Populist Disinformation: Exploring Intersections between Online Populism and Disinformation in the US and the Netherlands

The discursive construction of a populist divide between the ‘good’ people and ‘corrupt’ elites can conceptually be linked to disinformation. More specifically, (right-wing) populists are not only attributing blame to the political elites, but increasingly vent anti-media sentiments in which the mainstream press is scapegoated for not representing the people. In an era of post-truth relativism, ‘fake news’ is increasingly politicized and used as a label to delegitimize political opponents or the press. To better understand the affinity between disinformation and populism, this article conceptualizes two relationships between these concepts: (1) blame attributions to the dishonest media as part of the corrupt elites that mislead the people; and (2) the expression of populist boundaries in a people-centric, anti-expert, and evidence-free way. The results of a comparative qualitative content analysis in the US and Netherlands indicate that the political leaders Donald Trump and Geert Wilders blame legacy media in populist ways by regarding them as part of the corrupt and lying establishment. Compared to left-wing populist and mainstream politicians, these politicians are the most central players in the discursive construction of populist disinformation. Both politicians bypassed empirical evidence and expert knowledge whilst prioritizing the people’s truth and common sense at the center stage of honesty and reality. These expressions resonated with public opinion on Facebook, although citizens were more likely to frame misand disinformation in terms of ideological cleavages. These findings have important implications for our understanding of the role of populist discourse in a postfactual era.


Introduction
Populism and the uncontrolled spread of mis-and disinformation have been regarded as key threats to the functioning of representative democracy. Although populism and mis-and disinformation have been studied within separate research fields, we can identify an important conceptual affinity between these concepts (e.g., Waisbord, 2018). First of all, populism's antagonistic framing of the ordinary people versus the corrupt elites can be extrapolated to the attribution of blame to alleged inaccurate and dishonest media elites. Second, populism typically focuses on conflict and the people's feelings and experiences whilst circumventing or attacking empirical evidence and expert analyses. Although this does not mean that populism should be equated with the politics of disinformation, it does indicate that the central stylistic and framing elements of populism can give rise to a type of argumentation in which people-centric experiences are preferred over expert knowledge and empirical evidence.
Populist communication and mis-and disinformation may have similar political consequences. By shifting blame to the alleged 'corrupt' elites whilst emphasizing the centrality of the ordinary people, populist communication may polarize the electorate-cultivating an in-group of deprived people against other groups in society (Hameleers, Bos, & de Vreese, 2017;Müller et al., 2017). Literature on the political consequences of misand disinformation posits that people may be inclined to accept information that aligns with their partisan lenses, whereas they avoid or counterargue dissonant information (Thorson, 2016). As a consequence of such defensive motivations, polarization between opposing camps may be bolstered, placing people in fact-free populist echo chambers. In this article, we extend the conceptualization of the interconnectedness of populism and misand disinformation beyond their shared political consequences by focusing on two types of discursive relationships: (1) scapegoating the media as part of a populist communication strategy; and (2) populist disinformation as a discursive construction of fact-free, anti-elitist, and people-centric discourse. We rely on a qualitative content analysis of social media data collected in the US and the Netherlands to empirically explore the presence of these relationships. The central two-fold research question guiding this study is: (1) How are the media blamed for being dishonest and inaccurate; and (2) how are populist expressions related to a fact-free discourse?
Different actors in media, politics and society can directly spread (dis)information without the interference of media elites or journalistic routines, such as verification, accuracy and balance. Citizens can, for example, use social media to share their distrust in the media and politics, irrespective of the factual basis of their claims. Politicians may also use the oxygen of publicity provided by ungated social media to spread falsehoods across society. Geert Wilders in the Netherlands and Donald Trump in the US are two influential cases to consider in this regard: They are found to frequently blame the media for spreading lies that harm the ordinary people. Together, this article analyzes the discursive construction of populist disinformation by citizens and leading politicians in the US and the Netherlands. These countries are selected to compare how the discursive relationship between populism and disinformation is constructed in 'most different' media and political systems (e.g., Hallin & Mancini, 2004). Hence, the US has a bi-partisan political setting-which is mirrored in the ideological leaning of the press. The Netherlands, in contrast, is governed by a multiparty minority government, and the opposition consists of (smaller) left-and right-wing parties. Although some media outlets may have an ideological color, the Dutch press is less divided by ideological/partisan perspectives. In this setting, we aim to assess how robust and context-independent the discursive construction between populism and mis-and disinformation is.
