Shattering Populists’ Rhetoric with Satire at Elections Times: The Effect of Humorously Holding Populists Accountable for Their Lack of Solutions

The popularity of populist parties has increased worldwide, and disproportionate media attention for these parties arguably fueled their success. We empirically test our theoretical argument that political satire—with its humor, focus on internal contradictions, and lack of journalistic principles regarding objectivity and facticity—may be an effective antidote to populism’s success. Using a representative panel survey that functions as a natural experiment in the context of the Netherlands, we show that consumption of the satire show Zondag met Lubach (ZML) indeed lowered support for a right-wing populist party (PVV), its leader (Wilders), and the perceived capability of this party. Specifically, this study shows how satirically revealing the “weak spot” of populism (i.e., lacking concrete treatment recommendations) lowers populist support particularly among citizens already inclined to vote for populist parties. These findings have important democratic implications as they reveal that satire can have a real-life impact on the political decisions that voters make.

and satire), we develop and eventually test the theoretical argument that political satire may be particularly powerful to hold populist politicians accountable for their lack of concrete policy ideas and treatment recommendations. Moreover, we theorize and demonstrate under what conditions satire has the ability to break through the alleged echo chambers that are commonly assumed to boost the success of populist actors.
Although populist politicians and political satirists seem each other's natural opposites, this is partly a false juxtaposition: In terms of rhetoric, they have a lot in common. The communication styles of both populists (Aalberg, Esser, Reinemann, Strömbäck, & de Vreese, 2017;Krämer, 2014) and satirists (Baym, 2005;Gaines, 2007) are characterized by simplistic, short, and dramatized messages driven by common sense. Whereas populists use this rhetorical style to attribute blame to the establishment (Hameleers, Bos, & de Vreese, 2017), satirists employ it to uncover faulty lines of reasoning in politicians' rhetoric (Richmond & Porpora, 2019). Accordingly, satirists may be relatively effective in holding populist politicians accountable when compared to conventional journalists. After all, research has shown that traditional news coverage has rather spurred than halted the popularity of populist parties (Damstra, Jacobs, Boukes, & Vliegenthart, 2019;Mazzoleni, Stewart, & Horsfield, 2003). Although not all media outlets and journalists uncritically convey the ideas of populist actors (e.g., Wettstein, Esser, Schulz, Wirz, & Wirth, 2018), journalists are constrained to objectively report facts and observable reality (Borden & Tew, 2007). For this reason, it is complicated to criticize the unexpressed element of populism-that is, the lack of feasible solutions to deal with the problems foregrounded by populist politicians themselves-while also maintaining the perception of being impartial. Not afraid to violate norms of objectivity, satirists may be more effective at revealing this truth to voters (Richmond & Porpora, 2019). Thus, it can be argued that satire and conventional journalism need each other in a healthy deliberative democracy: The role perceptions of journalism strive toward truth-telling, objectivity and balance, which informs deliberative democracy (Waisbord, 2018). Less bound to these principles, satire can take a more active stance in attacking and challenging the claims made by politicians. Together, these formats inform citizens on external reality, whilst revealing issues in the performance of politicians-which can be used to hold them accountable.
The current study investigates the role of political satire in the 2017 Parliamentary elections of the Netherlands in which the right-wing populist party of Geert Wilders (Freedom Party, PVV) ended second in the official result. Wilders' rhetoric has followed a typical populist discourse by shifting blame to the (corrupt) establishment as well as cultivating a sense of urgency and crisis that only a strong leader like him can possibly solve (Aalberg et al., 2017;Hameleers, 2018;Jacobs & Spierings, 2019). Yet, any proposed solutions consistently remain abstract and unrealistic. This lack of solutions is exactly what the satirist television show Zondag met Lubach (ZML) addressed in the video "Geert Wilders concrete." This video sparked many social media reactions and became the show's third most-viewed clip (behind ZML clips that internationally went viral: Netherlands 2nd and a video about the NRA). As we consider this clip to represent a prototypical case of contemporary satirist content (Kilby, 2018;Richmond & Porpora, 2019) and the PVV to represent a most-likely case of a successful right-wing populist party (Aalberg et al., 2017), we believe that the developed theoretical framework and empirical findings may generalize to other multiparty contexts with popular satire shows and populist politicians.
The actual influence of satire-in general, but on populist politicians' popularity, specifically-is still unknown due to "the lack of long-term, longitudinal research designs" (Holbert, 2013, p. 311) and scholars' reliance on cross-sectional or experimental designs that mostly reveal "short-term, heuristics-based persuasive effects" (p. 310). We investigate this topic with a representative panel survey to disentangle cause and effect in a real-life setting. Our panel data allow scrutinizing (a) the effect of viewing ZML on the support for the PVV after the satire clip about Wilders aired, (b) how this eventually affected actual vote choice one wave later, and (c) which voters were affected in particular.

