Master Thesis
Persuasion in online populist communication: the role of emotions and personality
Graduate School of Communication Master Program Communication Science
Submission: 31.01.2020 Word count: 8200 Marion Molliet 12198188 molliet.m@gmail.com Supervisor: Dr. Linda Bos
Abstract
Populist communication is particularly persuasive given its emotional content. However, this relationship has not been tested in an online environment. The present thesis investigates whether emotional populist Facebook appeals positively impact populist attitudes. Based on the appraisal theory of emotions the study focusses on the role of anger and hope. Further, the importance of personality traits for the formation of political attitudes is taken into account by examining the role of Agreeableness. Using an online experiment conducted amongst Swiss residents (N=227), the findings point to the crucial role of anger. While being in an angry emotional state and exposed to anger cues does not influence populist attitudes, there is a connection between being an angry individual and having strong populist attitudes. Further, the results show that the populist attitudes amongst low agreeable individuals are boosted when exposed to a populist appeal. These findings suggest that personality and emotional traits play an important role in the formation of populist attitudes, whereas emotional states and cues do not explain the persuasiveness of the populist message.
Introduction
All across the world, populist parties have achieved notable successes at the ballot box. This is no longer an extraordinary event, but rather, part of daily politics: expressing right-wing populist ideas is now a “mainstream” (Mudde, 2019, p. 164) phenomenon. The situation is no different in Switzerland, referred to as a “populist paradise” (2008, p. 100) by Albertazzi. Its direct-democratic system makes it a promising place for populists because it allows them to put popular sovereignty (Mudde, 2004) into practice and push for a vote on topics that are not necessarily on the agenda amongst the political elite.
Multiple right-wing populist parties are present in Switzerland. Although the Swiss People’s Party suffered significant losses in the 2019 election, it still holds 26.5% of the seats in the lower chamber of parliament (Office fédéral de la Statistique [OFS], 2019a) – making it the first party of the country. Other smaller right-wing populist parties like the Ticino League or the Geneva Citizens Movement are also well established regionally. Understanding why right-wing populist ideas are so popular is therefore of growing interest in general, and in Switzerland in particular.
The success of those parties can in part be explained by their effective communication. Being exposed to populist messages in the media environment can impact citizens’ attitudes and make populist ideas more accessible and popular (Andreadis et al., 2019). In this
perspective, social media has affordances that are of great potential for populist
communication. It enables politicians to avoid the gatekeeping function of traditional media (Engesser et al., 2017), spread their ideas without barriers (Krämer, 2017) and communicate directly with citizens. It also allows a growing number of individuals, politicians, and citizens alike, to correspond with others all around the world and share points of view extremely quickly, and in a potentially anonymous manner, building “gathering spaces” for “otherwise dispersed people” (Gerbaudo, 2018, p. 750).
In this context, emotions contribute to the persuasiveness of populist communication because emotionally-charged messages tend to be shared more than neutral ones on social media (Stieglitz & Dang-Xuan, 2013). Further, emotional content provokes more attention and is therefore better internalized; it is more convincing than neutral content, and impacts opinions (Forgas cited in Stieglitz & Dang-Xuan, 2013, p. 222). In the political realm,
emotions can be strategically used to influence viewpoints (Brader, 2005). This is particularly suitable to populist communication as it is already inherently emotional (Ernst et al., 2019) through its “Manichean” view of the world which focusses on the opposition of “good and evil” (Wirz, 2018, p. 1117).
This being said, emotions do not necessarily have the same persuasive power on everyone. Personality traits could influence the effect of populist messages. In that respect, individuals scoring low on the personality trait Agreeableness are particularly drawn to right-wing populism (Ackermann et al., 2018; Bakker et al., 2016, 2020). Whether this means that online populist messages resonate more amongst this group remains an open question. Consequently, the following questions arise:
RQ1: To what extent can the effects of populist communication on populist attitudes be explained by emotions on social media? RQ2: Is the impact of populist communication on social media moderated by the level of Agreeableness?