The key findings of the qualitative content analysis indicate that both Trump in the US and Wilders in the Netherlands use social media to express their distrust in established institutions. These sentiments resonate with hostile media perceptions on the demandside of the electorate. Ordinary citizens use Facebook communities to express their closeness to the ordinary, honest people who share similar constructions of reality. Moreover, they mark their distance to lying elites and dishonest media outlets. These constructions are not voiced by mainstream or left-wing populist politicians, who express milder media criticisms that are more closely linked to misinformation attributions. These findings implicate that the discursive construction of populism and mis-and disinformation can be integrated on social network sites, where both politicians and ordinary people shape alternative versions of 'their' reality whilst discrediting the 'truths' disseminated by their opponents. An important theoretical implication is that constructions of 'truth' and 'fake' are driven by identity attachments and motivated reasoning rather than a deliberation of all available facts-augmenting polarized divides across society.

Populist Discourse and the Attribution of Blame to the Media
Populism revolves around the expression of a central divide in politics and society-the ordinary people are pitted against the 'corrupt' elites (e.g., Albertazzi & McDonnell, 2008;Canovan, 1999;Mudde, 2004;Mudde & Rovira Kaltwasser, 2017). As populism emphasizes that the ordinary people are not represented by the 'corrupt' and self-interested elites, populism relates to attributions of blame (Hameleers et al., 2017). More specifically, problems experienced by the ordinary people are allegedly caused by elites that are unwilling and unable to represent their 'own' people. Here, it is relevant to distinguish the ideational core of populism from host ideologies that may be associated with populism (also see Mudde, 2004;Mudde & Rovira Kaltwasser, 2017). Essentially, populism refers to a style, communication tactic, discourse, or (thin) ideology in which the ordinary people are framed in opposition to the corrupt elites. This core idea can be enriched with host ideologiessuch as nativism and anti-immigration sentiments on the right-wing and economic inclusion or anti-capitalism on the left-wing.
In this article, a communication approach to populism is taken (also see Aalberg, Esser, Reinemann, Strömbäck, & de Vreese, 2017;Jagers & Walgrave, 2007). More specifically, populist ideas come into being-and have real-life political consequences-when communicators (i.e., politicians, the media, citizens) emphasize populist ideas in their communication. Although a growing number of empirical studies are based on content analytic research on the expression of populist ideas in (online) media (e.g., Ernst, Esser, Blassnig, & Engesser, 2019;Schmuck & Hameleers, 2019;Waisbord & Amado, 2017), there is relatively little inductive research on the nature of populist discourse (but see e.g., Engesser, Ernst, Esser, & Büchel, 2017;Hameleers, 2019). Qualitative research indicates that populist ideas are frequently present in a fragmented way in the media-indicating that the different components of populist communication do not always co-occur as single frames or interpretations in texts . Extending this research, this article aims to explore how the central building blocks of populist discourse are represented in texts communicated by politicians and citizens, and how the expression of (fragments of) populist discourse resonate with the attribution of communicative untruthfulness or 'fake news. ' Misinformation can simply be defined as inaccurate or false information that is spread without the intention to mislead (e.g., Nyhan & Reifler, 2010;Wardle, 2017). Disinformation can be defined as the intentional (multimodal) doctoring, manipulation, or de-contextualization to reach a certain goal (e.g., Marwick & Lewis, 2017;Wardle, 2017). Although mis-and disinformation are different from populist communication, we can identify a discursive connection between the attribution of misand disinformation and the ideational core of populist blame attributions (i.e., using 'fake news' as a delegitimizing label or accusing politicians of spreading falsehoods). More specifically, populism's Manichean discourse bypasses elitist knowledge and expert opinion and stresses conflict, emotionalization and people centrism.
Populism's antagonistic view on society and politics has been associated with anti-media sentiments (e.g., Krämer, 2017). Hence, established media outlets can be regarded as part of the 'corrupt' establishment far-removed from the people's experiences. Populism's blame attribution strategy may thus apply to the attribution of causal responsibility to the media elite as well. Against this backdrop, we first of all identify a relationship between populist rhetoric and attributions of misand disinformation: Next to shifting blame to political elites, populist communication can shift blame to the established press or media elites for not representing the ordinary people's worldview (misinformation) or for deliberately lying to them (disinformation).
Although populism emphasizes a pervasive causal and moral divide between the ordinary people and the corrupt elites, most research has applied a rather limited conceptualization of the elites. Hence, the political elites on the national or supra-national are not the only elitist actors deemed responsible for causing the people's problems. By allegedly silencing the people's voice, and by promoting versions of reality that support the established political order, the mainstream media can be regarded as an important enemy of the people in populist discourse (e.g., Fawzi, 2019). We therefore need to extend our understanding of populist communication and shift our focus to the media elites and journalists as part of the people's enemy. As a first step, we thus conceptualize attributions of mis-and disinformation within a populist framework: Populism's antagonistic framing of a central opposition between ordinary and honest people and lying and corrupt elites may be extrapolated to media critique and hostility. Just like the political elites are held responsible for depriving the ordinary people, the media can be blamed for lying to the people, and deliberately misleading them by communicating misleading interpretations that suit their own political agenda.