The role of the media in populism's success
Populist politicians attribute blame to the political elites, which are scapegoated for not representing the ordinary people and their will (Aalberg et al., 2017). Often they juxtapose the "good" ordinary people to the "corrupt" elite (Mudde, 2004;Taggart, 2000). Populists frequently refer to a severe crisis of the nation or cultivate a threat to the ordinary people on a transnational level (Moffitt, 2017), while "the people" are presented as a relatively powerless in-group whose will should be-but is not yet-prioritized in political decision-making (Mudde, 2004;Taggart, 2000). In this setting, populist politicians introduce themselves as the "reluctant" leaders that represent the people and their general interests (Aalberg et al., 2017;Taggart, 2000). Yet, solutions of populist parties to solve the perceived crisis are oftentimes simplistic, vague, and unfeasible and normally lack a clear treatment recommendation. The 2017 campaign program of the Freedom Party (PVV), for example, promised to shut down all mosques and to put a halt on immigration from non-Western countries to "de-Islamize" the country. Yet, it never becomes clear how these aims should be achieved, or what resources are necessary to fulfill these promises-if it were constitutionally viable at all.
Journalists do not necessarily support populism by covering them uncritically. Indeed, many media criticize populist politicians and their policies (Reinemann, Stanyer, Aalberg, Esser, & de Vreese, 2019;Wettstein et al., 2018); for example, by emphasizing that populists' anti-immigration policies are inhumane and that the hostile tone toward outgroups crosses the boundaries of freedom of speech. However, research still finds that regular media play an important role in fueling the success of populist political parties (Damstra et al., 2019;Ernst, Esser, Blassnig, & Engesser, 2019;Krämer, 2014) by oftentimes providing a favorable platform to spread their ideas (Mazzoleni et al., 2003).
One reason is the abundant and sometimes disproportionate attention for these parties, which can be explained by the close affinity between populist discourse and media logic (Hameleers, Bos, & de Vreese, 2017;Mazzoleni et al., 2003): Media logic emphasizes the news value of conflict, dramatization, simplification, and personalization, which all resonate with populists' communication tactics (Krämer, 2014;Mazzoleni, 2008) and, thus, gives populist politicians an advantage in the competition for media attention (Lawrence & Boydstun, 2017). Even when populists do not actively seek access to the mass media, the mainstream press often acts as a passive disseminator of their direct communication on social network sites (Aalberg et al., 2017), because it attracts eyeballs.
While covering a populist party, moreover, journalists often refrain from addressing the unobservable pitfalls in their rhetoric-afraid to be called "fake," "liberal," or "elitist". Journalists and politicians in democratic societies, traditionally, interact in mutually respected role-relationships (Van Dalen, 2019), where one acknowledges the legitimacy of the other. Populist actors, however, discredit the conventional news media and attack their legitimacy by depicting them as "elitist" and "belonging to the establishment." To re-establish their legitimacy, journalists' reaction has been to follow their professional routines even more closely (Van Dalen, 2019), in particular, the objectivity norm, which thus hampers uncovering the unspoken truth (i.e., populists' lack of clear treatment recommendations). Journalists' ability to "objectively" produce critical coverage through interviews with conventional politicians is restricted, moreover, because the latter often fail to properly challenge the populist candidates: In many cases, politicians fear that debating about immigration or simply giving attention to populist actors would further mobilize voters based on anti-immigrant sentiments and they, therefore, prefer to remain silent about these issues (Lucassen & Lucassen, 2015). This not only creates a breeding ground for dissatisfaction among the electorate (Luyendijk, 2019), but also leaves journalists without the necessary sources to produce a critical story.
Although a great body of research has provided important insights into the effects of populist communication, little is known about the potential consequences of communicating what populist actors do not talk about. For the reasons sketched above, it is difficult for conventional journalists to uncover the lack of concrete treatments beneath populism's thick layer of blame-framing and crisis urgency. We investigate whether the satire genre that is not bound to norms of objectivity may, instead, be more effective in confronting people with the discrepancy between populism's cultivation of causes but lack of feasible solutions.

Shattering populists' rhetoric by satire
Various characteristics of the satire genre make it an effective instrument to hold politicians-populist politicians, specifically-accountable for hollow and overly simplistic rhetoric (see Paletz, 1990). Essential is satirists' freedom to operate without journalistic ethics, such as objectivity, immediacy, and truthfulness (Borden & Tew, 2007;Ö dmark, 2018), which clearly overlaps with populist's rhetorical style. In the absence of a need to only report factual occurrences (i.e., truthfulness), political satire's "epistemological insight is based on textual interplay between the said and unsaid" (Meddaugh, 2010, p. 380); thus, giving satire the opportunity to emphasize the absence of solutions in populist rhetoric.
Not obliged to follow journalistic codes allows satirists to take a clear stance and critically scrutinize the arguments of a political actor (Baym, 2005); hypocrisy and deliberate omission of information are more easily revealed when one does not aim for neutrality (Richmond & Porpora, 2019). Regarding climate change, for instance, satirists have been free to explicitly affirm the reality of global warming and target the skeptics (Feldman, 2013), whereas journalists for a long time presented a "false balance" due to journalistic norms and professional routines (Brüggemann & Engesser, 2017). Similarly, satirists are able to challenge the claims that populist politicians make, who tend to substitute emotion for rational thinking (Jones & Baym, 2010), without the (journalistic) fear to be blamed for a lack of objectivity.