Whilst the coverage and communication strategies of populist parties, as well as their influence on the electorate are well researched, the potential of social media should be further investigated. Given that sites like Facebook, Twitter and Instagram have a growing number of users, as well as affordances that suit populist communication particularly well (Gerbaudo, 2018), their potential for populist ideas must be the focus of more research. Usually studies in this field gather existing real world data, such as tweets, to analyse them (see Engesser et al., 2017; Ernst et al., 2017; Thiele, 2019) or present participants with a fixed fake content
mimicking what could be found on social media platforms (Hameleers & Schmuck, 2017). The former does not allow researchers to causally test the impact of specific type of messages and manipulate their emotional content, and the latter does not account for the fact that social media users can to a large part select what they want to see on those platforms. Moreover, while the impact of emotions for populist communication has been demonstrated (Wirz, 2018), this was done using an unusual issue in the context of right-wing populism, leading to limited external validity.
This thesis takes these considerations into account by testing not only the emotions that a populist Facebook post can trigger, but also the impact that being in a particular emotional state when seeing such a post can have. This was done using the subject of
increasing healthcare costs, which can be adapted in a convincing manner in the case of right-wing populism. Additionally, the experiment includes an interactive element, as participants could select a person who could be their Facebook friend, which is reflected in the stimuli.
Theoretical framework The characteristics of populism and populist communication
Populism is generally defined as a “thin-centred ideology” (Mudde, 2004, p. 544), lacking core values, and usually combined with a host ideology (Akkerman et al., 2014) that can be left- as well as right-wing. It opposes “the pure people” to a dishonest elite acting to the detriment of the former (Mudde, 2004, p. 543). For the supporters of populism, politics should be an expression of the “volonté générale of the people” (Mudde, 2004, p. 543), not under control of disconnected elites. On the right side of the political spectrum, populism is “exclusionary and identity-focussed” (Stavrakakis et al., 2017, p. 458) concretely denying “certain outgroups the right to be counted among the people” (Krämer, 2017, p. 1299). As right-wing populism is more prevalent in Europe in general and in Switzerland in particular, it
will be the focus of this study. The previously cited characteristics demonstrate the
“Manichean” outlook behind populism, distinguishing merely between “good and bad” (Van Hauwaert et al., 2019, p. 3) without further nuance.
While at first glance, populism is attributed to certain actors or parties it can also be viewed as a “communication phenomenon” focussed on particular “content” and using a distinctive “style” (de Vreese et al., 2018, p. 426). From this perspective, the – populist – way in which a message is communicated is the main point of attention, rather than the actor who engages in it. Therefore, a populist style can also be used by non-populist parties (Bernhard, 2017). De Vreese et al. (2018) identify three crucial components of populist communication: “(1) reference to “the People,” (2) a battle against the “corrupt” elite, and with a possible extension of (3) the identification of an out-group” (p. 427).
A platform for populist communication
To show its effects, populist communication must be publicised. In this context, traditional media have a crucial role to play: not only passing on populist ideas (Esser et al. cited in de Vreese et al., 2018) (populism through the media), but also engaging in so called “media populism” (Krämer, 2014, p. 48) (populism by the media) themselves using a populist style and content. These phenomena have been shown to have a notable impact. The way in which populist party leaders are presented and the volume of coverage in traditional media influences their electoral success (Bos, van der Brug, & de Vreese, 2010). More specifically, an increased visibility in the media can make them seem more “influential” (Bos, van der Brug, & de Vreese, 2011, p. 197). The populist style also impacts the way in which party leaders are perceived: it leads to a more effective image and legitimacy amongst “the lower educated, the politically cynical and the less politically efficacious” (Bos, van der Brug, & de Vreese, 2013, p. 204). Further, populist cues can positively impact attitudes against
The potential of social media
Given the large number of users and the possibility to avoid the gatekeeping function of traditional media (Engesser et al., 2017), social media might be of even greater potential than traditional media. In this vein, Krämer (2017) sees social media as a facilitator to
“communicate publicly, to organise and to mobilize”, as well as a place for “debate” (Krämer, 2017, p. 1304). Furthermore, as Karpf (2017) notes citing Trump’s example, social media can also be efficiently used to “set the agenda of the mainstream media” (p. 3). More specifically, Facebook has particular affordances, such as “higher levels of proximity, reciprocity,
unlimited space for messages, [and a] non-elite character” (Ernst et al., 2017, p. 1358), that are particularly suitable for populists to spread their points of view in an “unfiltered”
(Krämer, 2017, p. 1303) way. As such, it functions well to construct the populist’s imagined “community of ordinary citizens” and exclude out-groups (Hameleers, in press).