Different actors can express populist and anti-media sentiments through different media channels. In line with the recent body of research on the content and effects of online populist communication (e.g., Engesser et al., 2017;Ernst et al., 2019), this article focuses on the communication of populist ideas via social media platforms. Although most empirical research has focused on the direct communication of populist ideas by (populist) actors , online media may also create a discursive platform for ordinary citizens to communicate populist ideas (Hameleers, 2019). Combining these approaches, this article aims to understand how politicians and members of the 'ordinary' people use social network sites to express populist boundaries that blame the media for the people's problems. On the actor level, we are mostly interested in how radical-right wing populist actors that have theoretically been associated with the spread of disinformation (Bennett & Livingston, 2018;Marwick & Lewis, 2017) attribute blame to the media by accusing them of disinformation. However, although conceptual literature has regarded the disinformation order as a radical-right wing phenomenon, it remains an open question if, and if so how, the affinity between the ideational core of populism and discourses of (un)truthfulness or the radical right-wing component is the driving force of attributions of blame to the media. For this reason, we will contrast conceptually most likely cases of media scapegoating (Trump and Wilders as radical right-wing populists) to other cases (left-wing populists and mainstream politicians).
In this article, a 'most different' systems design was chosen to explore the extent to which populist disinformation is constructed in similar ways in national settings that differ on a number of relevant factors. Specifically, we compare a bipartisan country (the US) to a multiparty system with a minority coalition (the Netherlands) to investigate whether the perseverance of partisan divides shapes attributions of populist disinformation in different ways. In addition, affective polarization along partisan lines has mostly been associated with the US (e.g., Iyengar & Hahn, 2009), whereas it is much less central in public opinion, media, and politics in the Netherlands. Finally, the presence of right-wing populism in US and Dutch politics differs. Although most literature has reached consensus that the Dutch politician Geert Wilders can be regarded as a (radical) right-wing populist actor (e.g., Aalberg et al., 2017), there has been less consensus on whether Trump is a populist or radical rightwing leader (Inglehart & Norris, 2016). Irrespective of the different ways in which Trump has been classified, empirical research confirmed that he communicates populist and nativist worldviews (Hameleers, 2019). Again, this article aims to explore how similar populist disinformation is constructed in different national settings. Across these national settings, we aim to arrive at a more comprehensive understanding of how the media are cultivated as a scapegoat in populist discourse. Therefore, the following research question is introduced: RQ1: How are references to the media as a scapegoat for the people's expressed in social media content in the US and the Netherlands?

Populist Disinformation: The Resonance of Populism with Fact-Free Communication
The second type of relationship between populism and communicative untruthfulness proposed in this articlepopulist misinformation-describes the resonance between populist styles of communication and the expression of fact-free sentiments that bypass expert knowledge and empirical evidence. Among other things, the style of populism has typically been regarded as peoplecentric, conflict-focused, emotionalized, and based on common sense and gut feelings (Ernst et al., 2019;Hameleers et al., 2017;Schmuck & Hameleers, 2019). Here, it should be emphasized that mis-and disinformation should not be conflated with the absence of factual information and/or verified empirical evidence. Hence, we argue that the circumvention of empirical evidence and expert knowledge may give rise to the reliance of a type of argumentation that relies on experiences and opinions instead of verified information. As populism shares a similar communication strategy (e.g., Aalberg et al., 2017;Jagers & Walgrave, 2007;Krämer, 2014), the second type of affinity between post-truth communication and populism should be regarded on the content level: a preference for people-centric experiences over hard facts and base rate information. Although this does not mean that such type of information is necessarily false, it does connect to a type of communication that deviates from journalistic principles that strive for the truth (Waisbord, 2018): The people's opinions and experiences are less susceptible to verification and scrutiny than information presented as empirical evidence.
Populist communication bypasses the elites and experts as a source of knowledge and claims to give voice to the ordinary people and their concerns (e.g., Krämer, 2014). These stylistic elements may give rise to a specific type of communication that resonates with misinformation: Populist communication may present information that is not based on empirical evidence and/or expert opinion, but rather on the feelings and experiences of the people. Here, it should be emphasized that such forms of evidence-free communication are not necessarily wrong or inaccurate. In fact, one of the role conceptions of journalism is to mobilize the public, and interpret issues by establishing a link between events that happened and the people on the streets. This means that giving a voice to the people, and emphasizing their interpretations of and connections to issues, is actually a focal part of quality journalism.
