Populism and Informal Fallacies: An Analysis of Right-Wing Populist Rhetoric in Election Campaigns

Abstract

Populism is on the rise, especially in Western Europe. While it is often assumed that populist actors have a tendency for fallacious reasoning, this has not been systematically investigated. We analyze the use of informal fallacies by right-wing populist politicians and their representation in the media during election campaigns. We conduct a quantitative content analysis of press releases of right-wing populist parties and news articles in print media during the most recent elections in the United Kingdom and Switzerland in 2015. The results show that fallacies are used in more than a third of all analyzed texts and overwhelmingly co-occur with populist key messages. Moreover, fallacies occur more often in populist parties’ press releases than in news articles and are more common in Switzerland compared to the United Kingdom. This study confirms the argument that populist actors use fallacies in combination with populist claims.

1 Introduction

“Little Marco”, “Lyin’ Ted”, “Crooked Hillary”—assigning negative nicknames to opponents was one of Donald Trump’s trademark political strategies in the 2016 presidential election campaign. Not only did these nicknames show an impressive tendency to stick—they are also perfect examples for the use of informal fallacies by a politician with right-wing populist tendencies. More precisely, these degrading nicknames are typical ad hominem arguments: personal attacks on other people—instead of the content of their argument—that have nothing to do with the truth of the target’s statements.

A fallacy, such as an ad hominem argument, is an erroneous argument that seems to be correct but in reality is not (Joseph 1906; Walton 1995). While formal fallacies contain mistakes in their logical form, informal fallacies arise from their misapplication in a specific context (Bunnin and Yu 2004). Informal fallacies, thus, may have a correct logical form, but show other flaws: They are untrustworthy because one or more of the premises are irrelevant or insufficient. Informal fallacies mostly occur in persuasive discourses, where participants try to convince each other using strong arguments (Walton 2008, pp. 10–11). Since the study at hand is interested in communication by right-wing “populist” parties in election campaigns—a special type of persuasive discourse—informal rather than formal fallacies are investigated.

Earlier studies have demonstrated that several informal fallacies are strategically used in political debates to justify seemingly coherent and strong positions (Zurloni and Anolli 2013). Hence, politicians may use fallacious arguments as rhetorical tools to steer the discourse into the direction that serves their own interest best, also known as strategic maneuvering (van Eemeren and Houtlosser 1999). Without critical assessment, fallacies may be perceived as reasonable arguments and thus mislead, intentionally or not, audiences into believing certain things to be true that are not (van Eemeren and Houtlosser 1999; Wilson et al. 2012, p. 4). This is especially problematic if such misinformation is disseminated by the mass media (Southwell and Thorson 2015). Moreover, the uncritical dissemination of fallacies can have far-reaching consequences, as the literature on fallacies in connection with the Iraq war (Sahlane 2015; Wilson et al. 2012) or health communication (Heiss and Bates 2016) shows. Another problematic political phenomenon that is often brought into connection with the use of fallacious rhetoric is populism. As our entry example also illustrates, populist actors are often accused of using informal fallacies in their communication (e.g., Reisigl 2002). This is particularly relevant because the supporters of populist politicians may be specifically prone to take the populists’ statements as the truth and, furthermore, be more resistant to corrections by the mass media due to stronger media skepticism (Reinemann et al. 2017). Building upon this literature, we investigate how right-wing populist actors use informal fallacies as a rhetorical instrument.

Populism is a widespread phenomenon that has gained traction in many Western democracies during the last few decades. Many studies have dealt with this rise and spread of populism, especially with regard to right-wing populists in Western Europe (e.g., Aalberg et al. 2017; Albertazzi and McDonnell 2008b; Kriesi and Pappas 2015; Mudde 2004; Mudde and Rovira Kaltwasser 2012). Recent studies have investigated populist communication in the press (e.g. Akkerman 2011; Bos et al. 2011; Rooduijn 2014b) as well as in party publications such as press releases or party manifestos (e.g., Bernhard 2016; Pauwels 2011; Rooduijn and Akkerman 2015; Rooduijn et al. 2014). However, not in combination with the concept of fallacies.

Populist argumentation—the question of how populists attempt to justify their demands or try to generate support for their policies—is still rarely addressed from a political communication perspective. While it is often suggested that populists tend to use fallacious argumentation (Reisigl 2002) or a kind of “Stammtisch-discourse” (Mudde 2001, p. 37), this has not been systematically investigated. The extant research on populist rhetoric and argumentation has mainly been qualitative, employing, for example, critical discourse analysis or a linguistic approach (Geden 2006; Reisigl 2012; Wodak 2003, 2015; Wodak et al. 2013). Most scholars have focused on single countries (Luginbühl 2010; Reisigl 2002; Wodak 2003) or comparative case studies (Wodak et al. 2013). These investigations provide rich, in-depth, and contextualized insights into aspects of populist rhetoric. However, they also entail inherent problems, such as a lack of generalizability or overestimation of single, idiosyncratic observations (Esser and Hanitzsch 2012). While the aim of qualitative research of argumentative discourse is generally to provide interpretation and meaning, quantitative research seeks to predict, explain, and understand (Labrie and Schulz 2015). Thus, the ambition of this study is to combine qualitative concepts and quantitative methods.

To do so, an investigation of persuasive discourses in which populist actors are participating is necessary. Election campaign communication is a prime example of such discourses: parties and politicians try to convince voters and opponents to follow their arguments and support them politically. In that sense, election campaigns are prototypical events in which discursive cultures crystallize and concentrate (Esser and Strömbäck 2012, p. 308). While the mass media are appropriate objects of study to locate this rhetorical strategies, journalists are not neutral: They are the gatekeepers (Lewin 1947; White 1950) that necessarily have to select some stories and statements and discard others. They do so according to specific media logics (Altheide and Snow 1979; Esser 2013, pp. 166–174) and news values (Galtung and Ruge 1965; Lippmann 1922; Schulz 1976). In order to allow some variation in the extent of these journalistic “filters”, press releases by the parties are analyzed in addition to mass media content. Furthermore, “quality” or “up-market” papers might show different media logics than “tabloid” or “mass-market” papers. Accordingly, both types of media are sampled. Finally, since established political parties show higher news values than non-established ones—and thus better chances to pass the journalistic gate (e.g., Hänggli 2012)—both an established and a less established populist right-wing party will be analyzed.

Taken together, the paper thus aims to analyze how right-wing populist actors actually use fallacies in their election campaign communication. We further investigate how fallacies and populist communication are related, and whether specific characteristics of the party or media context affect the usage of fallacies by populist actors. The analysis is based on a quantitative content analysis of right-wing populist parties’ self-presentation in press releases and their representation in the print media during the most recent elections in Switzerland and the United Kingdom, comparing the Swiss People’s Party (SVP) and the United Kingdom Independence Party (UKIP). In addition to answering the research question and hypotheses, we also provide qualitative examples for the use of informal fallacies by the investigated actors. There is often the impression that populist ideology and fallacious argumentation go hand in hand in public communication. We now provide empirical evidence for this assumption.