Considering the amount of information individuals are subjected to every day, social media can be an important selection tool to assess which news pieces are relevant, trustworthy and worth spending time on through personal relations. These relationships are crucial when determining credibility of sources and “recommendations” (Metzger et al., 2010, p. 425) from relatives are valued amongst online users. Emotions play an important role in this perspective. Stieglitz and Dang-Xuan (2013) showed that on social media, emotional messages have a higher probability of being shared than “neutral ones” (p. 241). This is hypothesized to be linked to the high level of “attention” and “arousal” emotional messages produce (Stieglitz & Dang-Xuan, 2013, p. 218).
In this context, the assessment of relatives is not only useful for the selection of news, but also for the evaluation of its quality – and therefore how persuasive it is. This plays a key role in persuasion because the sources evaluated as trustworthy are significantly more
communication, it has been shown that the persuasive power of messages also hinges on the support for a source: if the source was not considered credible, populist messages had no impact on attitudes (Hameleers & Schmuck, 2017). This is relevant even if individuals do not consciously use social media to see political information, as even those who do not
deliberately use social media for news, are still “incidentally” exposed to it (Fletcher & Nielsen, 2018, p. 2450). All in all, through its affordances, social media can be part of the “self-socialization” (Krämer, 2017, p. 1305) process leading individuals to embrace a right-wing populist “Weltanschauung” and to display it to the world.
Populist attitudes and communication
As Ackerman notes, “populist attitudes represent the extent of agreement with the populist thin ideology” (cited in Wirz, 2018, p. 1119). These attitudes are important since they partly explain the endorsement of populist parties (Van Hauwaert & Van Kessel, 2018). While different measures exist to evaluate these attitudes on the individual level, anti-establishment, favouring the people’s sovereignty and the belief in the homogeneity of the people are commonly accepted elements (Akkerman et al., 2014; Hameleers & de Vreese, 2018; Schulz et al., 2018; Van Hauwaert et al., 2019; Wettstein et al., 2019). The first relates to the antagonism between the corrupt elite and the people, which is at the core of populism (Mudde, 2004). The appeal for the people’s sovereignty lies in the vision that it is a
“homogeneous, pure and virtuous” group forming “the basis of the good society” (Akkerman et al., 2014, p. 13) and therefore the people should decide when state affairs are at stake. Since the people are considered to be a homogeneous group, those who are different – immigrants for example – are thought to be a menace and should be excluded. These elements exemplify the Manichean outlook populism has on society: intrinsically good or bad without further nuance. This approach allows for a precise measure, as attitudes have to feature those three
elements in conjunction with one another to be considered populist. Solely being an advocate of the people’s sovereignty, for example, does not mean that one has populist attitudes.
How can these attitudes vary? Multiple studies have demonstrated that populist communication can impact the attitudes of those who are exposed to it. Hameleers and Schmuck (2017) have shown that assigning blame to elites and immigrants through social media boosts the populist attitudes of those who support the source of the post. Even after exposure to only one populist message Andreadis et al. (2019) found some small, but
significant effects on participants’ populist attitudes in their experiment. Müller et al.’s (2017) finding support those claims but add that the impact of the exposure to populist messages hinges on people’s prior attitudes, and that those messages can have a polarizing effect: individuals with populist attitudes are likely to become more populist, while the contrary happens to those who do not have populist attitudes. This points to a potential backlash effect of populist communication and advocates for further investigation.
The mechanism behind the influence of populist communication is “cognitive priming, the way media present an event or group may activate the audience’s interest in it, influence its perception, and make it cognitively more accessible thereby strengthening its relative weight in decision-making” (Andreadis et al., 2019, ‘Populist Attitudes’, para. 3). Extending this logic to social media, emphasizing different elements of populism on Facebook could impact the attitudes of users, leading to the following hypothesis: (H1) Exposure to populist
appeals on Facebook will lead to stronger populist attitudes than exposure to non-populist appeals.