Situated in an era of post-factual relativism where even the most basic facts that can be judged as false or true are debated, (political) communication should at least be founded on a true factual basis (e.g., van Aelst et al., 2017). As an assessment of the normative implications an underpinnings of communication that avoids facts whilst prioritizing the ordinary people's lifeworld reaches beyond the scope of this empirical endeavor, this article aims to explore if, and if so, how, populist communication actually gives rise to a communication tactic that avoids verified empirical evidence and experts whilst prioritizing conflict, emotions, and people's experiences. The research question that guides this focus reads as follows: RQ2: To what extent is populist communication used to circumvent elitist knowledge and empirical evidence whilst prioritizing experiences, conflict, and people-centrism as the focal point of reality?

Method
To answer these research questions, this article reports the result of two qualitative content analyses conducted in the US and the Netherlands. The article analyzes direct communication via Twitter (politicians) and Facebook (citizens). These two social media channels are chosen for different reasons. Different social media channels may correspond to different affordances (e.g., Valenzuela, Correa, & Gil de Zúñiga, 2018). Twitter may be used to acquire novel information, and can be used as a one-directional communication channel where followers receive updates from connections that are not necessarily reciprocal connections or 'friends.' Politicians frequently use Twitter accounts as they can reach a large number of followers, with whom they do not have to be connected, which makes it a suitable platform for elitist communication via weak-tie networks. Communication among ordinary citizens on Facebook is more likely to be based on strong-tie networks (e.g., Valenzuela et al., 2018). However, politicians use Facebook in a different way: they communicate their (personal and political) viewpoints without necessarily interacting with their followers. They do not personally know their followers, and there is no reciprocity in the online exchanges. Yet, Facebook may create a stronger perception of interactivity and community because citizens can respond to posts by politicians and interact more directly with fellow citizens that respond to the same original posts. Interaction between users is afforded by both Twitter and Facebook, but Facebook interactions typically allow for richer and more detailed discussions and less elitist interactions than the response sections offered by Twitter.
Based on these theoretical premises, data from politicians' Twitter accounts in the Netherlands and the US were scraped. Here, two 'most likely' cases to express populist attributions of blame were first of all selected: Geert Wilders in the Netherlands and Donald Trump in the US. To further explore if, and if so, to what extent, populist media critique, attributions of blame to the media, and the expression of populist disinformation is a radical-ring wing populist phenomenon, these cases were contrasted to the direct communication of left-wing populist actors (Bernie Sanders in the US and the leftsocialist politician Emile Roemer of the Socialist Party in the Netherlands) and mainstream politicians (Hillary Clinton in the US and Mark Rutte in the Netherlands). Even though not all scholars may agree on the classification of Bernie Sanders and Emile Roemer as left-wing populist, empirical evidence at least indicates that the communication tactics of these actors at times align with populist rhetoric. The key aim of case selection was to test the theoretical premise that populist disinformation mainly pertains to the radical right-wing, or whether it can also be associated with the political communication of left-wing populist and mainstream actors.
For the sample of Facebook communities used by ordinary citizens, the most-likely cases strategy was also employed: Publicly accessible communities that revolves around the native people and their distrust in the elites, or nationalist pages more generally, were used to get inductive insights into the construction of populist misand disinformation.

Sample
The sample frame reflected key electoral events in both countries: the national elections in the Netherlands and the presidential elections in the US. In the Netherlands, the most recent general elections were held on March 15, 2017. All original tweets by Geert Wilders in a two months pre-election and a two months post-election period were scraped (N = 1,065) and supplemented with a routine period in 2016 and 2018. In the same period, all 124 tweets of the left-wing politician Emile Roemer and a sample of 558 tweets of the prime-minister were selected. The key electoral event in the US took place on November 8, 2016. In this country, the four-months Twitter activity yielded 1,153 tweets by Donald Trump (excluding non-relevant entries and retweets). This sample was extended with 603 tweets of Bernie Sanders and 405 tweets of Hillary Clinton. In the US, the same routine period as in the Dutch case was used for reasons of comparability.
In each country, two publicly accessible Facebook community pages that reflected radical-right wing issue positions were sampled. In the Netherlands, these platforms for example revolved around the theme of 'getting back the native people's country' and anti-immigration sentiments. Similar authoritarian pages were sampled in the US (one patriotist community page and one nativist page was selected). Within these two communities, contributions published in exactly the same timeframe as the Twitter posts were sampled.
