Table A5. Our analyses show that the more successful the populist Right is in a country, the more negative the impact of anti-elitism, and the more positive the impact of the anti-immigrant message on political engagement. These results are plotted in the top panel of Figure 5, which indicates that messages with an anti-elite cue only have a positive impact in countries with (below) average populist Right success. In addition, it shows that the impact of the anti-immigrant cue is more positive in countries with more successful populist Right parties, to such an extent that the negative impact is absent in countries with large populist radical Right parties (such as Switzerland and Poland). There are, again, no significant three-way interactions: Complete populist messages are not moderated by right-wing populist success. The bottom panels of Figure 5 show the moderating impact of a country’s left-wing populist success. The positive impact of a message with an anti-elite cue is higher in countries with a stronger populist party, and insignificant in countries with no such party having electoral success. The negative impact of the anti-immigrant message is larger in countries with a stronger populist Left. The impact of the antirich cue is insignificant regardless of the presence of neither the populist Left nor Right.

Figure 5.

The moderating impact of populist vote share.

To answer RQ1, the supply-side of populist success in a country partially resonates in the persuasiveness of populist cues: Anti-immigrant messages demobilize voters more in countries with a more successful populist Left, and demobilize voters less when they resonate with the electoral success of the populist Right. Anti-elite messages, in contrast, only have mobilizing effects in countries with a less successful populist Right party, while they mobilize more voters in countries with a stronger populist Left.

Discussion

Despite the growing salience of populism, the effects of populist messages have only recently become an issue of greater interest to political communication scholars (see, for example, Bos et al. 2013; Hameleers et al. 2017; Matthes and Schmuck 2017). This study set out to address some of the most important open questions of prior research by (1) investigating the effects of different combinations of populist message elements, (2) comparing the effects of populist messages in countries with different supply-side conditions, and (3) focusing on behavioral outcomes. To do so, a large-scale internationally comparative experiment on the effects of populist messages was conducted in sixteen countries.

Our key findings suggest that people centrism or anti-elite cues on their own do not significantly promote or weaken political engagement. Instead, it is their specific combination that makes populist cues effective. In line with theories on social identity framing, the priming of a social identity (the ordinary people) and a salient ingroup threat (the elites) make populist communication persuasive (e.g., Mols 2012; Simon and Klandermans 2001). This supports literature on collective action and injustice framing (Gamson 1992). In line with this literature, social identity frames are persuasive when they cultivate a strong sense of ingroup threat while foregrounding credible scapegoats deemed responsible for the ingroup’s deprivation. Against this backdrop, the success of populist appeals with respect to political engagement is contingent on both the induction of a personally relevant threat and the presentation of an outgroup that can be credibly blamed. This supports the theoretical notion that the key to explaining the success of populist appeals is to look at the combination of message elements, and not just at the isolated prevalence of individual elements (e.g., Aalberg et al. 2017).

However, using complete forms of right- and left-wing populism by adding immigrants or the rich as culpable outgroups did not increase the level of political engagement. An anti-immigrant cue even produced the opposite effect: It reduced the likelihood that people would spread the article and become engaged. Complete populism, as conceptualized in extant literature (e.g., Jagers and Walgrave 2007), is, thus, not necessarily more effective in promoting political engagement. This finding can be explained based on the theoretical premises of injustice and collective action framing (Gamson 1992; Klandermans 1977). Specifically, the fit of the issue of decreasing purchasing power with responsibility attributions to nonelitist actors may not be seen as credible, and citizens may consequentially not be provided with fitting strategies to avert the threat to their ingroup (e.g., Mols 2012; Van Zomeren et al. 2008). After all, fighting a decline of purchase power might be seen as a problem that should primarily be handled by politicians. An important implication is that populist communication should be most effective when the topic, the framing of ingroup deprivation, and blame attributions all align—and provide a credible story for citizens to restore a subjective sense of ingroup injustice.