Emotions, populist communication and persuasion
Emotions can be defined as “internal, mental states representing evaluative, valanced reactions to events, agents, or objects that vary in intensity” (Ortony, Clore, & Collins cited in
Nabi, 1999, p. 295). Following appraisal theories, emotions can be provoked through different situational contexts – Lazarus labels these “core relational themes” (cited in Nabi, 1999, p. 297). Therefore, different situations or messages – which can be fabricated – should bring forward different emotions (Nabi, 1999). In order for a message to successfully bring about a specific emotion, it has to comprise “a core relational theme” and the receiver must view “it to be of personal relevance” (Wirz, 2018, p. 1117). Emotions can then influence the impact of messages in three ways: they can function as “cues”, as a “piece of evidence” or lead to biased thinking (Petty & Briñol, 2015, pp. 3–4) – that is, evaluating something more positively or negatively than it actually is.
Given the underlying emotional component of populist communication through its “Manichean” outlook on the world and its focus on the opposition of “good and evil” (Wirz, 2018, p. 1117) emotions have great potential in this domain. Indeed, Ernst et al. (2019) see “emotionality” (p.10) as one of the key elements of the populist style.
A focus on anger and hope
In order to assess the importance of emotions for populist messages on social media, two emotions are studied: anger and hope. In general, the use of anger has been found to provoke notable increases on political participation (Valentino et al., 2011), racist attitudes (Banks & Valentino, 2012) and to be linked to higher populist attitudes (Rico et al., 2017). More specifically, for populist communication, the significance of this emotion has been demonstrated by various scholars (Hameleers et al., 2017; Rico et al., 2017; Wirz, 2018), but has yet to be refined in an online environment. Though both anger and hope influenced attitudes towards the labour market when messages were presented in a populist style (Wirz, 2018), hope – and positive emotions in general – is widely overlooked in studies on populist communication and emotions. However, hope should not be neglected, as it can be used to
depict a worrisome situation and its potential solutions, which functions well in the context of populism as it is focussed on the oppositions between good and bad (Wirz, 2018)
The “core relational theme” of anger is the perception of “demeaning offenses against oneself or one’s loved” according to Lazarus (cited in Nabi, 1999, p. 297). The impact of anger lies in the connection between the way to appraise it and the blame attribution and confrontation characterizing populist communication (Rico et al., 2017). Speaking about harm inflicted on the people is at the core of populism (Mudde, 2004), and employing anger works particularly well in this context. It can be used to address all sub-dimensions of populism in a credible way by calling out the elite and its behaviour against the people (Wirz, 2018), demanding the sovereignty belonging to the people and pointing out to the differences between the people and other groups, such as immigrants.
In a previous study on populist communication, anger was appraised only by
employing trigger sentences that made use of appraisal theories (Wirz, 2018). Emotions can also be provoked when asking individuals to describe what made them angry (Valentino et al., 2011). In order to enhance the emotional reaction of participants, this thesis combines those two techniques. As a result, the following is hypothesized: (H2) exposure to populist appeals
on Facebook will make users angrier when they are already in an angry emotional state and lead to stronger populist attitudes than when they are in a relaxed emotional state.
For hope, the core theme is “fearing the worst but yearning for better” and therefore, to create hope, “one must appraise a positive future outcome as possible” (Lazarus cited in Chadwick, 2015, p. 599) while the current situation should be alarming. This combination of negative and positive aspects match the Manichean perspective behind populism representing the world in an unnuanced opposition between god and bad (Wirz, 2018). In this context, hope cues can credibly be used to blame the elites for current problems while presenting “the pure people and populist actors” (Wirz, 2018, p. 1120) as the solution. Hope cues should thus
present “an opportunity” as well as “recommended actions to take advantage of the
opportunity” (Chadwick, 2015, p. 602). Curato (2016) showed that hope can be used in an effective way in populist communication – even when combined with fear – as it motivates citizens to take their future into their own hands and act.
As for anger, a combination of two methods eliciting emotions (Valentino et al., 2011; Wirz, 2018) was tested. Consequently, it is hypothesized that: (H3) exposure to populist
appeals on Facebook will make users more hopeful when they are already in a hopeful emotional state and lead to stronger populist attitudes than when they are in a relaxed emotional state.