2 Theoretical Background

2.1 Populism, Populist Communication, and Populist Rhetoric

Populism is often described as a “notoriously vague term” (Canovan 1999, p. 3). While the term populism is regularly used in everyday discourse, the nature of the phenomenon itself is contentious, showing a lack of definitional consensus and conceptual clarity (Taguieff 1997, p. 11). Recently, scholars increasingly agree to conceive populism as a thin and less elaborate ideology (Abts and Rummens 2007; Albertazzi and McDonnell 2008a; Kriesi 2014; Mudde 2004; Stanley 2008; Wirth et al. 2016) and to understand it as a set of ideas (Hawkins 2009; Rooduijn 2014b; Rooduijn and Akkerman 2015; Taggart 2000). This refers to the fact that any thick ideology can be attached to this thin ideology (and thus creating specific types of populism, such as left-wing or right-wing populism). A popular definition of populism is provided by Mudde (2004, p. 543), who describes populism as

an ideology that considers society to be ultimately separated into two homogenous and antagonistic groups – ‘the pure people’ versus the ‘corrupt elite’, and which argues that politics should be an expression of the volonté générale (general will) of the people.

This definition includes the core characteristics of populist ideology: A Manichean and moralistic juxtaposition between the “good people” and the “evil elite” (Rooduijn and Akkerman 2015, p. 2), a demand for the empowerment and sovereignty of the people, and a monolithic conception of the people with a common understanding of the world. The latter implies that certain out-groups—specific segments of the population—are excluded from the “good people” and seen as a threat or burden to society (Albertazzi and McDonnell 2008a, p. 6; Jagers and Walgrave 2007, p. 324). This exclusionism is not necessarily a core characteristic of populism, but instead specific of right-wing populism in Western Europe (Rooduijn 2014a). By defining the Others, the thin populist ideology becomes a thick right-wing ideology. Following this conceptualization, right-wing populism consists of four core concepts: people-centrism, anti-elitism, popular sovereignty, and the exclusion of others.

Despite this growing consensus, defining populism as an ideology also has its critics (see Aslanidis 2016). Other authors conceive populism as a communication style, discourse, or frame (Aslanidis 2016; Bos et al. 2011; Canovan 1999; Cranmer 2011; Hawkins 2009; Jagers and Walgrave 2007; Laclau 2005). However, these approaches are not mutually exclusive. As Kriesi (2014, p. 364) states: “The populist ideology manifests itself in the political communication strategies of populist leaders.” This means that ideology may be the content and style may be the form of populist communication (Engesser et al. 2017). Based on the conceptualization of populism as a thin ideology, the content of populist communication—or populist key messages—consists of the same four core dimensions.

Other characteristics attributed to populist communication are more concerned with rhetoric or style. Reisigl (2002, p. 166ff), focusing on the right-wing Freedom Party of Austria (FPÖ), identifies characteristic principles of populist rhetoric such as “black- and white” rhetoric, reduction of complexity based on drastically simplified representations, a harsh tone with opponents or calculated ambivalence. Similar rhetorical styles have been identified by other scholars as typical for populist communication (Abts and Rummens 2007, p. 407; Geden 2006, p. 22; Luginbühl 2010; Wirth et al. 2016, p. 42; Wodak 2013).

Reisigl (2002, p. 185) further mentions another important aspect of right-wing populist rhetoric: its specific form of argumentation. For, according to the author, populists’ reasoning often violates certain principles and guidelines, such as logical validity, which are essential for a sound argument (Reisigl 2002, p. 185). Following argumentation theory, if an argument violates one of these rules, it is considered a fallacy. Reisigl (2002, pp. 186–196) lists several argumentation patterns—he calls them “topoi”—typical for right-wing populism that qualify as so-called fallacies.

2.2 Informal Fallacies

The concept of fallacies—which is mainly used in argumentation theory—and their investigation date back to Aristotle (Hamblin 1970, p. 12). Similar to populism, there is a certain variety in how a fallacy has been defined conceptually. In this study, we follow the conceptualization of informal fallacies by Walton (e.g., Walton 1995). Informal fallacies are arguments with flaws that do not concern formal logic but are caused by an incorrect use of the underlying argumentation scheme in the specific context (Bunnin and Yu 2004, p. 248; Walton 2008, p. 15). Informal fallacies thus violate the rules of a critical discourse, which ensure that actors involved in a discourse argue in a rational way (Habermas 2006; van Eemeren 2015, p. 214; van Eemeren and Grootendorst 1984). One example for an informal fallacy is the “ad populum” fallacy, which, according to Walton (1995, p. 225), describes the claim “[…] that a belief must be true or an action must be right because most or all people accept it or approve of it.” It is an appeal to emotions, usually as an attempt to sidetrack what should be the real issue of an argument (Walton 2008, p. 110). This is very reminiscent of Reisigl’s (2002, pp. 186–196) “Topos des Volkes” [topos of the people], which he sees as typically populist argumentation pattern.

Following Walton (2008), we analyze the argumentation schemes linked to different types of informal fallacies. These fallacies are chosen based on an extensive literature review as well as various pre-tests and will be described in further detail in the method section, including examples. The following 15 fallacies are included in the study at hand: Petitio principii, ad hominem, ad consequentiam, slippery slope,ad baculum, false cause (‘post hoc ergo propter hoc’), ad misericordiam,ad ignorantiam, false analogy, hasty generalization, straw man, ad populum, ad verecundiam, many questions, and red herring. These fallacies cover a wide range of different fallacies, corresponding to different aspects of a persuasive discourse (e.g., attacks, wrong premises, appeals to emotions, flaws related to induction and statistics, etc.). For example, the petitio principii, ad ignorantiam, ad consequentiam and many questions refer to wrong premises; ad misericordiam and ad populum are appeals to emotion. By doing so, we hope to find a wide range of fallacious statements in the raw content analysis material.

2.3 Context: Media Channel and Party Characteristics

Political communication by parties during election campaigns is a prime example of persuasive discourse, in which communication cultures and rhetorical strategies in politics crystallize in a short period of time. Additionally, it is known beforehand when elections will occur, which makes scientific inquiries more easy to prepare (Esser and Strömbäck 2012, p. 308). Hence, election campaigns are suitable objects of study for the analysis of fallacious argumentation by politicians.

The mass media are an adequate channel to investigate politicians’ argumentation, since the media cover parties’ campaigns to a great extent. However, journalists are not neutral and the mass media are not a passive conduit that simply passes party messages to the public. Journalists actively select stories and statements, while discarding others—they are the gatekeepers (Lewin 1947; White 1950) that decide what to publish or not. They do so according to media logics comprised of economic, professional, and technological aspects (Altheide and Snow 1979; Esser 2013, pp. 166–174). For example, since media are often organized as private, for-profit corporations, aspects of popularity and interestingness drive some journalistic decisions (economic elements of media logic). Furthermore, professional values such as objectivity, a certain public service attitude, etc., are crucial determinants of journalistic content (professional elements). Technological aspects are also relevant since audio-visual media allow for different modes of presentation than pure print media.

The concept of news values (Galtung and Ruge 1965) corresponds closely to media logic but solely describes characteristics of stories and statements and whether these characteristics render certain stories more attractive for the news than others. For example, geographic and cultural proximity, emotionality, surprise and recency, established and known actors and sources, crime and damage, etc., are classic news values that render stories containing such elements more newsworthy than stories without such elements. Taken together, the concepts of media logic and news values emphasize the selectivity inherent in any news published through mass media.