The sampling strategy on Facebook was two-staged. More specifically, original posts had multiple replies that contained relevant information connected to the research question. Based on principles of maximum variation and saturation (e.g., Braun & Clarke, 2013;Glaser & Strauss, 1967), ten original posts in each community was found to be sufficient for saturation (meaning that an additional sample of new posts did not yield additional findings). For every post, the first ten replies were selected (ordered on date). Again, saturation was assessed by coding additional replies after the first ten. In some cases, the analysis of additional replies yielded additional insights, which were included in the analysis. Together, 20 posts and 215 replies were analyzed in the Netherlands. 20 posts and 234 replies were analyzed in the US. To contrast these pages to negative cases, we added one left-wing community page in each country. These pages mainly reflected an anti-corporation perspective, whilst articulating a more inclusive perspective on the people (which is in contrast to the authoritarian emphasis of the radical right-wing pages).

Analysis
All data were analyzed at the level of tweets, Facebook posts, or replies. The Grounded Theory approach was used to analyze the data in a step-by-step approach (Braun & Clarke, 2013;Charmaz, 2006;Glaser & Strauss, 1967). The coding process was selective in the sense that only excerpts that were relevant in light of the research questions were coded. First of all, open coding was applied to label segments of tweets, Facebook posts, and responses in light of the sensitizing concepts (i.e., discursive constructions of truth, fake, misinformation, disinformation, populism). Here, it should be noted that the coding process did not aim to classify information as misand/or disinformation on the content level (which would require fact-checking). Rather, we looked at how politicians and citizens referred to information spread without the intention to mislead (misinformation) or claims that are deliberately untrue (disinformation). Further, we looked at the type of argumentation used to make claims about reality: was empirical research quoted? Were expert analyses referred to? Did the politician or citizen refer to experiences and common sense as argumentation/evidence for issue positions?
During the second step of focused coding, this extensive list of codes (500+) was reduced by merging unique open codes, reformulating codes to higher levels of abstraction, and raising codes to categories. Codes were grouped and ordered based on their variety. In this process, piles of codes related to the construction of truth, the attribution of blame to (mainstream) media, falsehoods, and populism were made. These groups were used when conceptualizing dimensions that captured variety in the concepts of interest. Finally, during the step of axial coding, connections between these groups were made. The research questions were guiding during this final step of data reduction. More specifically, constructions of populism were connected to discourses of truth, reality, and disinformation. The outcomes of the three-stage analysis strategy are depicted in a conceptindicator-model (see Figure 1).

Validity and Reliability in Qualitative Text Analysis
It has been argued that the measures to ensure validity and reliability used in quantitative (content) analyses are not suited to the completely different nature, aims, and scope of qualitative (text) analysis (see e.g., Braun & Clarke, 2013). Responding to these different demands, all steps of coding have been discussed with a peer that was less involved in the study, but familiar with qualitative text analysis. The raw data files of 25 tweets and 25 Facebook posts were also coded independently by this second researcher. After this sample was coded twice, differences in the labeling of segments (open coding), the merging and grouping of codes (focused coding), and the conceptual connection between emerging dimensions (axial coding) was discussed extensively. Although minor differences in the allocation of open codes and the subsequent process of data reduction were identified, the final core themes that emerged from the raw data were similar, and resulted in the same answers to the two research questions.

How Right-Wing Populists Cultivate a Divide between the Honest People and Lying Press
Both Trump and Wilders scapegoated the traditional press for withholding the truth to the ordinary people. The media, and mainstream media in particular, were blamed for spreading lies that deprive the people of the truth. This can be exemplified by one of Wilders' tweets: "Most of the media channels have just one aim: to hurt me and the Freedom Party with their lies. Fortunately, we are stronger than the lies they are spreading" (Wilders, 2017b). Similar anti-media sentiments were expressed by Trump: "Not only does the media give a platform to hate groups, but the media turns a blind eye to the gang violence on our streets!" (Trump, 2017c).
Even more explicitly, Trump (2017b) actively refers to the cluster of media channels he distrust as the so-called 'fake news' media-which he regards as the greatest enemy of the American people: "The FAKE NEWS media (failing @nytimes, @NBCNews, @ABC, @CBS, @CNN) is not my enemy, it is the enemy of the American People!" References to the 'danger to our country' or 'the American people' explicate the discursive connec-  tion between populism and accusations of disinformation: Because the media are not reporting accurately on the facts that happened, and as they deliberately distort the truth, the native people are threatened severely. Wilders further emphasized the need to start a revolution to remove the elites in politics and media. A further analysis of the discourse used to frame the media as a culpable, elitist outsider reveals a clear distinction between mis-and disinformation. More specifically, Trump's and Wilders' references to the media emphasize that the media's dishonesty and inaccurate reporting is goal-directed and deliberate. As Trump (2017b) puts it: "FAKE NEWS media knowingly doesn't tell the truth. A great danger to our country. The failing @nytimes has become a joke. Likewise @CNN. Sad!" These accusations of disinformation further point to an alleged political goal or hidden agenda of the news media: "Crooked Hillary colluded w/FBI and DOJ and media is covering up to protect her. It's a #RiggedSystem! Our country deserves better!" (Trump, 2016). According to Trump, many news media outlets reside with opposed partisans. In this reading, these media outlets are propaganda machines that promote and uncritically disseminate the political agenda of the Democrats whilst disregarding, attacking or strategically neglecting the Republicans. Wilders further blames the media for selfcensorship, and for denying the 'real' problems facing the nation: the Islam. As he puts it: "Fear and selfcensorship of the coward media that ignores Islam as the greatest danger to our nation" (Wilders, 2017a).