Agreeableness and populist attitudes
Psychological traits can help to explain political behaviour on the individual level (Gallego & Oberski, 2012; Kalmoe et al., 2018; Kam & Simas, 2012). In the context of
populism, multiple authors found that low agreeable citizens tend to show stronger support for populist parties than high agreeable ones (Ackermann et al., 2018; Bakker et al., 2016, 2020). Low agreeable people have the tendency to be less risk-averse, trusting, cooperative,
“sympathetic and altruistic” and like “conflict” (Mondak & Halperin, 2008, p. 346). This trait is commonly assessed with the Big Five measurement, which has been tested in multiple national contexts and languages (Schmitt, Allik, McCrae & Benet-Martinez cited in Ackermann et al., 2018).
The relationship between Agreeableness and populist support is hypothesized to hinge on the matching characteristics of those two constructs (Bakker et al., 2020): low agreeable citizens are susceptible to the properties of populism for three reasons: first, the conflictual perspective of the world presented by populist does not repel them (Ackermann et al., 2018), in opposition to high agreeable individuals who tend to avoid conflict. Second, low agreeable
people tend to have a more negative view of out-groups compared to high agreeable ones (Turner et al., 2014), which is why exclusionary populism often resonates with them. Third, given the distrust characterizing low agreeable individuals (Mondak & Halperin, 2008), the anti-elitist component of populism also appeals to them (Bakker et al., 2016). It is therefore hypothesized that (H4) exposure to populist appeals on Facebook will lead to stronger
populist attitudes amongst low agreeable individuals than amongst high agreeable ones.
Theoretical model and hypotheses
All things considered, populist appeals are expected to have a direct effect on populist attitudes (H1), as well as an indirect one through the moderated mediation of anger (H2) and hope (H3). This means that the participants who are already in an angry or hopeful emotional state are foreseen to become angrier or more hopeful when exposed to populist appeals, and therefore have stronger populist attitudes. Finally, Agreeableness should moderate the impact of the populist appeal (H4) and results in stronger populist attitudes for those who score low on the trait. Figure 1 depicts the model that is tested in this study.
Figure 1
Moderated mediation model: the influence of populist appeals on attitudes
Methods
To test the hypotheses, a 2 x 3 factorial experiment with Factor Type of appeal as between-subjects variable (two levels, namely: populist and non-populist) and Factor Type of emotion as between-subjects variable (three levels, namely: anger, hope and control) was conducted.
Participants
The participants were current or former Swiss residents above 18 years old. The experiment was not restricted to people with Swiss nationality, in order to keep the potential pool of participants as wide as possible. It was performed in French. Participants were
Populist vs. non-populist appeal Populist attitudes H1 H4 H2 H3 Emotional state hopeful Agreeableness Anger Hope Emotional state angry
recruited trough social media and personal contacts. The data collection period lasted 2.5 weeks in December 2019.
After all incomplete responses were removed, there were 229 participants.1 Two
additional participants were removed because they were not part of the target group. The analyses were conducted with a sample of 227 participants. The mean age was 41.22 years (SD = 15.57). 47.1% of respondents were female, 52.9% were male. In terms of education, 61.3% of respondents had at least a Bachelor – resulting in a highly skewed sample.2
Procedure
The experiment was conducted online. First, informed consent was obtained. Then, participants were asked about their background and the control variables: their political orientation, the party they would vote for, their interest in politics, their Facebook use and issue positions on three topics. One of those was healthcare, in order to assess participants’ views on the issue used for the stimuli. After that, participants were randomly assigned to one condition of the Factor Type of emotion. They completed a 1.5 minutes emotion-inducing task, asking them to describe an event that made them feel angry, hopeful or relaxed (control condition). A first manipulation check asking participants to rate their emotional state was introduced. Next, the images of four women and men (see Appendix A) were presented to the participants. They chose one person who could be their Facebook friend. This step was introduced to reduce bias towards the stimulus and to add an interactive component.
1 While 353 participants started filling in the experiment, only 229 finished it. This was mostly due to the
complexity of the questions requiring that participants write down their thoughts during the emotion-inducing task, as well as answer numerous questions.