As theorized in extant literature, contextual factors have an impact on the effects of populism (e.g., Stanyer et al. 2017). Interestingly, whereas the level of immigration does not make a difference, the activating potential of specific message cues varies with the national level of unemployment. The higher the level of unemployment, the stronger the motivating effect of anti-elite cues and the stronger the demotivating effect of anti-immigrant cues. This may again be explained in the light of collective action framing: Citizens need to be offered a credible scapegoat in the various national settings (e.g., Klandermans 1977). Politicians are the most salient actors entrusted with decision-making power in economic policies, and should, thus, be held accountable. Therefore, higher unemployment makes people more susceptible to blame directed at politicians. This is in line with research on the augmenting role of ingroup deprivation on political engagement (Simon and Klandermans 2001). Extending this mechanism, our study demonstrates that when the threat to the ordinary people becomes more severe in the light of a country’s contextual factors, people become more motivated to restore “their” ingroup’s status by engaging politically.

Finally, our study shows that the overall political climate makes a difference for the success of populist cues. The success of anti-elite and anti-immigrant cues is contingent on the success of left- and right-wing populist parties in a country. In line with the theoretical notion that political engagement is primed when the ingroup threat is politicized (e.g., Simon and Klandermans 2001), our results indicate that the salience of populism on the Left and Right augments the effects of populist communication. The relationship between the strength of right-wing populist parties and populist cues may be explained by the fact that there are several countries in which right-wing populists are very successful or even part of national or regional governments (e.g., Poland, Switzerland, Austria, Norway). This makes right-wing anti-elite cues less convincing, but anti-immigrant cues, which should be especially accepted in these countries, not as unconvincing as in other countries. The results for the contextual effects of left-wing populist success support this line of argumentation. A stronger left-wing populist climate increases the already negative effect of socially less accepted anti-immigrant cues. Populist communication, thus, resonates with the opportunity structures in countries.

This study has several limitations. First, we only tested the impact of populist messages using one specific economic topic and one specific fictional source. It could be argued that the effects of messages would be even stronger for issues typically owned by left- or right-wing populist parties. For example, migrant-initiated crimes or corruption among bankers as topics might produce stronger effects on political mobilization because the issue and blame cue fit better together. Second, the samples in the various countries were varied but at least in some countries, survey companies had a hard time reaching low-education and income groups in their online panels. Further analyses will need to take this into account. Third, this paper focused on main and interaction effects of populist cues and between-country differences based on selected contextual factors. Further analyses will need to look at the moderating and mediating influences of additional individual-level variables, shedding more light on the underlying mechanisms that drive the effects of populist cues on behavioral outcomes. Despite these limitations, this study shows how combining populist message elements and contextual country factors can help to better understand what makes populism so successful across Europe.

Supplemental Material

ONLINE_APPENDIX_(1) – Supplemental material for Start Spreading the News: A Comparative Experiment on the Effects of Populist Communication on Political Engagement in Sixteen European Countries

Supplemental material, ONLINE_APPENDIX_(1) for Start Spreading the News: A Comparative Experiment on the Effects of Populist Communication on Political Engagement in Sixteen European Countries by Michael Hameleers, Linda Bos, Nayla Fawzi, Carsten Reinemann, Ioannis Andreadis, Nicoleta Corbu, Christian Schemer, Anne Schulz, Tamir Shaefer, Toril Aalberg, Sofia Axelsson, Rosa Berganza, Cristina Cremonesi, Stefan Dahlberg, Claes H. de Vreese, Agnieszka Hess, Evangelia Kartsounidou, Dominika Kasprowicz, Joerg Matthes, Elena Negrea-Busuioc, Signe Ringdal, Susana Salgado, Karen Sanders, Desirée Schmuck, Jesper Stromback, Jane Suiter, Hajo Boomgaarden, Keren Tenenboim-Weinblatt and Naama Weiss-Yaniv in The International Journal of Press/Politics

Footnotes

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The data collection of Poland was supported by the National Science Center, Poland, Grant No. 2015/18/M/HS5/00080. The data collection of Romania was supported by the Doctoral School within the National University of Political Studies and Public Administration. The data collection of the Netherlands was supported by the Amsterdam School of Communication Research. The data collection of Norway was supported by the Department of Sociology and Political Science at the Norwegian University of Science and Technology. The data collection of Ireland was supported by the Faculty of Humanities and Social Sciences at Dublin City University. The data collection of Portugal was supported by the Portuguese Foundation for Science and Technology (FCT): IF/01451/2014/CP1239/CT0004.