Recent forms of political satire, such as Last Week Tonight with John Oliver or Zondag Met Lubach, address the audience mostly with literal commentary and in a didactic manner (Young, 2013) by using playful but pedagogical critiques and clear juxtapositions (Waisanen, 2009). Accordingly, their message is unlikely to be misinterpreted (Boukes, 2019), which is in contrast to traditional forms of political parody (e.g., LaMarre, Landreville, & Beam, 2009;Matthes & Rauchfleisch, 2013). Hence, it may affect the audience across-the-board (Morris, 2009) and decrease overall support for satirized politicians (Becker, 2012;Warner et al., 2018).
Moreover, satire's semiotic application of "logical continuity" (Gaines, 2007) and "Socratic irony" (Warner, 2007) convincingly reveals the internal contradictions and inconsistencies between what politicians promise and what they actually (cannot) accomplish. Demonstrating what politicians say at one moment and how this differs from another occasion, not only has a humorous effect but also allows the audience to draw its own conclusions regarding a politician's reliability; the satirist not even has to offer his own interpretation, a short silence suffices (Gaines, 2007;Warner, 2007), which arguably strengthens the persuasive power of the message.
In particular concerning populist politicians who tend to promise more than they can deliver (i.e., they do not have feasible solutions), revealing the discrepancy between promises and reality should be an effective message strategy. Making the vagueness in their rhetorical claims obvious to the audience and identifying the narrative shifts of politicians, satirists can demonstrate that populist politicians evade answers and hide their interests (Waisanen, 2009). Satire, thereby, allows to restore the boundary between fact and opinion that is frequently obscured in the current political atmosphere (Jones & Baym, 2010;Richmond & Porpora, 2019). Next, we specify three components of populist support that may be affected negatively by consuming a satirical criticism directed at a populist party.

Three dimensions of populist party's popularity
We test the effect of a critical satire on three dimensions of populist party support: (a) people's propensity to vote for the populist party, (b) the attitude toward its party leader, and (c) the perceptions of how capable the populist party is in managing the topic of which they are the issue-owner (i.e., immigration). The first is important, because voting decisions reflect citizens' primary role in a democratic society (Strömbäck, 2005), and is usually assumed (but not often tested) to be the ultimate consequence of attitudinal effects.
Party leader attitudes are especially important in the realm of a populist party, because leaders of such parties are at the center stage of their party, more so than for other parties (Aalberg et al., 2017). Usually, populist parties are led by charismatic leaders who try to convey a self-image of being far removed from all other established politicians (Eatwell, 2003) and being the true representatives of the ordinary people (Taggart, 2000). Hence, populist parties and their electoral success (Stevens, Karp, & Hodgson, 2011) are shaped by, and oftentimes unthinkable without, their political leader. This is definitively also the case for Geert Wilders and his Freedom Party (PVV).
Perceptions of the issue-owned topic, finally, have been shown to be another crucial indicator regarding citizens' vote intentions (Smith, 2010). As right-wing populists are perceived as pundits on the issue of immigration (Smith, 2010), they should evoke the public perception of being able to successfully solve problems related to this issue. If satirists are able to punch holes in right-wing populists' promises by surfacing the lack of successful solutions to solve immigration threats, this marginalizes their reason to exist.
The theory above specified how political satire-with its substantive criticism addressed via effective message strategies-could decrease the popularity of populist parties. Having distinguished three dimensions of this popularity, we test the following hypotheses: H1: Consumption of a political satire show that explicitly holds a populist party and its leader accountable for a lack of solutions negatively affects (a) the propensity to vote for this party, (b) attitudes toward the party leader, and (c) perceptions of the party's capability to handle the issue of immigration.

Preaching to the choir or having a real impact?
Effects of regular political communication are often assumed to be minimal due to patterns of media consumption and persuasion that are driven by confirmation bias (Bennett & Iyengar, 2008): People select messages  and uncritically accept information that reassures their pre-existing beliefs (Bos, Van der Brug, & de Vreese, 2013;Müller et al., 2017), while selectively avoiding or counterarguing uncongenial messages (Knobloch-Westerwick, Mothes, & Polavin, 2017). Stronger effects may be expected in the context of political entertainment (Holbert, Garrett, & Gleason, 2010) for reasons of message selection and processing.
The alleged perseverance of echo chambers and confirmation biases implies that people with populist attitudes tend to select attitude-consistent information (Fawzi, 2019) and avoid or counter-argue news from conventional news media . In reality, though, most citizens are still exposed to a wide range of news sources: Filter bubbles still seem more a theoretical construct than an empirical fact (Zuiderveen Borgesius et al., 2016). Even more than conventional news, nevertheless, political satire is likely to stimulate cross-cutting exposure. For many people, the reason to tune in to satire is not to see balanced and nuanced news (Feldman, 2007) nor to retrieve information (Lee, 2013), but rather for reasons of entertainment and escapism (Diddi & LaRose, 2006). Especially in less polarized political contexts than the United States (e.g., multiparty systems as the Netherlands), this leaves possibilities for citizens to diversify the ideological nature of their media diet.