2 Since in 2018, only 28.8% of the 25-64 years permanent Swiss residents had a diploma from a higher education
The participants were then randomly exposed to the stimulus: a Facebook post by their potential Facebook friend, with a populist3 or non-populist appeal.4 They had to stay on this
page for at least 30 seconds.
After that, participants completed a second manipulation check concerning their emotional state. Questions measuring populist attitudes followed. At the end of the
questionnaire, the manipulation checks and questions about the post were presented. Finally, participants were thanked and debriefed.
Manipulation of Type of emotion and appeal
Regarding the experimental manipulation of emotions, participants were asked to complete an emotion-inducing task (see Appendix E) where they had to describe in writing something that made them feel angry, hopeful or relaxed (control condition). This method is adapted from Valentino et al.’s (2011) technique, where participants were asked this but in the context of a political campaign. According to the authors, this technique has the advantage of producing specific emotions (Valentino et al., 2011). This method also avoids one of the pitfalls of showing participants an image aiming to stimulate anger, as an image can very likely also lead to fear or other related emotions.
Concerning the appeal, populist or non-populist, participants were exposed to one of two different Facebook posts about the increasing costs of healthcare in Switzerland, one with populist elements and the other without. The subject was chosen because it is not inherently populist or owned by a populist party but can easily be presented in a populist manner. The posts consisted of two fictitious texts of comparable structure and length, as well as a link to a
3 A pre-test among seven French-speaking participants showed that the stimulus was not perceived to be
considerably more populist in the populist condition than in the non-populist one. The populist stimulus was adjusted accordingly, and two questions were added to the manipulation check to capture the variance between conditions more precisely.
4 Regardless of the Facebook friend participants chose they all saw the same stimuli, only the source varied. All
real article about increasing health costs in 2020. Only the headline of the article and the cover image were visible for participants. Le Matin, a popular newspaper in Switzerland not typically pertaining to the left nor right side of the political spectrum, was the source of the article. In both conditions, the texts contained cues expected to evoke anger and hope, according to the appraisal theory (Nabi, 1999) described previously. Two manipulation checks measuring the emotional state of participants were performed: one after the emotion-inducing task and the other after the stimuli.
The text of the stimuli is presented below (see Appendix B for the visuals). The two posts had an identical structure, but differed in the following elements: in the populist
condition phrases blaming the elites and appealing to the sovereignty and homogeneity of the people were present (underlined) – according to the elements of populist attitudes (Wettstein et al., 2019), while in the non-populist condition these parts were removed and replaced with more factual information (in square brackets) (Conseil fédéral, 2019):
“Year after year nothing changes, the health insurance costs increase more and more without Berne even lifting a finger! Meanwhile, the political elites join managing boards of directors of insurances and fill their pockets while [and] the middle class pays. 374,40 francs on average per month per adult is too much! [For a family with two kids that is almost 1000 francs per month.] It is about nothing less than knowing which health care system we want for our country: controlled by the elites in Bern or focussed on the well-being of the Swiss people [out of control or truly focussed on the well-being of the population]. We cannot endorse this system that works at the expenses of the taxpayers, it needs to change and this will not happen if we wait for the elites but it will happen through the people [and solutions exist: lowering the cost of drugs, more prevention to prevent rather than cure diseases, reduce the number of visits to emergency services, giving more value to general practitioner to reduce unnecessary visits to specialists]!”
Measures
Moderator
Agreeableness. This variable was measured using a random subset of four validated
measures by Plaisant, et al. (2010) (see Appendix C for the items wording). Participants rated to what extent they were kind, helpful, looking for quarrel and contemptuous (M = 5.56, SD = .84) on a scale ranging from not at all (1) to completely (7). The last two items had to be recoded, as they were reverse coded. An exploratory factor analysis showed that the scale was unidimensional (one item had an Eigenvalue above one), which explained 48.08% of variance in the four items (see Appendix D). The scale was of low reliability (α =.62) but could not be significantly improved by removing an item. This is in part due to the use of a reduced scale, that does not render the full complexity of this psychological trait (Bakker & Lelkes, 2018).