Notes

ORCID iD

Ioannis Andreadis

Supplemental Material

Supplemental material for this article is available online.

Author Biographies

Michael Hameleers is an assistant professor in Political Communication at the Amsterdam School of Communication Research (ASCoR) at the University of Amsterdam, The Netherlands. His research interests include populism and the media, framing, and the role of identity in media effects.

Linda Bos is an assistant professor in Political Communication at the Amsterdam School of Communication Research (ASCoR) at the University of Amsterdam, The Netherlands. Her research interests include electoral behavior, quantitative research methods, (right-wing) populism, and political communication.

Nayla Fawzi is an assistant professor at the Department of Communication Studies and Media Research (IfKW) at Ludwig-Maximilians-University Munich, Germany. Her research interests include populism, politics-media relations, media literacy, and cyber-mobbing.

Carsten Reinemann is a professor and chair of Political Communication at the Department of Communication Studies and Media Research (IfKW) at Ludwig-Maximilians-University Munich, Germany. His research interests include populism, extremism, political journalism, and media effects.

Ioannis Andreadis is an associate professor in the School of Political Science at the Aristotle University of Thessaloniki, Greece. He is interested in public opinion, voting, quantitative methods, and web-based research.

Nicoleta Corbu is a professor in the College of Communication and Public Relations (SNSPA) at the University of Bucharest, Romania. Her interests include European identity, framing, education policies, and research methods.

Christian Schemer is a professor of Communication at the Institut für Publizistik, Johannes Gutenberg University Mainz, Germany. His research focuses on media use and effects, political persuasion, advertising research and methods.

Anne Schulz is a researcher at the University of Institute of Mass Communication and Media Research at the University of Zurich, Switzerland. Her research interest include populism and media effects.

Tamir Sheafer is a professor in the Departments of Political Science and Communication and is the dean of the Faculty of Social Sciences at the Hebrew University of Jerusalem, Israel. His research focuses on personalization, political values, and media effects in elections.

Toril Aalberg is a professor and Head of the Department of Sociology and Political Science at NTNU in Trondheim, Norway. Her research interests include comparative politics, election campaigns, media effects, public opinion, and the relationship between media and politics.

Sofia Axelsson is Deputy Chief Analyst in the Department of Political Science at the University of Gothenburg, Sweden.

Rosa Berganza is professor and chair of Political Communication at the Department of Communication Sciences 2 at the University of Rey Juan Carlos Madrid, Spain. Her research focuses on comparative studies of media and journalism, election coverage and campaigning, media effects, public opinion and women and the media.

Cristina Cremonesi is a researcher in the Department of Cultures, Politics, and Society at the University of Turin, Italy. She is working on populism, political communication, and public opinion.

Stefan Dahlberg is a senior lecturer in the Department of Political Science at the University of Gothenburg, Sweden. His research focuses on representative democracy, democratic legitimacy, political parties and voting behavior.

Claes H. De Vreese is a professor and chair of Political Communication in the Amsterdam School of Communication Research (ASCoR) at the University of Amsterdam, The Netherlands. He chairs the Social Science Council of the Royal Academy of Sciences (KNAW) and is the editor of Political Communication.

Agnieszka Hess is a senior researcher in the Institute of Journalism, Media and Social Communication at the Jagiellonian University in Krakow, Poland. Her research interests include populism, mediatization, and the role of NGOs in democracy.

Evangelia Kartsounidou is a PhD candidate at the School of Political Sciences of the Aristotle University of Thessaloniki, Greece. Her research interests include stances and attitudes of political elites and web-based research methods.

Dominika Kasprowicz is a researcher in the Institute of Journalism, Media and Social Communication at the Jagiellonian University in Krakow, Poland. Her research interests include populism, strategic party communication, and extreme right-wing parties.