This raises another issue: There might be different media logics depending on the type (respectively, the business model) of media that is investigated. I.e., the proportion and composition of the single dimensions (economic media logic, professional media logic, and technical aspects) might differ between various media. For example, so-called “tabloid” or “mass-market” newspapers might emphasize different news values than so-called “quality” or “up-market” papers. For example, the former might place stronger emphasis on news values such as conflict, human interest stories, or might use bigger, brighter, and more provocative pictures. In order to grasp the full extent of fallacies and populism in different media outlets, an analysis of both tabloid as well as quality newspapers is therefore necessary.

According to professional aspects of media logic such as objectivity, relevance, quality, etc., one could think that journalists might detect and filter out fallacious arguments. This would be disadvantageous for the study at hand since we would not find any fallacies in the mass media if that were the case. However, this issue can be turned into an advantage by analyzing not only mass media content with regard to fallacies but also considering the “input side” of news during elections: Press releases. They are released by the parties and politicians themselves and thus do not have to pass any journalistic filters and gates. By analyzing both of these channels of communication, conclusions can be drawn about the difference that the presence of journalists and the respective filters make.

Finally, the concept of news values also implies that aspects of the political parties could influence their newsworthiness. Obviously, small fringe parties that have never had any seats in parliament or the executive are less relevant for the news media than large catch-all parties. Incumbency in itself is a news value that provides a certain newsworthiness to incumbent actors and their statements. Accordingly, established populist parties might have better chances to pass journalistic filters than non-established ones—no matter whether fallacious statements are present or not. Thus, it is again desirable to vary this characteristic by selecting both established and non-established populist actors for the analysis.

2.4 Research Questions and Hypotheses

As elaborated above, there are many overlaps between aspects of populist rhetoric and fallacies that have been discussed theoretically or described qualitatively in specific cases. However, the occurrence and use of fallacies in populist communication has not yet been investigated quantitatively. The first question that is addressed in this study is therefore:

• RQ1: Which fallacies do right-wing populists actually use and to which extent?

Next to this open research question, the paper postulates four hypotheses, which investigate factors on different levels of comparison that may influence the use of fallacies in the communication of populist actors and in the media. First, we want to examine how the use of fallacies is related to the use of populist communication. As mentioned above, several authors have suggested that populist actors tend to use fallacies in their argumentation (Luginbühl 2010; Reisigl 2002; Wodak 2003, 2013). We assume that populist actors are not only more prone to use fallacies but, moreover, that these fallacies are specifically used in combination with populist key messages in an effort to justify or defend these populist claims. Therefore, we expect to find more fallacies in press releases and news articles that also contain populist statements (as opposed to texts not containing populist rhetoric). We therefore formulate the first hypothesis:

H1

Fallacies occur more often in news articles and press releases that contain populist key messages.

Second, we expect to find differences across the various investigated communication channels. Due to the mentioned journalistic selection processes and gatekeeper effects, we expect to find fewer fallacies in newspaper articles than in press releases, whose content is solely decided by the populist parties. This expectation is closely connected to our overall argument that political actors use fallacies as a rhetorical instrument to justify their populist claims.

H2

Press releases contain more fallacies than newspaper articles.

We assume that this difference holds true across different parties. However, established parties might have better chances to pass journalistic filters since they show a higher amount of newsworthiness than non- or less-established ones. Thus, articles about established populist parties such as the Swiss SVP should contain more fallacies than news about non-established populist parties such as the British UKIP. However, this difference should mainly hold true for the news articles (where journalistic filters are present) and less so regarding the press releases by the parties. This leads to the third hypothesis:

H3

Swiss news articles about the SVP contain more fallacies than British news articles about UKIP.

Furthermore, based on Mazzoleni (2008, 2014), who argues that highly commercialized and tabloid media are more prone to populism due to their specific media logic, we also expect the occurrence of fallacies to be higher in tabloid newspapers compared to quality newspapers. Tabloid media are more market oriented and more focused on mass audiences than elite media. Moreover, commercial logic seems to favor a critical stance towards the elite and a focus on “common citizens” (Hallin and Mancini 2004, p. 278; Rooduijn 2014b, p. 730). Following this rationale, we also expect tabloid media to be more susceptible to adopt informal fallacies. This leads to the fourth hypothesis:

H4

Articles in tabloid media contain more fallacies than articles in quality newspapers.

3 Method

3.1 Sample and Data Collection

We conducted a quantitative content analysis during election campaigns in Switzerland and the United Kingdom, focusing on articles in daily newspapers and press releases from the most successful populist party¹ in each country. The two countries both have prominent populist parties whose political positions are comparable. The two parties under investigation are the Swiss People’s Party (SVP) and the United Kingdom Independence Party (UKIP). Both parties can be considered right-wing and nationalist and both are regularly classified as populist (Ernst et al. 2017; Stanyer et al. 2017). Most importantly, though, they differ in the key party characteristic deduced in the theory section: The SVP is an established party with the highest amount of seats in the Swiss parliament since almost 20 years and is also present in the executive Federal Council, while UKIP holds almost no political mandates and has never had any position in the national executive. Accordingly, these two parties are ideal candidates for the study at hand: They are functional equivalents with regard to many characteristics but differ in a key feature. Such a design of analysis allows differences between the two parties in terms of populism and fallacies to be inferred back to the status of (non-)establishment that they hold.

We collected data 3 months prior to and 2 weeks after each country’s national parliamentary election. This allowed us to analyze populist communication during election campaigns, a time when not only the parties’ interest in communicating with potential voters is particularly high but also the media’s attention to politics (Koopmans 2004, p. 372). The parliament of the United Kingdom was elected on May 7 2015, resulting in a window of analysis reaching from February 7 to May 21. The Swiss federal election took place on October 18 2015, which means that the data collection for Switzerland started on July 18 and ended on November 1.

The press releases were obtained through the party websites. All releases within the periods of investigation were analyzed, resulting in 110 texts from the United Kingdom and 80 from Switzerland. Criteria for the selection of the print media were their reach as well as their editorial line. In order to compare the use of fallacies in quality papers and tabloid media, we collected articles from two quality newspapers and two up to three tabloid newspapers per country (see Table 1).

Table 1 Newspaper sample

Relevant news articles from these print outlets were identified through the Factiva database using a specific search string. We are only interested in the use of fallacies by politicians of the two selected parties and not by journalists. Furthermore, we assume that politicians who communicate more actively on their party’s website will also tend to appear more often in the print media. On the one hand, this assumption helps us to identify and include relevant actors of the respective parties. On the other hand, it allows us to increase the comparability between statements in press releases and news articles. Based on these considerations, an article had to fulfill the following criteria: (a) mention of UKIP respectively SVP, (b) mention of a UKIP respectively SVP politician who had published two or more texts on his or her party’s website in the investigated time period, and (c) include a statement of said politician, be it paraphrased or an actual quote. From this pool of potential articles, a randomized sample of 40 articles was drawn for each newspaper, resulting in 200 articles from the United Kingdom and 160 articles from Switzerland. Overall, the sample comprised of 360 articles and 190 press releases (n = 550).

4 Operationalization

The coding of the data mainly concerned variables related to populist communication and informal fallacies, whose operationalization is elaborated now in more detail. Populist communication was coded based on the four theoretical sub-dimensions. Informal fallacies were coded based on 15 different types. Both, the four dimensions of populist communication and the different types of informal fallacies, were coded individually as dummy variables and are henceforth considered as variables or categories of the quantitative content analysis.