Dependent variable
Populist attitudes. The variable was measured using three dimensions, each described by four
items (see Appendix C for wordings): anti-elitism, demand for popular sovereignty and homogeneity of the people (Wettstein et al., 2019). Participants indicated the extent to which they agreed with the statements, on a scale ranging from completely disagree (1) to
completely agree (7). An exploratory factor analysis demonstrated that the scale was
three-dimensional, three components had an Eigenvalue above 1, explaining 69.58% of the variance in the 12 original items (see Appendix D). The scale was reliable (M = 3.89, SD = 1.11, α = .86). 5
Randomization checks
Participants were randomly distributed across conditions in the following way: 33 participants in the anger and populist condition, 48 in the anger and non-populist condition, 42 in the hope and populist condition, 30 in the hope and non-populist condition, 36 in the control and populist condition, 38 in the control and non-populist condition.
The randomization across conditions proved to be successful. Eight One-way
ANOVAs with Type of emotion (anger, hope and control) and Type of appeal (populist and non-populist) as independent variables and relevant variables as dependent variables, showed there were no significant differences across conditions for these variables (see Table 1).
Table 1
Summary of ANOVAs
Sum of
Squares df Mean Square F p
Partial Eta Squared Age 1555.07 5 311.01 1.29 .268 .03 Political interest 32.68 5 6.54 1.89 .097 .04 Political orientation 5.794050 5 1.158810 .48 .788 .01 Facebook use 8.375 5 1.675 .42 .836 .01
Friendship with author of
the post 19.578 5 3.916 1.26
.282 .03
Education 25.784 5 5.16 2.09 .068 .05
Agreement with the post 13.23 5 2.65 .97 .439 .02
Opinion health costs 15.33 5 3.07 1.43 .214 .03
*p < .05., **p < .01., ***p < .001.
Two Chi Square tests demonstrated that distribution of gender X2 (5, N = 227) = 3.25,
p = .661 and support for a populist party X2 (5, N = 227) = 3.56, p = .615 were comparable
Manipulation checks
Emotions after the emotion-inducing task
Directly after completing the emotion-inducing task, participants were asked to indicate the extent to which they felt angry (M = 2.57, SD = 1.91), hopeful (M = 4.08, SD = .15) and relaxed6 (M = 4.17, SD = 1.75) on a scale ranging from not at all (1) to very much
(7).7
A One-way ANOVA with Type of emotion (anger vs. hope vs. relaxed) as the independent variable and the first anger manipulation check as the dependent variable was conducted to check if participants were angrier in the anger condition than in the hope and control conditions. A significant effect of Type of emotion on the anger manipulation check was found F (2, 224) = 24.37, p < .001, η2 = .18. Post-hoc Bonferroni test showed that
participants were angrier in the anger (M = 3.64, SD = 2.09) than in the hope (M = 2.11, SD = 1.61, p < .001) and control (M = 1.84, SD = 1.41, p < .001) conditions.
A One-way ANOVA with Type of emotion (anger vs. hope vs. relaxed) as the independent variable and the first hope manipulation check as the dependent variable was conducted to see if participants were more hopeful in the hope condition than in the anger and control conditions. A significant effect of Type of emotion on the hope manipulation check was found F (2, 224) = 6.82, p = .001, η2 = .06. While post-hoc Bonferroni test showed that
participants were more hopeful in the hope (M = 4.49, SD = 1.47) than in the anger condition (M = 3.54, SD = 1.8, p = .002), there was no significant difference between the hope and the control (M = 4.27, SD = 1.69, p = 1) conditions. Conclusively, the manipulation of hope through the emotion-inducing task was not successful.
6 Two other emotions were tapped but they are not relevant here. 7 These are the overall results, not the ones per condition.
Emotions after the exposure to the stimuli
After seeing the stimuli, participants were asked to rate their emotions for a second time, leading to these results: angry (M = 3.15, SD = 1.78), hopeful (M = 3.36, SD = 1.56) and relaxed (M = 3.73, SD = 1.52).8
A One-way ANOVA with Type of emotion (anger vs. hope vs. relaxed) as the
independent variable and the second anger manipulation check as the dependent variable was conducted to investigate if participants were angrier in the anger condition than in the hope and control conditions. This was not the case, F (2, 224) = 2.28, p =.105, η2 = .02. Post-hoc
Bonferroni showed there was no significant difference between the anger (M = 3.48, SD = 1.88), the hope (M = 2.94, SD = 1.77, p = .186) and control (M = 2.97, SD = 1.64, p =.225) conditions.