Jörg Matthes is professor of Communication Science in the Department of Communication at the University of Vienna, Austria. His research focuses on advertising effects, the process of public opinion formation, news framing, and empirical methods.

Elena Negrea-Busuioc is a researcher at the College of Communication and Public Relations (SNSPA) at the University of Bucharest, Romania. Her academic interests include European Communication and discourse analysis.

Signe Ringdal is a PhD candidate at the Department of Sociology and Political Science at NTNU in Trondheim, Norway. Her research interest include populist parties and communication.

Susana Salgado is a professor of Political Communication and Media and Politics in the Institute of Social Sciences at the University of Lisbon, Portugal. Her research interests include comparative media studies, media and democratization, media and elections, and Internet and politics.

Karen Sanders is a professor and Head of the School of Arts and Humanities at St Mary’s University, Twickenham, United Kingdom. She is interested in public sector communication, populist communication, and communication ethics.

Desirée Schmuck is a post-doctoral researcher in the Department of Communication at the University of Vienna, Austria. Her research interests include right-wing populism and political advertising, the effects of political communication on (young) people’s attitudes as well as on green advertising.

Jesper Strömbäck is a professor in Journalism and Political Communication in the Department of Journalism, Media and Communication at the University of Gothenburg, Sweden. His research interests include public opinion, political news journalism and political communication.

Jane Suiter is director of the Institute for Future Media and Journalism (FuJo) and Associate Professor in the School of Communications at Dublin City University, Irleand. Her research interests focus on political communication and the media (traditional and new).

Hajo Boomgarden is a professor for Empirical Social Science Methods in the Department of Communication at the University of Vienna, Austria. His research interests focus on media coverage of politics, media effects on political attitudes and behaviours and advances of content analysis.

Keren Tenenboim-Weinblatt is a professor and senior lecturer in the Department of Communication and Journalism at The Hebrew University of Jerusalem, Israel. Her research focuses on cultural and political dimensions of journalism from a comparative perspective, media and conflict, mediated memory, and the intersection of journalism, political communication and popular culture.

Naama Weiss-Yaniv is a PhD candidate in the Department of Communication and Journalism at The Hebrew University of Jerusalem, Israel. Her research interest include populist parties and communication.

References

Aalberg

Toril

Esser

Frank

Reinemann

Carsten

Strömbäck

Jesper

de Vreese

Claes H.

, eds. 2017. Populist Political Communication in Europe. London: Routledge.

Bennett

W. Lance

, ed. 2008. Civic Life Online: Learning How Digital Media Can Engage Youth. Cambridge: MIT Press.

Bimber

Bruce

. 2001. “Information and Political Engagement in America: The Search for Effects of Information Technology at the Individual Level.” Political Research Quarterly 54 (1): 53–67.

Bos

Linda

van der Brug

Wouter

de Vreese

Claes H.

2013. “An Experimental Test of the Impact of Style and Rhetoric on the Perception of Right-Wing Populist and Mainstream Party Leaders.” Acta Politica 48 (2): 192–208. doi:10.1057/ap.2012.27.

Caiani

Manuela

della Porta

Donatella

. 2011. “The Elitist Populism of the Extreme Right: A Frame Analysis of Extreme Right-Wing Discourses in Italy and Germany.” Acta Politica 46 (2): 180–202. doi:10.1057/ap.2010.28.

Elchardus

Mark

Spruyt

Bram

. 2016. “Populism, Persistent Republicanism and Declinism: An Empirical Analysis of Populism as a Thin Ideology.” Government and Opposition 51 (1): 111–33. doi:10.1017/gov.2014.27.

Eurostat. 2015. “Level of immigration.” http://ec.europa.eu/eurostat/tgm/table.do?tab=tableandinit=1andlanguage=enandpcode=tps00176andplugin=1.

Gamson

William A.

1992. Talking Politics. Cambridge: Cambridge University Press.

Hameleers

Michael

Bos

Linda

de Vreese

Claes H.