4.1 Populist Communication

Populist communication is operationalized building upon Jagers and Walgrave (2007) and Cranmer (2011), distinguishing four dimensions of populist key messages: The first dimension, people-centrism means that a political actor speaks in the name of the people and claims to defend its will (advocacy), or claims to be accountable to the people and refers to the importance of responding to what is portrayed as the people’s will (accountability). The second dimension is anti-elitism, which refers to criticism against a broadly understood establishment. Third, restoring sovereignty means that a political actor demands popular sovereignty by advocating for direct democratic elements. Finally, the fourth dimension refers to the exclusion of specific social groups and perceiving such a group as a threat. For each of these four aspects, multiple indicators are defined to measure the populist key messages in a given text. Indicators for people-centrism are references to the “common people”, advocacy, and accountability. Anti-elitism is divided into criticism of the political establishment, criticism of the state as an institution, criticism of the political system, and criticism of the media. Indicators for restoring sovereignty are the demand of direct-democratic instruments and references to a popular initiative or referendum as justification for an argument. Finally, exclusion comprises excluding specific social groups and emphasizing the homogeneity of the people by portraying specific social groups as a threat. For each category, we code on the text level whether a given populist key message is present in an article or not. Based on these indicators, a dichotomous populism variable is created indicating whether a text includes at least one indicator of one of the four dimensions of populism.

4.2 Fallacies

We analyze 15 informal fallacy types based on their respective argumentation patterns. In the following, the definition and operationalization for each individual type of fallacy is explained in-depth.

4.2.1 Petitio Principii

Whenever an argumentation follows circular reasoning, meaning the premise that leads to a conclusion is based on the conclusion itself, we speak of petitio principii, also known as “begging the question” (Ikuenobe 2002; Tindale 2007; Walton 1995). In persuasion dialogue, each participant seeks to prove his or her conclusion based on premises that are accepted by the other participants. If the premises are not known or better established than the conclusion to be proved, the argument is not useful (Walton 2008, pp. 64–65). There are different ways to construct such fallacious arguments, but usually they follow the form that “A is true because B is true, and B is true because A is true”. For example, a right-wing populist might argue that there are many criminal immigrants, since many immigrants break the law—which they do because of their criminal nature. In this case, the conclusion is just a rephrased version of the premise that conveys the same message in other words.

4.2.2 Ad Hominem

Attacks on other people—rather than their argument—are in the focus of this fallacy (Walton 2008, p. 170). It contains all attacks by a politician on their opponent that (a) try to damage the integrity of a person or capitalize on a presumably flawed character (“abusive ad hominem”, Walton 2008, p. 171), (b) focus on a supposed inconsistency of action and advocated position of an opponent (“circumstantial ad hominem”, Walton 2008, p. 177), or (c) stress some sort of bias on their opponent’s side that make them unable to decide and speak objectively (“attack on arguer’s impartiality”, Walton 2008, p. 185). Insults and other instances of defamation are not counted as fallacy if they are not directly linked to an argumentation (Tindale 2007; van Eemeren 2015; Walton 1995, 1998; Zurloni and Anolli 2013). For example, a political opponent could argue that the government does not tell the truth about the amount of unemployment, since it fears for its re-election. That is an example of an “attack on the arguer’s impartiality” (Walton 2008, p. 185). If the opponent argued that the president does not tell the truth simply because he is a notorious liar, it’s a direct attack on the president’s character (“abusive ad hominem”, Walton 2008, p. 171).

4.2.3 Ad Consequentiam

This fallacy is coded whenever a conclusion is reached solely based on the (un)desirability of a consequence (e.g., Walton 2008, p. 24). Walton (2008, p. 24) also calls this “argument from consequences”, because it cites foreseeable negative or positive consequences of a proposed action as the premise. The underlying argumentation scheme for this fallacy is an argument, where the conclusion that A should (not) be brought about is based on the premise that “if A is brought about, good (respectively bad) consequences will plausibly occur” (Walton 2008, p. 26). Such arguments are regarded as ad consequentiam fallacy if (a) the probability or plausibility that the cited consequences might occur is weak, (b) there is no evidence presented to support the claim that these consequences will occur if A is brought about, or (c) there are also negative (respectively positive) consequences that need to be taken into account (Walton 2008, pp. 26–27).

4.2.4 Slippery Slope

Slippery slope argumentation schemes are arguments that include a sequence of steps or a chain argument that rationalizes small differences (Walton 2008, pp. 315–316). A characteristic of such an argument is that a first step in a certain direction is described as invariably involving a whole series of small steps that are not to be stopped once the first step is taken and will finally result in a very negative consequence. Walton (2008, p. 22) describes the following example of a slippery slope fallacy: “[…] a proposal to permit legalized abortion in some cases might be criticized by arguing that such a step would lead to loss of respect for human life, which would eventually lead to concentration camps to eliminate people who are not useful to the economy.”

4.2.5 Ad Baculum

This category refers to consequences packaged in threats made by politicians and is often also referred to as an “appeal to force” (Walton 2008, p. 117). This argumentation scheme applies, if a politician uses an open or implicit threat directed at the respondent as premise for a conclusion. Thus, it violates aims and rules of reasonable dialogue (Walton 2008, p. 117). In distinction to ad consequentiam, an ad baculum argument occurs if a threat is specifically targeted to the respondent and the negative consequences will be brought about by the politician who is making the threat (Walton 2008, pp. 121–123). An illustration is the following example: “You will certainly support our reinvestment plan, right? Because otherwise, your job is on the line.”

4.2.6 False Cause

This category covers cases where a politician tries to establish a causal link between events that happen sequentially or at the same time. In that sense, this fallacy has to do with induction and statistical reasoning. The “post hoc ergo propter hoc” fallacy wrongly interprets weak statistical correlations or co-occurrence as strict causal connection (Walton 2008, p. 20). By doing so, the arguer does not provide enough evidence why these co-occurring events should be causally linked or fails to recognize alternative explanations (Tindale 2007; Walton 1995, 2008). Usually, besides a statistical correlation that excludes intervening variables, the two further criteria for causal conclusions are that the cause precedes the effect and that a sound theory must be able to explain the causal link (e.g., Cook and Campbell 1979, p. 6). The classic example for this is the correlation between the decline in stork populations and the drop in human birthrates; while such a correlation can be shown statistically, it does not imply that the stork brings babies, since intervening variables are responsible for the correlation (i.e., both the stork population and the human birthrates decline in urban regions).

4.2.7 Ad Misericordiam

This fallacy includes attempts to capitalize on sympathy or pity. As such, it is an appeal to emotions (Walton 2008, p. 106). Instead of outlining or responding to an argument, politicians try to position themselves in a way that generates these emotions with the audience or readers. This can be used to distract from the real issue, prevent a discussion or obtain approval for their own position (Walton 2008, p. 128). Characteristic for this type of fallacy are e.g. mentioning human suffering or focusing on the fate of individuals (Tindale 2007; Walton 1995, 1997, 2008). For example, a lazy student that fails all exams and does no homework whatsoever might try to convince his teacher to let him pass because he would otherwise have to join the army—thus appealing to the teacher’s pity (Walton 1997, p. 18).

4.2.8 Ad Ignorantiam

The lack of evidence is key for this fallacy. It covers any argumentation that is based on the absence of proof rather than evidence supporting the claim (e.g., Walton 2008, p. 56). There are two ways how this fallacy can be applied: either a position or claim must be true because it has not been shown to be false; or a position or claim is not true because it has not been proved (yet) (Cummings 2015; Tindale 2007; Walton 1996, 2008). An obvious example is the argument that God must exist because no one has ever been able to disprove his existence, despite many people trying to do so (Walton 2008, p. 57).