The references used by Trump and Wilders to describe the media climate further confirm the discursive framing of an alleged climate of disinformation as a key threat to the native people. Trump, for example, uses adjectives as dishonest, rigged, dirty, crooked, and fake to denote that the media are an enemy of the people. Wilders refers to the media as unworldly, disgusting, cowards, or left-wing elitist. Here, it is important to note that Trump is much more selective in attributing blame to the media than Wilders. Whereas Wilders seems to scapegoat the 'news media' as a whole, Trump (2017a) clearly distinguishes between platforms that show a bias against his political viewpoints and media platforms that do report on the facts accurately: "The fake news media is going crazy with their conspiracy theories and blind hatred. @MSNBC & @CNN are unwatchable. @foxandfriends is great!" Hence, channels that are in line with Trump's political agenda are credited, and incongruent media are regarded as biased and blamed for their deliberate spread of dishonesty. For Wilders (2016), the divide is mostly based on a cleavage between the people's reality and the distorted worldviews of the left-wing elitist media: "A new all-time low for the left-wing media scums. Disgusting!" Contrasting the discursive constructions expressed by the right-wing populist leaders to other political actors, it can be confirmed that hostile media sentiments and accusations of disinformation do not spill over to left-wing populists or the mainstream. In the US, Bernie Sanders did not voice hostile media sentiments. Specifically, disinformation, 'fake news,' or related accusations were not addressed to the established press or other sources of information. The discourses of (un)truthfulness voiced by him did emphasize an antagonism between the people's reality and the lies spread by his political opponent Trump (at least in the pre-election period). In the pre-and post-election period, Clinton did not explicitly engage in discourses of (un)truthfulness. In the Dutch case, results are similar: The left-wing populist politician Emile Roemer did not voice anti-media sentiments, although he did emphasize a divide between hard-working ordinary people and corporate elites. Finally, the Dutch prime-minister Mark Rutte did not engage in any populist or anti-media discourse. Taken together, our findings illustrate that attributions of blame to the media, and using 'fake news' or disinformation as a delegitimizing label, was restricted to the direct communication of the two radical right-wing populist leaders in our sample. Although left-wing populists do stress a divide between the truth of the people and lying political opponents, these reality constructions reflect partisan and divides instead of blame attributions to the (established) press.

A Populist Conception of Truth and Reality: The People Know Best
The second type of relationship between populism and discourses of mis-and disinformation conceptualized in this article-populist disinformation-can be identified clearly in the direct communication of Trump and Wilders. In the direct Twitter communication of both Trump and Wilders, expert opinion is oftentimes neglected and discredited, whereas the ordinary people are regarded as the most reliable source of honest and accurate information. In the Netherlands, Wilders (2018b) frequently refers to common sense and the knowledge of the ordinary people to disregard expert knowledge on climate change, also referred to as "climate nonsense" (Wilders, 2018b) by Wilders and his followers: "Ordinary people confront the king with climate nonsense of our cabinet. Where do people get 15,000 euros from?" Wilders (2018a) also cultivates a divide between the representation of left-wing elitist parties and ordinary people represented by his Freedom Party: "The Freedom Party represents all ordinary people despite their color. The Greens only represent the white, leftwing elites." Common sense is used to depict the truth, without referring to any empirical evidence, numbers, or sources: "What is sure is that the ordinary Dutch person can pay for all this nonsense. The rest of it is based on lies and deception" (Wilders, 2018d). This type of evidence that prioritizes common sense and the ordinary people is used to interpret any kind of issue, for example the (failing) expenditures of the government: Electricity more expensive. VAT goes up. Rents higher. But billions of euros go to Africa. The ordinary Dutch people can bleed as a cause of the mistakes of the gang of our governmental leader. Give these billions to the hardworking Dutch citizens! (Wilders, 2018c) Hence, hard claims, as well as causal connections that resonate with a populist anti-elitist divide and threats to the ordinary people, are made without any references to evidence, statistics, numbers, or expert opinion.