Once exposed to a counter-attitudinal message, moreover, people are less likely to respond by means of motivated reasoning if it is presented in a humorous manner. In contrast to traditional political messages, laughter may relieve tension (Paletz, 1990). Thereby, political humor will appear less confrontational and may function as a playful environment that allows partisans to see the other party not as opponent or enemy (Jones & Baym, 2010) but rather as legitimate discussion partners from the same community (Paletz, 1990).
Two cognitive mechanisms explain how satire disrupts counter-arguing when processing humorous messages. First, perceiving satire as entertainment, people may be transported into the satire's narrative and, hence, less critically evaluate its content as this would disturb enjoyment (Boukes, Boomgaarden, Moorman, & de Vreese, 2015;Nabi, Moyer-Gusé, & Byrne, 2007). Second, understanding satirewith all its absurdity, incongruity, and irony-requires a high cognitive load, which leaves insufficient cognitive resources to critically scrutinize the arguments in the satirical message (LaMarre, Landreville, Young, & Gilkerson, 2014;Young, 2008). Satire, thus, decreases both the motivation and ability to counter-argue. Accordingly, we identified two groups that would be most susceptible to a satire that provides a counter-narrative to populist rhetoric: (a) people intending to vote for a populist party and (b) politically cynical people. These are the people who have most sympathy to lose for these political actors.
Just like voters for any other party, citizens with populist attitudes are motivated to vote for right-wing populists because they agree with their issue positions (Smith, 2010;Van der Brug & Mughan, 2007). When satire overcomes populist voters' selective exposure biases, they will however be confronted with a reality check: The issues their leaders should be able to deal with are unlikely to be executed in a satisfactory manner. When exposed to a critical satire about this party, PVV supporters may consequentially (negatively) update their beliefs about the party more strongly than the people who are already unlikely to vote for them (Van Spanje & Azrout, 2019). Following Holbert and Park's (2019) typology of moderated relationships, we expect a contingent moderation in such a way that the popularity of the populist party only decreases amongst the probable voters for this party: H2: Being a probable voter for the PVV is a contingent moderator of the relationship between consumption of a political satire that explicitly holds the populist party and its leader accountable for its lack of solutions and the (a) vote intentions, (b) party leader attitude, and (c) issue-owner capability of this party, such that the statistically significant and negative effect of satire consumption is evident only for probable PVV voters. The effect of consuming this satire is nonsignificant for voters unlikely to vote for PVV.
Populist rhetoric is most effective among politically cynical citizens (Bos et al., 2013): People who are more skeptical about the political system or politics in general are most likely to support populist worldviews that emphasize a divide between the "ordinary people" and the "untrustworthy elite" (Aalberg et al., 2017). As they distrust mainstream politics, they perceive populist parties as a viable alternative; after all, populists promise to punish and replace the unresponsive elite. When satirists reveal that a populist political party is-even more than their mainstream counterparts-making strong claims and undeliverable promises, especially these cynical people may be disappointed to learn that this party will not be capable of keeping its promises either (Bos et al., 2013). We therefore expect a contributory moderation such that the negative effect on populist party popularity increases in strength (i.e., becomes more negative) for higher levels of political cynicism: H3: Political cynicism is a contributory moderator of the relationship between consumption of a satire that explicitly holds the populist party and its leader accountable for its lack of solutions and the (a) vote intentions, (b) party leader attitude, and (c) issue-owner capability of this party, such that the effect of consuming this satire is statistically significant and positive for all viewers, but stronger as citizens are more politically cynical. 15 March 2017). Between Wave 2 (27 January) and Wave 3 (2 March), an episode of satire show Zondag Met Lubach (ZML) featured a clip of 19 minutes that critically scrutinized the discourse and campaign promises of Geert Wilders' Freedom Party (PVV). Bearing a "striking resemblance" with Last Week Tonight (Nieuwenhuis, 2018, p. 78), ZML's host is sitting behind a news desk while delivering monologues about current affairs. The monologues have a high joke density with many ironic statements and a clear argumentative logic in which criticisms are systematically listed, often ending with a concrete call-to-action. The ZML episode that aired on 19 February 2017 (i.e., 11 days before the data collection of Wave 3) was seen by almost 1 million television viewers (15% of all viewers in the country), 1 and was the fourth most online-viewed program of the month. 2 The specific clip about Wilders in this episode went viral with 2.6 million views on YouTube, 3 and more than 2,000 retweets on Twitter. 4 Considering the small population of the Netherlands (17 million), its reach was considerable.
The episode of 19 February carried a clip entitled "Geert Wilders concrete," which step-by-step explained that this populist politician mainly articulates problems and blame attributions, but never clarifies how these societal issues could be solved. Concretely, Lubach (i.e., the host) mentioned that Wilders wants to close down mosques and Islamic schools as well as forbid the Quran. Lubach argued that the execution of this scheme, however, requires a dramatic change of the constitution, which is unlikely to happen. In addition, the clip shows that Wilders himself acknowledged in an interview that abolishing the Quran would be infeasible. Moreover, the satire clip emphasizes inconsistency in Wilders' rhetoric regarding the banning of books by comparing it to his position regarding Mein Kampf. The video also displays that Wilders does not come up with any solutions on Twitter (where he is very prominent), in interviews, in Parliamentary debates, nor in voting advice applications where other parties usually explain their standpoints (but the PVV refuses to do so).