4.2.9 Faulty Analogy

This fallacy deals with problematic comparisons. This includes proverbial comparisons of apples and oranges, meaning two things are compared to one another despite a weak actual analogy between the two (e.g., Walton 2008, pp. 305–315). Additionally, the fallacy comes into play whenever politicians compare two objects that share one characteristic and thus reason these objects have to be alike in a number of other properties, too. However, these proposed shared characteristics are not as evident as portrayed. A third form of faulty analogy is neglecting important differences between the compared objects (Tindale 2007; Walton 1995, 2008, 2013; Zurloni and Anolli 2013). For example, someone might argue that smoking cigarettes is just like taking arsenic, as both are related to death. The conclusion would be that if someone does not want to take arsenic, then that person should also not smoke (Walton 1995, p. 60).

4.2.10 Hasty Generalization

This covers cases where politicians jump to conclusions. In that sense, it is based on incorrect induction and statistical reasoning (Walton 2008, pp. 246–247). One way to do this is to conclude a rule or principle based on very few examples, neglecting context information or peculiarities of these cases. The other way is to insist that a given rule be applied to all cases, leaving no room for outliers or exceptions (Tindale 2007; Walton 1995, 2008). As such, a “hasty generalization” has to do with insufficient sample sizes (Walton 2008, p. 254). For example, a poll of eight people taken in a pub is not a relevant sample to draw conclusions about public opinion regarding specific topics (Walton 1995, p. 52).

4.2.11 Straw Man

This fallacy refers to arguments in which politicians depict their opponent’s position in an abusive way in order to make it look preposterous or even hazardous (Walton 2008). It is also coded if a politician outlines only part of the opposing argument, makes their case against it and concludes that the position as a whole has been disproved. Finally, we also count as straw man instances where an opponent’s position has been outlined wrongly by politicians, if there is a correction—either by the opponent themselves or a journalist—within the same article (Tindale 2007; Walton 1995, 2008, 2013). For example, a politician might try to argue against moderate environmental protection by suggesting that the proponents of such measures are suffering from technophobia.

4.2.12 Ad Populum

This fallacy covers instances of attempted reinforcement of political claims by referring to the fact that something is very popular, or the will of the people. In that sense, it is an appeal to emotions (Walton 2008, pp. 107–108). For coding the newspaper articles, it is crucial that the alleged will cannot be verified within the boundaries of the given text, i.e. it contains no survey information on the specific subject. Given that Switzerland regularly holds referenda on specific political issues, they are not counted as absent evidence of the will of the people (Tindale 2007; Walton 1995, 2008). An argument from popularity follows the form that “A must be true, because everybody accepts that A is true” (Walton 2008, p. 113).

4.2.13 Ad Verecundiam

This category focuses on the wrongful or inappropriate use of expert opinions (“appeal to authority”, Walton 2008, pp. 223–224). It is coded when (a) there is doubt that the quoted expert really is an expert in the respective subject or field [e.g., citing Charles Darwin in a discourse about morality (Walton, 1995, p. 46)]; (b) the expert’s opinion is quoted wrongly or partially and thus potentially distorted; or (c) there is no evidence who the quoted expert is. The latter example also includes citing figures from studies and statistics that remain without a source, thus prohibiting verification (Walton 1995, 2008). Note that citing experts is not fallacious per se—it can be legitimate if the expert opinion is relevant and reasonable for the subject at hand (Walton 2008, p. 211).

4.2.14 Many Questions

This applies if a question contains one or more assumptions that would have to be accepted by the respondent in the event of a direct response and could have a negative effect on him (Walton 2008). Central to the fallacy is that the premise within the question is presented in a way that makes it look as if it were already confirmed—though in reality it is not—and is clearly to the disadvantage of the politician’s opponent. In principle, this is a trap for the opponent, because no matter how he or she responds, he or she is immediately in trouble (Tindale 2007; Walton 2008). The most famous example for this type of fallacy is the question: “Have you stopped abusing your spouse?” (e.g., Walton 2008, pp. 46–50); no matter whether the respondent answers “yes” or “no”, he confirms that he does abuse or has abused his spouse (Walton 2008, p. 40).

4.2.15 Red Herring

This fallacy covers deviation from the topic at hand. In that sense, this fallacy describes a rhetorical diversionary tactic (“smoke screen”) that aims at distracting the discursive opponent (e.g., Walton 2008, p. 97). It applies whenever politicians switch the subject without a clear or outlined link to the original topic. This fallacy is also coded when politicians dodge questions of journalists without giving a reason for it (Tindale 2007; Walton 2008). This is a very common type of fallacy and there are many examples. A simple example is an employer who argues against wage raises by stressing how hard he works to make an outstanding product.

Each listed fallacy is individually measured using a dichotomous variable. Based on these variables, a dichotomous fallacy variable was created that indicates whether at least one fallacy is present in a text. This approach is preferred to an additive index because only 8.0% (n = 44) of all texts contain more than one fallacy and only 2.4% (n = 13) more than two.

4.3 Reliability

Two extensively trained coders coded all material by hand with regard to the variables related to populism and informal fallacies. The validity and reliability of the codebook were ensured in a number of steps. First, to warrant the validity, the codebook is strongly based on theory and all items and their theoretical foundation were elaborately discussed with the two coders in a thorough training. Second, the codebook was reviewed and revised based on two pre-tests. After each pre-test, critical variables were discussed using concrete examples. The results of these discussions were included in the codebook as additional explanations and instructions. Finally, an inter-coder reliability test was conducted based on 17 press releases and 17 news articles (n = 34). The reliability for the relevant populism and fallacy variables is overall satisfactorily high. The average Brennan and Prediger’s kappa across all variables is .81 (see Table 2). Although these reliability scores are satisfactory, remaining uncertainties regarding individual variables were again discussed. The largest source of error for the anti-state and anti-politics variables were multiple occurrences within an article. Therefore, additional instructions were agreed upon for these cases. With regard to the ad hominem fallacy, the greatest difficulty was the differentiation from mere insults. If for example a populist called someone a liar, it was not always clear if this was a general defamation or if the accusation referred to a concrete situation or issue. To ensure reliability, we decided to code all accusations in this regard as fallacy, if they were not supported by evidence and if they were used to invalidate an argument from the opponent.

Table 2 Reliability scores

5 Findings

5.1 Descriptive Results and Qualitative Examples

To answer the first research question of which fallacies populist actors use, we first investigate which fallacies actually occur in the analyzed sample and to what extent. Overall, almost a third (32.5%, n = 179) of all analyzed texts include at least one fallacy. A simple one-sample T Test comparing this value against zero shows that fallacies occur significantly more often than not at all in the investigated texts (t(549) = 16.28, p < .001). Thus, we can conclude that the fallacies found in the analysis at hand do not arise randomly. The 95% confidence interval shows that the probability is 95% for the value to be between 28.6 and 36.5% in the population. This means that in about a third of all analyzed texts, the investigated right-wing populists argue incorrectly according to argumentation theory.

However, the distribution varies greatly between the different fallacy types. As Table 3 shows, ad consequentiam fallacies are used the most, followed by ad hominem, ad populum, and hasty generalization fallacies. petitio principii, in contrast, did not occur at all in the analyzed press releases and news articles. Thus, with regard to our first research question, it is important to note that while the use of fallacies is quite extensive overall, the different types of fallacies are not applied equally.