A similar discourse construction of reality can be identified in Donald Trump's populist expressions. Trump (2018a) actively defends the political agenda he pursues as the agenda governed by the common sense of the American people: "Our agenda is NOT a partisan agenda-it is the mainstream, common sense agenda of the American People." Moreover, Trump explicitly refers to 'facts' and 'the truth' without giving any type of empirical evidence to support these truths. In these references to the truth and the centrality of the ordinary people, the two types of relationships between accusations of disinformation and a populist framing of truth and reality oftentimes co-occur in single interpretations: The Fake News hates me saying that they are the Enemy of the People only because they know it's TRUE. I am providing a great service by explaining this to the American People. They purposely cause great division & distrust. They can also cause War! They are very dangerous & sick! (Trump, 2018b) Cultivating the people's truth is not restricted to the discourse of radical right-wing populists. The left-wing populist actors in our sample emphasized that ordinary or native people are right, whereas elitist outsiders (i.e., corporations) are breaking their promises by lying to the people. This can be illustrated by the following tweet of Sanders (2016): "Time and again Native Americans have seen the government break solemn promises and corporations put profits ahead of their sovereign rights." Although people-centrism was less central in Emile Roemer's (2016) discourse, references to the majority of the people and their will were implicitly articulated by the Dutch left-wing populist politician: "The people know it-2/3 of all Dutch people agree with the Socialist Party: we need to let the profiting billionaires pay!" Turning to the mainstream politicians, Clinton in the US and Mark Rutte in the Netherlands do identify many references to the "American people" or "Dutch voters." Yet, these references do not cultivate a cleavage between common-sense and people-centric realities contrasted to the elite's lies.

The Audience's Perception of a Cleavage Between the Truthful 'Us' and Dishonest 'Them'
Both types of relationships between populism and misand disinformation identified in the Twitter communica-tion of Trump and Wilders are confirmed with the analyses of the public's discourse on Facebook. In this section, only the differences between politicians' reality constructions and citizens' interpretations on the demandside will be discussed in more detail. In US public discourse, citizens made a less fine-grained distinction between trustworthy and rigged media outlets compared to Trump. Hence, the media, opposed partisans, and governmental institutions were frequently lumped together as an elitist outsider that did not comprehend the people's lifeworld: "Those that are white getting in trouble for hate (racist) crimes and yet the far left communist Democrat controlled media never seem to report these hate crimes against the whites" (Facebook user, February 15, 2017). In the Netherlands, people mainly attributed blame to the media elites for looking away, or for being 'blind' to see the real problems and experiences of the ordinary people: "They make sure that this is not seen on TV. But the police is willing to join [protests initiated by the people]" (Facebook user, March 12, 2017). In addition, Dutch citizens focalize an overall sense of distrust in both the media and expert knowledge: "They are all shouting to tell exactly the same story. All the best to them: there is no one who trusts media, students or other so-called experts" (Facebook user, July 17, 2017).
The epistemic and moral boundary between the innocent and honest ordinary people and the lying elites was more salient on Facebook than reflected in the direct communication of both politicians. Similar to politicians' discourse, however, people refer to 'the truth' and 'reality' without using empirical evidence or facts: "That's the truth. People with jobs don't vote Democrat unless they just don't understand what goes on in this world" (Facebook user, April 6, 2017). In the Netherlands, this divide was further stressed by cultivating the divide along ideological lines. The left-wing was regarded as dishonest and far-removed from reality, whereas the 'real' ordinary people did know what was going on in society: "Like our Facebook page to show these left-wing idiots that reported our previous page that we are right. They do not want to see what is really going wrong here" (Facebook user, August 8, 2017).
The analysis from the negative cases-the leftoriented pages-reveal that people-centrism and a focus on the common sense of the ordinary people is a common theme on these community pages as well. On these pages, the reality constructions and lies of corporate and political elites are contrasted to the ordinary people's experiences. Here, we see a left-wing populist construction in which the hardworking ordinary citizen is juxtaposed to the self-interested elites. Media critique is salient on these pages as well, but it takes on less hostile forms. See, for example, the following statement voiced on a Dutch Facebook community: "The media do not report accurately. They present a worldview that does not take these factors into account" (Facebook user, September 8, 2017). Although the hostile media critique on the rightwing populist pages may be considered as accusations of disinformation, the left-wing pages more closely reflected attributions of misinformation.

Discussion and Conclusion
The alleged uncontrolled spread of dishonest or inaccurate information in today's fragmented media environment may have severe political consequences (van Aelst et al., 2017). More specifically, the epistemic status of factual information increasingly becomes the focal point of heated debates, and the acceptance of information may be driven by defensive and consistency motivations rather than veracity (e.g., Nyhan & Reifler, 2010). Extending this line of argumentation, this article has proposed a two-fold relationship between populism and mis-and disinformation: (1) the attribution of misand disinformation to the (media) elites; and (2) populist disinformation as a communication style that avoids empirical evidence and expert analysis, whilst placing common sense and the ordinary people at the center stage of reality. Two qualitative content analyses in the US and the Netherlands were conducted to provide indepth insights into the affinity between populism and mis-and disinformation: How are social media platforms providing a discursive opportunity for politicians and ordinary citizens to express populist boundaries between the truthful us and the dishonest them?