The video then argues that one could still accept such undeveloped policy ideas if a political party would be newly established. However, it had been 4,552 days (i.e., 12 years) since Wilders' left the Liberal party and founded his own Freedom Party (PVV). Still, he and his party remain vague about concrete policies and avoid any debate about this. A concrete example in the video clip is PVV's party manifesto, which only consists of 259 words that fit on one A4-page (similar to U.S. letter). In the end, Lubach acted in an activist manner comparable to many U.S. satirists (Kilby, 2018): Lubach encouraged citizens and journalists to ask Wilders "How?" and introduced the hashtag #HoeDan, which people could use in reply to Wilders when he posted something on Twitter.
Sample Data collection was conducted by polling firm Kantar. A representative sample of adult citizens was recruited from the general voting population of whom the voting In the last wave, 12.8% of the respondents reported to have voted PVV (n ¼ 173), which closely matches the 13.1% in the actual election result.

Measurements 5
Independent variable: Consumption of ZML was measured in Wave 2 by asking how frequently people watched the show in an average month (i.e., this is a weekly program): (a) never, (b) once or twice, or (c) three to five times per month. For analytical purposes, a binary variable was created by categorizing those who "never" watched ZML (73.6%) apart from those who once per month or more frequently watched ZML (26.4%). The exposure level in our independent variable closely reflects the known viewing data of the specific clip about Wilders and the PVV. In that sense, we regard our study as a natural experiment in which one group (one could call this the "likely treatment condition") is compared to a "control condition" that was unlikely to receive the treatment. Yet, it needs to be stressed that the allocation to groups is nonrandom and partly driven by selective exposure, for which we control by means of lagged dependent variables (see discussion on "Controls" below). Dependent variables: Voters' support for the right-wing populist party was tapped with three separate measurements. First, the propensity to vote PVV was measured on a 10-point scale from "I will never vote for this party" (0) to "I will certainly vote for this party some day" (9). This question was repeated in all four waves (Wave 1: M ¼ 2.06, SD ¼ 3.01; Wave 2: M ¼ 2.31, SD ¼ 3.16; Wave 3: M ¼ 2.21, SD ¼ 3.20; Wave 4: M ¼ 2.07, SD ¼ 3.10).
Attitude toward party leader was measured in Waves 2 and 3 using a latent construct of four items all tapped on a 0-6 scale measuring how much Wilders was perceived as (a) trusted, (b) sympathetic, (c) competent, and (d) a true leader. The scale proved to be reliable in both waves (Wave 2: a ¼ .93, M ¼ 1.67, SD ¼ 1.68; Wave 3: Finally, the perception of the party's capability to handle the issue of immigration was measured using a scale from 0 (not suitable at all) to 6 (very suitable). The question was asked in Wave 1 (M ¼ 2.25, SD ¼ 2.20) and repeated in Wave 3 (M ¼ 2.36, SD ¼ 2.26).
Additional analyses examined whether and how eventual vote choice in the Parliamentary elections of 2017 (Wave 4) was affected. We created the dichotomous variable indicating whether people voted for the PVV (1; 12.8% of sample) or not (0; 87.2%).
Moderators: Initial intention to vote for the populist party was constructed by combining two repeated binary items from Waves 1 and 2: (a) whether people would vote PVV if there had been elections today or (b) considered voting PVV in the upcoming election. A dummy variable was created indicating whether people did not intend to vote PVV (70%, n ¼ 877) or indicated at least once to intend voting for PVV (30%, n ¼ 375). Although conceptually related, this moderator is measured in time before the dependent variables. The temporal order allows using it as a valid moderator to distinguish people who intended to vote for the populist party (probable PVV voters) and those who did not consider to do so before the satire broadcast.
Political cynicism was measured in Wave 2 using a combination of five items (a ¼ .92, M ¼ 3.34, SD ¼ 1.44) all tapped on a 0-6 scale (recoded, so higher score imply more cynicism): (a) satisfaction with how democracy is working in the Netherlands, (b) democracy works fairly in the Netherlands, (c) democracy works efficiently in the Netherlands, (d) politicians take sufficient account of the views of Dutch citizens, and (e) I have a lot of faith in Parliament. Both moderators-initial intention to vote PVV and political cynicism-did not correlate strongly (r ¼ .45, p < .001), indicating that these are different constructs.
Controls: Each of our models includes a lagged dependent variable. Lagged dependent variables are identical measurements of the outcome variable(s) obtained at an earlier moment in time (t -1 ). Including these as independent variables in a prediction, thus, controls for respondents' initial status and only leaves variance to be explained that accounts for changes over time in the dependent variable. By including the lagged dependent variable, it is unnecessary to employ additional control variables to account for between-individual variation-this is already captured in the lagged variable. To be sure, however, that it is the consumption of ZML and not demographic characteristics nor consumption of conventional news, we control for whether people voted PVV in the last national election (2012; 7.0%), age, education, and consumption of television news, online news, and newspapers.