Table 3 Distribution of fallacies

Before statistically testing our hypotheses, let us have a closer look at some actual examples of fallacies found in the newspaper articles in order to further disentangle the general research question. First, every twelfth article or press release contains ad consequentiam fallacies (n = 46). It seems to be a rather common type of argument in political discourse during election campaigns. For example, in a press release on Oct 13, 2015, the SVP’s Toni Brunner argues:

So, what is the advantage for you, dear voters, what is the advantage for Switzerland if you choose SVP on Oct 18 [the polling day]? As a citizen, a Switzerland that is intact and worthy to live in; […] as inhabitants of our country, more security, and fewer burglaries and thefts.

The premise (“vote for SVP”) is only supported because of supposed positive consequences of the conclusion that are not directly connected or supported by evidence.

Second, ad hominem attacks are rather frequent, too—5.8% of all coded articles show such attacks (n = 32). Attacking the political opponent is a typical strategy in election campaigns, thus this result is sensible and intuitive. It also relates to the populist dimension of anti-elitism. An example by UKIP’s Nigel Farage is found in the Sun (April 16, 2015):

Mr. Farage […] said that anyone calling his party racist were ‘part of the establishment’.

It is a direct attack on the accusers identifying racist tendencies in UKIP, rather than an actual dispute about whether there are such tendencies or not.

Third, the ad populum fallacy is present in 5.6% of all articles (n =31). Considering that politicians—and especially populists—like to evoke the will of the people and popular opinions, this is a comprehensible finding. For example, in a UKIP press release on April 3, 2015, Nathan Gill MEP, UKIP’s development spokesman, said:

It is clear that the people of the UK are fed up of salving the consciences of the political establishment with their money. Everybody has the right to give to charitable organisations and most of us do. But to have the EU pick our pockets to give money to schemes that are regularly shown to be corrupt and without any serious benefit is outrageous.

The conclusion that the EU supports corrupt schemes is in no way supported by the premise that the British people are fed up with the political establishment.

Fourth, hasty generalizations are used in every twentieth article or press release (n =29). Given that generalizations based on small, unrepresentative samples are frequent in natural language, it is no surprise that politicians use such rhetorical tricks in election campaigns (whether intentionally or not). Consider the following example found in a Daily Mirror article on March 3, 2015:

UKIP health spokeswoman Louise Bours said the re-validation tests do not work. She added: ‘Harold Shipman would have passed re-validation with no problems at all.’

Re-validation tests in the United Kingdom are tests for doctors and nurses that check whether their medical skills are up-to-date; Harold Shipman was a general practitioner in the United Kingdom who was convicted of mass murder. By generalizing that he would have passed re-validation, and that thus the whole process of re-validation is invalid, Louise Bours commits the fallacy of broad conclusions based on a very narrow sample.

Fifth, the straw man fallacy is used in 4.4% of all instances (n =24). Since misrepresenting the position of a political opponent could be considered a rather popular rhetorical strategy among politicians, it is somewhat surprising that this fallacy does not occur more often. In a press release on September 18, 2015, the Swiss SVP provides a prime example of such a strategy:

Federal Councilor Sommaruga concluded that ‘we want to show that we are in solidarity with Europe’ at today’s media conference. This is exactly the point: Once again, the Federal Council just wants to please the EU. The SVP has no understanding for this kind of policy.

Councilor Sommaruga’s statement referred to the redistribution of refugees across Europe and the SVP statement deviates from that argument by misrepresenting the intention of the Federal Council.

Sixth, the ad ignorantiam fallacy is present in 3.8% of all articles or press releases (n = 21). UKIP Economics Spokesman Patrick O’Flynn provides an example of this fallacy in a UKIP press release (April 28, 2015) that compares UKIP tax policies to those offered by Labour: Trying to position UKIP as the party representing the working class, O’Flynn is paraphrased arguing that Labour offers working people a “total insult” tax policy because they never said that they would raise the tax free personal allowance. In other words, this single part of tax policies that Labour has not made a statement on, leads him to draw a conclusion about their overall tax policy. Mr. O’Flynn focuses on the absence of any pledge by Labour for a specific aspect of tax policies rather than providing evidence with regard to Labour’s overall tax policy.

Seventh, ad verecundiam fallacies appear in 3.5% of all instances (n =19). Given that cited expert opinions are often used in political discourse, it is startling that such opinions are not cited fallaciously more often. Consider, for example, the following statement by UKIP’s Deputy Leader Paul Nuttall (UKIP press release, March 11, 2015):

It is estimated that counterfeit tobacco products will cost the Treasury billions a year and at the same time increase the risk to smokers. Monitoring by retail and manufacturing bodies will also become unreliable because of the black market.

This is a typical example of a vague reference to some expert or study, without further specifying the actual source (and thus the credibility) of the statement.

Eight, the fallacy of ad baculum occurs in 2.5% (n = 14). A Daily Mirror article (May 6, 2015) reports on derogatory comments made by UKIP candidate Robert Blay. Additionally to claiming that his rivaling Tory candidate Ranil Jayawardena was “not British enough” to be in Parliament, he included this threat:

If this lad turns up to be our Prime Minister I’ll personally put a bullet in him.

This is a drastic and exaggerated appeal to threat that is not directly related to the election campaign or Mr. Jayawardena’s suitability for office. It must be acknowledged that such extreme threats are rare in politician’s public speeches. Furthermore, although this statement is a clear threat and Mr. Blay states that he would bring about the negative consequences himself, the argument is not addressed directly at the respondent. However, as our material showed, most threats in news articles or press releases are addressed indirectly to the recipient due to the mediated form of discourse. This is and may explain why adbaculum fallacies are rather rare in comparison to more general threats that fall into the category of ad consequentiam.

Ninth, the false analogy is present in 1.3% of all texts (n = 7). Considering that analogies are a common strategy in political and other persuasive discourses, it is surprising that this fallacy does not occur more often. To provide an example out of our content analysis material, let us have a look at the following statement by the SVP’s Adrian Amstutz (Tagesanzeiger interview, October 16, 2015). He argues that foreign aid should be used to help directly on-site, but that the overall budget must not be expanded:

Mr Glättli [MEP for the Green Party] wants to leave the development aid cake untouched and speak in favor of millions more for emergency aid. I think we should use part of the existing foreign aid budget of a whopping three billion Francs for aiding the countries in crisis. Every fire-brigade commander pulls his forces together in a blaze, instead of spreading them out for combating small bush-fires.

He compares the refugee situation to a blaze and suggests that foreign aid money must now be re-allocated, just as a commander would consolidate his forces. This comparison is problematic for several reasons. First, while a blaze is a temporary and one-dimensional problem that can be solved comparatively simply, the refugee situation has many causes, drivers, catalyzers, and characteristics that demand simultaneous and diverse attendance. While a blaze is concentrated in time and space, Mr. Amstutz admits that there are several different countries in crisis at the root of the refugee crisis. Second, foreign aid has many aims other than “combatting” the refugee situation. Thus, the analogy neglects important differences between the compared objects. Third, while Mr. Amstutz uses this analogy to argue for a re-allocation of the foreign aid budget, the same analogy could be used as an argument for additional funds. This is, however, what he is arguing against. In Mr. Amstutz’s defense, figurative analogies are usually simplified and more pointed.