2017. “‘They Did It’: The Effects of Emotionalized Blame Attribution in Populist Communication.” Communication Research 44 (6): 870–900. doi:10.1177/0093650216644026.

10.

Hox

Joop J.

Moerbeek

Mirjam

van de Schoot

Rens

. 2010. Multilevel Analysis: Techniques and Applications. London: Routledge.

11.

Jagers

Jan

Walgrave

Stefaan

. 2007. “Populism as Political Communication Style: An Empirical Study of Political Parties’ Discourse in Belgium.” European Journal of Political Research 46 (3): 319–45. doi:10.1111/j.1475-6765.2006.

12.

Laclau

Ernesto

. 2005. On Populist Reason. London: Verso.

13.

Matthes

Jörg

Schmuck

Desiree

. 2017. “The Effects of Anti-Immigrant Right-Wing Populist Ads on Implicit and Explicit Attitudes: A Moderated Mediation Model.” Communication Research 44 (4): 556–81. doi:10.1177/0093650215577859.

14.

Mazzoleni

Gianpietro

. 2008. “Populism and the Media.” In Twenty-First Century Populism: The Spectre of Western European Democracy, ed. Albertazzi

Daniele

McDonnell

Duncan

, 49–64. Basingstoke: Palgrave Macmillan.

15.

Moffitt

Benjamin

Tormey

Simon

. 2014. “Rethinking Populism: Politics, Mediatisation and Political Style.” Political Studies 62 (2): 381–97. doi:10.1111/1467-9248.12032.

16.

Mols

Frank

. 2012. “What Makes a Frame Persuasive? Lessons from Social Identity Theory.” Evidence & Policy: A Journal of Research, Debate and Practice 8 (3): 329–45. doi:10.1332/174426412X654059.

17.

Mudde

Cas

. 2004. “The Populist Zeitgeist.” Government and Opposition 39 (4): 542–64. doi:10.1111/j.1477-7053.2004.00135.x.

18.

Polletta

Francesca

Jasper

James M.

2001. “Collective Identity and Social Movements.” Annual Review of Sociology 27 (1): 283–305.

19.

Rooduijn

Matthijs

. 2014. “The Mesmerizing Message: The Diffusion of Populism in Public Debates in Western European Media.” Political Studies 62 (4): 726–44. doi:10.1111/1467-9248.12074.

20.

Simon

Bernd

Klandermans

Bert

. 2001. “Politicized Collective Identity: Asocial-Psychological Analysis.” American Psychologist 56:319–31. doi:10.1037/0003-066X.56.4.319.

21.

Stanyer

James

Salgado

Susana

Strömbäck

Jesper

. 2017. “Populist Actors as Communicators or Political Actors as Populist Communicators.” In Populist Political Communication in Europe, ed. Aalberg

Toril

Esser

Frank

Reinemann

Carsten

Strömbäck

Jesper

de Vreese

Claes H.

, 353–64. London: Routledge.

22.

Taggart

Paul

. 2000. Populism. Buckingham: Open University Press.

23.

Tajfel

Henri

. 1978. “Social Categorization, Social Identity, and Social Comparisons.” In Differentiation between Social Groups, ed. Tajfel

Henri

, 61-78. London: Academic Press.

24.

Tajfel

Henri

Turner

John C.

1986. “The Social Identity Theory of Inter-Group Behavior.” In Psychology of Intergroup Relations, ed. Worchel

Stephan

Austin

William G.

, 33-47. Chicago: Nelson-Hall.

25.

Van Zomeren

Martijn

Tom

Postmes

Spears

Russell

. 2008. “Toward an Integrative Social Identity Model of Collective Action: A Quantitative Research Synthesis of Three Socio-Psychological Perspectives.” Psychological Bulletin 134 (4): 504–35. doi:10.1037/0033-2909.134.4.504.

26.

World Bank. 2016. “Unemployment, Total (% of Total Labor Force) (Modeled ILO Estimate).” http://data.worldbank.org/indicator/SL.UEM.TOTL.ZS.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

2.36 MB