Results
Propensity to vote, party leader attitudes, and issue-owner capability Table 1 shows the results of ordinary least squares (OLS) regression models predicting the effects of Zondag met Lubach (ZML) viewing. ZML consumption has negatively affected support for the populist party on all three dimensions. First, the propensity to vote PVV in Wave 3 decreased significantly among the viewers of ZML (b* ¼ À0.03, p ¼ .049). To verify that it was the ZML broadcast between Waves 2 and 3 that caused the effect, and not a general tendency among ZML viewers in election times, we tested whether this effect did not occur in similar analyses (with adjusted lagged dependent variables) for Wave 2 before which no such clip was broadcast. Indeed, no significant effect of ZML consumption on propensity to Note. Cells contain unstandardized regression coefficients (b) with standard errors (SE) in parentheses, standardized regression coefficients (b*) and probabilities (p; two-tailed).
vote PVV was detected in that wave (p ¼ .221). When propensity to vote PVV in Wave 4 was estimated with Wave 2 as lagged dependent variable (i.e., status before the satire broadcast), the effect of ZML consumption was still significant (b* ¼ À0.04, p ¼ .008) indicating that the effect was also detectable on the long term (i.e., 4 weeks later). In Wave 3, negative effects of ZML consumption were also yielded on the attitude toward its party leader Geert Wilders (b* ¼ À0.03, p ¼ .022) as well as on the perceived capability of the PVV to handle the issue of immigration (b* ¼ À0.03, p ¼ .047). Altogether, this provides evidence in line with Hypothesis 1 abc . Table 1 also shows that none of the other media variables affected populist party support. The only control variable that had an effect was education; as the election came nearer, support for the PVV and its leader especially decreased among the higher educated.

Moderation by vote intention and political cynicism
We test the hypothesized moderation effects by including interaction effects with (a) the initial intention to vote PVV and (b) political cynicism. Regarding the first, we find a significant interaction effect on the propensity to vote for PVV (b ¼ À0.57, SE ¼ 0.25, p ¼ .021). Figure 1 visualizes this interaction effect: The propensity to vote only decreased significantly among those who already considered voting PVV (p ¼ .022) but not among those who did not consider voting for this party (p ¼ .552). Thus, the negative effect of ZML was contingent on people's intention to vote for PVV and was insignificant for people unlikely to vote for PVV. Figure 1 Visualized interaction effect of ZML consumption and the initial intention to vote this party on propensity to vote PVV. So, whereas Hypothesis 2 has been confirmed regarding the propensity to vote (H 2a ), the interaction between ZML viewing and initial vote intention was insignificant for the other two dimensions of party support. No moderated relationship was found when the attitude toward party leader Wilders was examined (b ¼ À0.12, SE ¼ 0.14, p ¼ .392) nor when the perceived capability of the PVV to handle the issue of immigration was the dependent variable (b ¼ À0.21, SE ¼ 0.26, p ¼ .418). We therefore reject H 2b and H 2c .
A similar pattern of findings was yielded for Hypothesis 3. The interaction effect of ZML consumption with political cynicism was insignificant regarding both the perception of PVV's capability of handling the issue of immigration (b ¼ 0.07, SE ¼ 0.07, p ¼ .325) as well as the attitudes toward its party leader Wilders (b ¼ À0.06, SE ¼ 0.04, p ¼ .091). Thus, H 3b and H 3c were rejected. The latter finding, though, demonstrates a marginal tendency among the more politically cynical citizens to become more negative about Wilders due to the consumption of ZML.
In line with H 3a , the moderation by political cynicism was, again, significant regarding the propensity to vote PVV (b ¼ À0.20, SE ¼ 0.07, p ¼ .004). Figure 2 visualizes the marginal effect of ZML viewing on the propensity to vote PVV for different levels of political cynicism. It shows that the negative effect only becomes significant at cynicism levels that are little higher than the scales' middle point (i.e., political cynicism-values > 3.20). Hypothesis 3, thus, is not completely supported; the negative effect on vote propensity was indeed found to increase in strength for higher levels of political cynicism. However, the effect of ZML was actually contingent on political cynicism (i.e., not contributory); the effect was not significant for all ZML viewers, but only for those who were relatively more cynical about politics.

Indirectly affecting vote choice
We additionally analyzed the effect that ZML consumption had on the actual vote choice for PVV in the Parliamentary elections of 2017. A logistic regression (with the same set of controls and vote choice in the previous election of 2012 as lagged dependent variable) showed that ZML consumption had a negative but only marginally significant impact on the real vote choice too (b ¼ À0.46, p ¼ .060): The odds to vote PVV became 0.63 times less likely among ZML viewers. This effect was not conditional upon the intention to vote PVV in the previous waves (p ¼ .982) nor political cynicism (p ¼ .434).