Tenth, red herring fallacies occur in 1.1% of all texts (n =6). Out of these six occurrences, a prime example is provided by Nigel Farage (interview in the Sun, April 12, 2015). The journalist asks: “Why do you think single mothers and disabled women are bearing the brunt of cuts?”, to which Mr. Farage replies:

It’s unacceptable. There is enough money to go round, but spending in this country is all wrong. The gap between rich and poor has grown, not shrunk. David Cameron keeps talking about the ‘recovery’, but for most people around this country, there is little sign of a recovery and we are not helping the most vulnerable. Instead we are giving away £5 million every day to the EU and around £1 billion a month to foreign aid, much of which is intercepted by dictators.

He clearly deviates from the question and topic of single mothers and disabled women to argue against government spending and the EU.

Eleventh, the false cause (or post hoc ergo propter hoc) only appears in 0.5% (n = 3) instances. In a press release on April 16, 2015, UKIP’s Tim Aker responds to Thurrock’s Conservative MP Jackie Doyle-Price argument that increasing foreign aid helps to cut immigration:

We’ve seen foreign aid and mass immigration both increase under this Tory government, so this incoherent argument from Thurrock’s Tory MP once again demonstrates how badly out of touch she is.

The argument he tries to refute is that foreign aid should help to cut immigration by arguing that both aspects have risen under the Tory government. However, no conclusions about any true causal relations between the two aspects can be drawn on the basis of this statement, since they simply co-occurred during that time—but correlation does not mean causation. There is no clear temporal lag between cause and effect in this argument and Mr. Aker fails to further clarify the theoretical link between these two phenomena. Furthermore, it is not directly related to his actual argument that one should vote for him:

In a few weeks voters can cast their verdict on Conservative and Labour levels of Mass immigration and foreign aid giveaways by voting UKIP and having an MP who puts local people first, not those abroad.

Twelfth, the many questions fallacy also only occurs in 0.5% (n =3). The Swiss SVP provides this example of a classic rhetorical trap in a press release (September 2, 2015) discussing the issue of international cooperation to combat tax evasion and banking secrecy. The press release discusses negotiations about international administrative assistance to exchange banking data and provides this conclusion:

Did the Federal Council and the other parties – except the SVP – get pulled over the barrel yet another time here?

The premise in the question (that the Swiss executive and all parties except the SVP have been pulled over the barrel in the past) is presented in a way that makes it look like it has already been confirmed. The question cannot be answered without confirming this premise—if negated, the answer still confirms that this has happened in the past.

Furthermore, the fallacy of a slippery slope occurs even less frequently, namely only in two instances arguments were found that indicate a slippery slope. In one article (article in 20 Minuten, October 1, 2015) about a planned popular initiative to ban the burqa in Switzerland, SVP MP Sebastian Frehner argues in favor of a prohibition to wear a burqa that

if clothing has a negative impact on integration and thus on society in a country, then it is time to act. ‘Women wearing a burqa do not find jobs here. They don’t learn our language, hence cannot integrate and then get dependent on social welfare.’

Typically for slippery slope arguments, there is some legitimate room for attack and defense that make the individual steps seem reasonable (Walton 2008, pp. 316–317). For example, women wearing a burqa may actually have a lower probability of finding a job, and if someone does not find a job, they have a higher probability of getting dependent on social welfare. However, also typical for a slippery slope argument are the supposed inevitability of the sequence and the exaggeration of the consequence that eventually states that women wearing burqas have a negative impact on integration and society as a whole in a country. This connection does seem overstated considering that according to a report by the Swiss Federal Council² only a handful of women in Switzerland actually wear a burqa and most of them are tourists from the Gulf region or wives of the diplomatic corps.

The ad misericordiam fallacy also only occurs twice. Again, we can turn to the SVP for an example. In a Blick article (October 2, 2015) about a panel discussion, Roger Köppel, SVP candidate and editor of the weekly news magazine “Die Weltwoche”, argues that

[…] he actually does not have any free time, he wants to continue with his newspaper and see his young family. ‘I need to write “Daddy loves you” on my election poster ads and hang them up at home, in order for my children to still recognize me at all.’

This is clearly an appeal to sympathy for his family and the respective values. Finally, the fallacy of petitio principii does not occur at all in our content analysis.

6 Hypotheses

After providing these real-world examples out of our content analysis raw material, we turn to the quantitative tests of our hypotheses. With regard to the first hypothesis, we investigate the relation between the use of fallacies and the use of populist key messages. Hypothesis 1 postulates that more fallacies are expected in texts that also contain populist statements. This complies with the overall assumption that fallacies are used to justify or defend populist messages. The findings clearly show that fallacies are more likely to occur in texts in which populist key messages occur, too (φ_c= .168, p = 0.001). Table 4 compares the contingencies of the two dummy variables for populism and fallacies: It lists the numbers of texts that show both fallacies and populist messages, only one or the other, or none. As such, the contingency table aids to answer the main research question by providing further insights about the relationship between fallacies and populist statements: While fallacies occur in about 15% of texts without populist statements, the value for texts with populist statements is more than twice as high (36.1% of texts with fallacies show a populist key message). Thus, while populist key messages are often used without fallacious arguments, they are significantly more likely to occur together. An example for the co-occurrence can be found in a Daily Mail article (February 12, 2015). Nigel Farage argues with regard to UKIP’s voters:

Table 4 Fallacies by populism

I actually think people are looking for the right type of people – they are sick of the college kids of the political class who are in hock to the big corporates. We are different. We are unashamed to be patriotic.

This anti-elitist and people-centrist statement is also an ad hominemattack on the political class, as well as an ad populum fallacy, arguing that something is bad because many people believe so. As a second example, a UKIP press release (May 18, 2015) quotes UKIP health spokesman, MEP Louise Bours, as reacting to Tory plans regarding the National Health Service (NHS), the public health services in the UK:

The Tories want people to forget they have been in power for five years already and almost decimated the NHS in that time. Cameron can offer as many extra GPs as he likes for services at weekends, but without a plan like UKIP has for recruiting the extra GPs, the promise is meaningless.

It is a populist appeal to the people with traces of calls for popular sovereignty, as well as an ad consequentiam fallacy rejecting the PM’s plan solely on the basis of projected bad consequences. These examples further show that populist actors actually use fallacies in combination with populist key messages in their argumentation.

Hypothesis 2—differences between press releases and news articles—and Hypothesis 3—differences between established and non-established parties—are analyzed jointly in the following. Figure 1 plots the mean values of the fallacy index for news articles respectively press releases in the two countries. These values correspond to the percentages of texts that contain at least one fallacy. As the figure shows, fallacies appear more often in press releases than in news articles across both countries. Furthermore, fallacies are used more often in Switzerland, regardless of the channel.