First of all, we found that both Trump in the US and Wilders in the Netherlands expressed a populist boundary between the dishonest, inaccurate, and fake media and ordinary native people that were victimized by the media's dishonesty. These attributions tie in with disinformation: The media were accused of deliberately distorting reality to promote their own biased political agendas. The language used by both politicians further indicate that the media are blamed for looking away, and denying the problems experienced by the ordinary people. There was one noteworthy difference between Trump's and Wilders' anti-media discourse: Wilders mostly attributed blame to the media in general, whereas Trump more specifically blamed the media outlets that did not support his partisan views. This finding can be interpreted as a stronger hostile media bias in the US. Here, media outlets that expressed incongruent viewpoints were regarded as biased against the views of the people (e.g., Vallone, Ross, & Lepper, 1985) and the truth in general. Donald Trump thus selectively blamed and credited sources to defend the partisan views he communicated to his followers. This finding can be explained in light of the different media discourses at the time of data collection. Although most legacy media were critical toward Wilders and his policies (i.e., there are at least no clear indications that certain media were systematically more favorable toward Wilders as compared to other outlets), Trump could more clearly rely on the US partisan media system: Certain media may be explicitly negative in their coverage, whereas others may be systematically more favorable because of a political parallelism.
In support of the theoretical notion of the expression of fact-free and people centric-communication in populist discourse (e.g., Waisbord, 2018), both politicians clearly avoided expert knowledge, statistics, verifiable facts or evidence, and relied on common sense and the people's truth as evidence for the populist claims they made. There was little room for balance or opposing viewpoints, and the populist discourse was generally one-sided and presented as the only reality opposed to the 'fake news' presented by opposing politicians and media sources. An important implication of these findings is that social network sites, such as Twitter, provide (populist) politicians with a platform to express disinformation to strategically attack the politicians they oppose. This may eventually increase polarized divides in society, and raise levels of political distrust and cynicism among the electorate (Marwick & Lewis, 2017). Explicit attacks targeted at the news media were only found on the radical-right, and did not spill over to the communication tactics of left-wing populists or mainstream politicians in the two countries. Left-wing populists did, however, emphasize people centrality and attributed dishonesty to their political rivals and the (corporate) establishment. Hence, discourses of untruthfulness can be connected to populism in general, whereas the explicit reliance on common sense and emotions as the focal point of reality and blame attribution to the 'lying' established press is a communication tactic of radical right-wing populists in particular. In that sense, it seems that the authoritarian element that sets right-wing populism apart from the thin-cored ideology of populism (Mudde, 2007) can be associated with hostile critique on the established press and a circumvention of expert knowledge and empirical evidence. But how is populist disinformation shaped by the public on social media?
The content analysis of citizens' discourse on Facebook largely confirms the findings of the politicians' discourse, pointing to an alignment of populist interpretations between the supply and demand-side. The difference mainly revolves around the type of moral and epistemic cleavage emphasized by the public. More specifically, Dutch citizens were more likely to cultivate a divide between their in-group and the ignorant left-wing people. In the US, the political and media elites were frequently lumped together, whereas Trump articulated a more fine-grained distinction between the 'fake news' media and politicians of the opposed party. The analysis of the negative cases-left-wing oriented Facebook community pages-revealed that emphasizing the people's truth is not restricted to radical right-wing populist interpretations. However, media critique was less hostile and focused more on unintended false information (misinformation) than intentional deception (disinformation). These findings indicate that citizens communicating their political perspectives on different platforms do distinguish between attributions of mis-and disinformation.
Despite providing important new insights into how mis-and disinformation can be situated in populist dis-course, this study has some limitations. First of all, the empirical study only focused on two social media platforms. Future research may extend the analysis to different platforms (i.e., including commentary sections of mainstream outlets) and political actors (i.e., distinction between populist and mainstream actors may be relevant). Second, the qualitative and inductive findings presented in this article provided important first insights into how populist disinformation manifests itself online, but may be extended with (automated) content analytic research that also provides insights into the relative salience of, and relationships between, different forms of populist sentiments targeted at the media.
Despite these limitations, this article contributes to our understanding of the current post-factual media era and populist zeitgeist-and the interconnectedness of these communicative phenomena-indicating in what ways different actors can use social network sites to express a pervasive divide between the 'honest' people and 'the others.'