Structural equation modeling (SEM) was applied to examine ZML's indirect effect on the actual vote that people casted during the election (correlation matrix of inserted variables is available in the Supporting Information). A SEM model was built that contained the vote for PVV in 2012 as a lagged dependent variable and the following exogenous variables: propensity to vote (Wave 2), party leader attitude (Wave 2), and PVV's perceived capability to handle immigration (Wave 1). These variables were repeated as potential mediators in Wave 3 (after the ZML episode). As outcome variable in Wave 4, we estimated the dummy variable voted for PVV. First, a model was constructed that achieved a close-to-perfect model fit: v 2 (2) ¼ 0.94, p ¼ .627; Comparative fit index (CFI) ¼ 1.00, standardized root mean square residual (SRMR) ¼ 0.00, and the root mean square error of approximation (RMSEA) ¼ .00 (90% confidence interval [CI] [0.00, 0.04]). Figure 3 gives a visual illustration of this model. To account for the dichotomous nature of the outcome variable (having voted PVV or not), we have used the Markov Chain Monte Carlo option in AMOS 25.0 to apply Bayesian estimation (328 burn-in observations, i.e., 25% of sample; and 100,000 analysis samples). 6 The fit measure also proved satisfactory (posterior predictive p ¼ .50; convergence statistic ¼ 1.02). Estimates of the direct effects confirmed the previously presented findings: ZML consumption negatively affected the propensity to vote (b ¼ À0.22, 95% CI [À0.42, À0.02]), the attitude toward Wilders (b ¼ À0.12, 95% CI [À0.23, À0.01]), as well as the PVV's capability to handle the issue of immigration (b ¼ À0.21, 95% CI [À0.40, À0.03])-all measured in Wave 3.

Discussion
The media are often regarded as a central factor in fueling the success of populists (Damstra et al., 2019;Mazzoleni et al., 2003) due to the resonance of populist rhetoric and media logic (Krämer, 2014). By connecting two distant streams of literature on populism and political satire, we approached the populism-media interdependency from a different perspective. Concretely, this study expands the theoretical reasoning (e.g., Richmond & Porpora, 2019) on why the satire genre is particularly powerful in holding populist politicians accountable for their hollow rhetoric: In contrast to journalists (see Van Dalen, 2019), satirists do not have to establish their legitimacy through factual and objective reporting (Borden & Tew, 2007;Jones & Baym, 2010): This allows satirists to reveal the "weak spot" of populists (i.e., a lack of concrete policy solutions) (Meddaugh, 2010). Moreover, the humor may release tension among citizens that otherwise would feel offended (Paletz, 1990) and it restricts critical processing; consequently, partisans cannot immediately counter-argue the information by means of motivated reasoning (LaMarre et al., 2014;Nabi et al., 2007;Young, 2008). Employing panel survey data, we investigated the case of ZML's episode about the populist party of Geert Wilders (PVV) and examined whether this piece of satire mitigated the party's success in the setting of an upcoming election. Concretely, we find that viewers of ZML lowered their propensity to vote PVV, developed more negative attitudes toward its party leader, and perceived the party as less capable after the ZML episode that addressed the lack of concreteness in the Wilders' rhetoric. Indirectly, it even decreased the likelihood that people voted PVV in the 2017 election, which demonstrates the real-world impact that satire can have. This study thus demonstrates that-also outside of experimental settings or cross-sectional designs-satire may negatively affect attitudes toward individual political candidates (Baumgartner, Morris, & Walth, 2012;Warner et al., 2018;Young, 2012), and that it is particularly the supporters of a satirized politician that amend their opinions (Becker, 2014;Xenos, Moy, & Becker, 2011;Young, 2004). Thereby, it adds another dimension to the satire literature and verifies the external validity of existing work on this genre: Satire not only influences knowledge (Hardy, Gottfried, Winneg, & Jamieson, 2014) but also electoral preferences in real-world conditions, and it does so with long-term persuasive effects (Feldman & Borum Chattoo, 2019).
Opposite to what is generally known about regular political communication, the propensity to vote decreased most strongly among people who already intended to vote PVV or were most politically cynical. Satire may thus relativize the impact of populist parties and their rhetoric: People who are likely to vote for a right-wing populist party are also most inclined to correct their political preferences when satirists inform them about the faulty lines of reasoning in populist discourse. A key implication of the current study is that the format of satire can break through echo chambers and patterns of motivated reasoning to successfully highlight discrepancies and inconsistencies in (populist) politicians' reasoning. Theoretically, our findings thus indicate that satire can be an effective format of political criticism to persuade citizens that are initially inclined to support the political party under attack.
However, these citizens were not affected more strongly regarding their attitudes toward the political leader nor the populist party's capability to handle immigration issues. This indicates that whereas satire has differential effects on propensity to vote (behavioral outcome), the perceptions of the party and its leader are affected uniformly (attitudinal outcome). Arguably, this is due to a bottom effect on the first: People who were initially not intending to vote for the populist party (M ¼ 0.53, SD ¼ 1.52) as well as those citizens who were least politically cynical (lowest 10%; M ¼ 0.75, SD ¼ 1.90) had such an aversion against the PVV, that their propensity to vote for this party could barely get any lower. The critical satire, accordingly, could not exercise a negative effect on top of that.