Fig. 1

Fallacies by country and communication channel

To test whether these differences are statistically significant, a two-factor analysis of variance (ANOVA) is conducted with the fallacy index as dependent variable and country and communication channel (press release vs. news article) as independent factors. ANOVAs are a simple means to compare the internal variance within specific groups to the external variance between the groups. It is thus a statistical way of comparing arithmetic means of specific groups to each other (adhering to the variance and numbers of cases at hand) and identifying statistically significant (i.e., not randomly occurring) differences. On the one hand, this confirms that fallacies occur significantly more often in parties’ press releases (41.1%) than news articles (28.1%; F(1, 550) = 10.04, p < .01, η² = .018). On the other hand, this reveals that fallacies are used significantly more often in Switzerland (40.4%) than in the United Kingdom (26.5%; F(1, 550) = 11.28, p < .01, η² = .020). The interaction between country and channel, however, is not significant (F(1, 550) = 0.03, ns). This means that both Hypothesis 2 and 3 can (partially) be confirmed: Press releases contain more fallacies than news articles (regardless of the country) and Swiss texts contain more fallacies than British ones (regardless of whether they are press releases or news articles). Thus, with regard to the third hypothesis, there is some restraint: As the country effect is not restricted to news articles, it seems that this difference is due to an overall higher use of fallacies by the Swiss SVP. This means that the SVP as an established party not only passes the journalistic filters more easily than the UKIP (despite the use of fallacies), but that they also use more fallacies in their press releases.

Finally, to answer Hypothesis 4, the arithmetic means of the fallacy index are compared between tabloid and quality media. A single factor ANOVA shows that the difference between the two groups is not significant (F(1, 360) = 2.08, ns). Moreover, contrary to our hypothesis, on average there are more fallacies in quality media (M = 31.9%) compared to tabloid media (M = 25.0%). While not significant, this could be attributed to the sample of the study: Since only statements by UKIP and SVP politicians are coded, quality papers might use such statements more often in order to debunk the populist statements (this train of thought is further elaborated in the conclusion). Nevertheless, Hypothesis 4 must be discarded.

Taken together, the analysis has revealed several key findings. Firstly, it is interesting to see which fallacies are used more often and which less often. For some fallacies, it is surprising that they do not appear more frequently, e.g., false analogies or red herrings. Furthermore, fallacies and populist statements significantly co-occur. Finally, the amount of fallacies differs between the investigated parties (the SVP uses more fallacies than the UKIP) and between party press releases and news articles (more fallacies in press releases), but not between tabloid and up-market newspapers. In the following, these findings are further discussed and contextualized.

7 Discussion and Conclusion

The aim of this article is to investigate the use of informal fallacies by right-wing populist politicians (in press releases) and their representation in the press during election campaigns. On the one hand, we are interested in what fallacies populist actors use and how fallacies and populist communication are related. On the other hand, we investigate the influences of specific characteristics of the party and media context on the usage of fallacies by populist actors.

First, our investigation shows that populist politicians generally use fallacies extensively. Almost a third of all analyzed texts include at least one fallacious argument. This is a telling result by itself that empirically supports the often-made assertion that populist actors have a tendency to use fallacious reasoning. However, the occurrence varies greatly between the different fallacy types. On the one hand, as might be expected, populists often resort to fallacy types that are theoretically closely connected to populism such as the ad populum fallacy. On the other hand, even more often they draw on fallacies such as ad consequentiam or ad hominem that are described in the literature as typical and specifically powerful for political deliberations in general (Walton 2008, p. 24/174). Thus, populist actors seem to use a wide range of different fallacies that correspond to different aspects of a persuasive discourse.

With regard to our first hypothesis, the findings confirm that fallacies occur overwhelmingly in combination with populist key messages. The amount of fallacies in texts that also show populism is more than twice as high as the amount of fallacies in texts without any populist statements. Moreover, textual examples from our corpus show that populist actors actually use informal fallacies coupled with populist key messages within the same argument. These findings confirm the theoretical assumption that there is a close relation between populist communication and the use of informal fallacies.

With regard to contextual factors, the results support our hypothesis that fallacies appear more often in press releases than in news articles in both countries. This supports the notion that populist politicians use fallacies as strategic rhetorical means in their own communication channels, while journalistic selection processes and gatekeeper effects attenuate their occurrence in the print media. Furthermore, while we expected more fallacies in Swiss news articles than in the British print media, our findings reveal that the SVP also uses more fallacies than UKIP in their press releases. Thus, the differences between the two parties may not only be explained by the degree of establishment that a party shows. We expected that more established parties should be more newsworthy than non-established ones, and thus pass the journalistic gatekeepers more easily. However, this difference not only manifests itself in news articles but also with regard to the press releases to begin with. A possible explanation could be that the high newsworthiness of the SVP reciprocally reinforces their boldness in using fallacies in their press releases in an interdependent process. However, this would need further investigation.

Furthermore, no significant differences were found between tabloid and quality media, indicating that the editorial line does not influence how often fallacies occur in news articles. This holds true across both parties (SVP and UKIP), as well. This is reminiscent of other studies on populism, which found that tabloid media are not more populist than quality media (Bos et al. 2010; Rooduijn 2014b). However, this finding may be due to the sampling of the study at hand: Since only statements by the right-wing populist UKIP and SVP politicians are included in the data, quality papers—especially left-leaning papers—may quote the respective politicians with the intention to ‘debunk’ their statements. Such instances of journalists trying to expose fallacies were also found in the analyzed material—mostly in up-market newspapers. For example, in a Guardian article (April 18, 2015), Nigel Farage is quoted as stating that “foreigners with HIV cost the NHS up to £25’000 per year and that 60% of people diagnosed with the virus every year in the UK were born abroad.” The Guardian journalist contrasts this statement with a counterargument by Allan Anderson of the HIV charity ‘Positively UK’ that “Farage’s figures are wrong and that he was fueling ‘misinformation and stigma’ surrounding people who deserved support.” Furthermore, the journalist amplifies this refutation of Mr. Farage’s statement by quoting two additional sources, ‘Public Health England’ (PHE, an executive agency of the Department of Health and Social Care) and ‘Avert’ (a global charity for information and education on HIV and AIDS): “PHE’s latest figures show that 54% of people in the UK with HIV were born abroad and the cost of treatment is put at £18,000 per year by the charity Avert.” This is an example of a journalist, who by quoting other sources, attempts to debunk what seems to be a fallacious claim that includes both hasty generalization and ad verecundiam fallacies. It demonstrates that the media do not necessarily disseminate apparent fallacies uncritically but that they may also reveal and criticize them. This would be a very interesting aspect for future research.

The study at hand shows certain limitations. First, only two right-wing populist parties are included in the sample. Thus, our analysis only includes statements by actors of these parties and neglects potential fallacies by politicians of other parties, journalists, or other actors. Although populism is mainly an issue of the political right in Europe, it would be interesting to see if the results also hold true for left-wing populist actors. Furthermore, in order to determine clearly whether populist actors are more likely to use fallacies than other politicians are, it would be necessary to compare so-called “populist” parties to “non-populist” parties as a baseline. Finally, since our study is among the first to jointly investigate populism and fallacies in a quantitative analysis, our operationalization of the individual fallacy types could be further improved and our approach remains in great part descriptive. Hence, we see our study as a starting point for more explanatory investigations on the use of fallacies by politicians, possibly comparing more parties, countries, and different communication channels.

Nevertheless, our results provide an interesting first insight into how right-wing populist actors use informal fallacies in press releases during election campaigns and how these fallacious statements are disseminated by the news media. Moreover, our study provides both quantitative and qualitative empirical evidence that reinforces the popular notion that populist ideology and fallacious argumentation are often combined. This is not trivial, given that politicians such as Donald Trump are often accused of resorting to fallacious reasoning. This is highly problematic from the point of view of democratic theory, since the use of fallacies renders a critical discourse and thus political deliberation more difficult.