Opinions regarding party leader Wilders and the party's capability, however, may have been more ambivalent: Although citizens may be sure that they will never vote for the populist party, they could still see some value in the PVV's strict ideas on immigration or have some trust in Wilders' leadership capabilities. Watching ZML's satire, then, could negatively influence these evaluations among all citizens. Overall, this corresponds with literature on responsibility attribution (Bellucci, 2014;Marsh & Tilley, 2010): Holding politicians accountable for their failures can indeed lower the likelihood to cast a vote on this party, especially among those who initially considered doing this; for accountability to have an effect on vote intentions, however, people need to consider voting for this party in the first place. Evaluations of the party leader or the party's capabilities may still be affected across-the-board, because even its opponents may hold certain positive perceptions that could be overruled by the satire.
Previous research has found that satire may sometimes work counterproductive by increasing the affinity for those who are actually being mocked (Baumgartner & Morris, 2008). Other research would predict that satire's persuasive influence is limited, because humorous messages are easily discounted (Boukes et al., 2015;Nabi et al., 2007) due to the perception that these are less informational or credible and, therefore, should not be taken too seriously (Feldman & Borum Chattoo, 2019;LaMarre & Walther, 2013). However, these findings mainly occurred for lighthearted, Horatian styles of humor (LaMarre et al., 2014). Message discounting or message uncertainty, arguably, is unlikely for more explicit formats as Zondag Met Lubach, Full Frontal, or Last Week Tonight (Holbert, Hmielowski, Jain, Lather, & Morey, 2011;Landreville, 2015) in which the host delivers an overly clear message using a hard-boiled approach while seriously questioning political situations (Nieuwenhuis, 2018). Such satire is more likely to encourage political action (Becker, 2013) and make a real political impact (Boukes, 2019). As the current study demonstrates, the critical message in this type of satire is difficult to misinterpret and therefore unlikely to generate sympathy for the satirized politicians; by contrast, it reduced support for the satirized populist politician and his party.
A question one may raise is whether satire was not just making people cynical about politics in general (Hart & Hartelius, 2007), including people's perception of the populist party. Previous research found that satire may decrease trust in political candidates and evoke cynicism toward the electoral system (Baumgartner & Morris, 2006;Guggenheim, Kwak, & Campbell, 2011). This was not the case in the current study: In contrast, the repeated measurement of political cynicism was actually affected negatively by the consumption of ZML (b* ¼ À0.05, p ¼ .010), thus making people less cynical about politics. This strengthens the evidence that ZML viewers became critical of the satirized populist party, in particular.
The standardized effects may appear weak, however, this is the result of conservative model estimates and is probably a realistic estimation of reality in which many other factors determine people's vote choice too (Mak & Sidman, 2008). Lagged dependent variables were included as controls, which accounted for a major share of the variance, but are required to draw solid conclusions about causality.
Moreover, the ZML consumption measurement will have suffered from unreliability, not only because people find it difficult to recall media exposure (Prior, 2009) but also because respondents may have seen this video clip on other places than in ZML-it went viral online and people may not have realized that it belonged to ZML. This may all have had a downward bias on the observed effect strengths.
Our main conclusion is that satire has the potential to hold (populist) politicians accountable, which is partly due to the genre's lack of striving for neutrality (Borden & Tew, 2007;Richmond & Porpora, 2019). The question is how generalizable these findings are to other national contexts. For example, satire's influence may be restricted in the United Kingdom, where satire is expected to follow journalistic rules of impartiality, which may limit its critical potential (Bailey, 2018). Arguably, political context matters too. Citizens who intend to vote PVV have alternative parties in the Dutch multiparty system to consider if they want stricter migration policy or remove elites from power. Thus, party identity will be less strong and, therefore, less threatened by a critical satire than would be the case in polarized majoritarian electoral systems with only two major political parties (e.g., United States). Under those polarized conditions, it seems less likely that people expose themselves to such satire (e.g., Stroud & Muddiman, 2013) than in the Netherlands, where satirist Lubach is not perceived left or right but rather as an outside observer. It is also important to consider the transferability of our findings to other types of populist parties. Leftwing anti-establishment parties, for example, have been electorally successful in the Americas and Southern Europe. Although the issue positions of such parties are different (i.e., economic inclusionism), the core ideology of people-centrism and antielitism is similar to right-wing populists. Hence, satire may also offer an effective debunking of left-wing populists' rhetoric that contains blame attributions without clear treatment recommendation. However, this is all conditional upon the presence of satire shows in a country, which probably relates to the level of press freedom.
A question for future research is whether and how the satire has affected journalistic coverage of Wilders (e.g., Abel & Barthel, 2013) and what consequences it had for how political opponents treated Wilders in debates. The satire host, Arjen Lubach, explicitly called upon journalists and politicians to ask Wilders how he wanted to accomplish his apparent unfeasible plans. Any effects on these elite actors is especially likely when the humor has substantive elements (i.e., not just personalityoriented, see Meddaugh, 2010) as was the case in this item of ZML. Additionally, one may investigate whether ZML has weakened Wilders' leverage on social media; after all, his replies were flooded by #HoeDan responses-demonstrating the activating power of satirists' calls-to-action (Bode & Becker, 2018)-which may have distracted his followers from the message that Wilders was trying to convey.

Supporting Information
Additional Supporting Information may be found in the online version of this article.
Please note: Oxford University Press is not responsible for the content or functionality of any supplementary materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article.