A One-way ANOVA with Type of emotion (anger vs. hope vs. relaxed) as the independent variable and the second hope manipulation check as the dependent variable was conducted to check if participants were more hopeful in the hope condition than in the anger and control conditions. They were not, F (2, 224) = .72, p =.489, η2 = .01. Post-hoc
Bonferroni showed there was no significant difference between the hope (M = 3.54, SD = 1.63) the anger (M = 3.26, SD = 1.64, p = .794), and control (M = 3.3, SD = 1.39, p =1) conditions. These results suggest that the emotion-inducing task was partly successful at provoking emotions, while the Facebook post was not.
Type of appeal
Participants were asked if they thought the author of the post was explicitly accusing the elites of not acting to reduce health costs (M = 4.78, SD = 1.7), explicitly indicating that the people are the solution to this issue (M = 3.5, SD = 1.75) and explicitly naming the people as the victim of increasing health costs (M = 5.31, SD = 1.68).9 Participants had to indicate
8 These are the overall results, not the ones per condition. 9 These are the overall results, not the ones per condition.
this on a scale ranging from not at all (1) to completely (7). An exploratory factor analysis showed that the scale was unidimensional (only one component had an Eigenvalue above 1), explaining 54.3% of the variance in the three original items (see Appendix D). The scale proved to be of low reliability (α = .57, M = 4.53, SD = 1.25) but could not be substantially improved by deleting items.
An independent-samples t-test was performed to compare Type of appeal in the populist and the non-populist conditions. There was a significant difference in the scores for the populist (M = 4.78, SD = 1.32) and the non-populist (M = 4.29, SD = 1.14) conditions,
t(225)=3.03, p <.01. The manipulation was therefore successful for the Type of appeal, this
means that participants who were exposed to the populist stimuli perceived it to be more populist than those who were exposed to the non-populist stimuli.
Results Main effect on populist attitudes
In order to test the first hypothesis (H1) stating that participants who were exposed to populist Facebook posts were expected to have stronger populist attitudes than those exposed to a non-populist post, an Independent sample t-test was performed. There was no significant difference in the scores for the populist (M = 3.93, SD = 1.03) and the non-populist (M = 3.84,
SD =1.18) conditions, t (225) =.58, p = .566. As a result, H1 was rejected.
Moderated mediation: the impact of emotional state and reaction
The populist post was expected to provoke more anger for users who were already in an angry state, and therefore lead to stronger populist attitudes for angry participants than for relaxed ones (H2). Similarly, the populist post was presumed to lead to higher levels of hope when users were already in a hopeful state and thus result in stronger populist attitudes for
hopeful participants than for relaxed ones (H3). To test these hypotheses, PROCESS model number 7 with 5000 bootstraps samples to estimate the bias corrected bootstrap confidence intervals (BCBCI) was used (Hayes, 2013).
In order to test H2, populist appeal was the independent variable, emotional state anger the moderator, emotional reaction angry the mediator and populist attitudes the dependent variable. Emotional state hope was used as a control. The model predicting emotional reaction was not significant, F = (4,222)= 1.46, p = .215, R2=.03. The interaction
effect of populist appeal and anger state on emotional reaction proved to be insignificant too (b = -.16, p = .756), all other results can be found in Table 2. Thus, H2 was rejected10.
However, the model predicting populist attitudes was significant, F = (3,223) = 6.95, p < .001, R2=.09, and the effect of reaction angry on populist attitudes was also significant (b =
.18, p < .001), which means that a 1-point increase in anger leads to a .18 increase in populist attitudes.
10 As an alternative, a moderation using PROCESS model 1 with emotional state anger as a moderator and a
mediation using PROCESS model 4 with emotional response angry as mediator were tested. They did not yield more significant results.
Table 2
Moderated mediation for predicting emotional reaction angry and populist attitudes
Effects b SE 95% BCBCI t p Constant on reaction angry 2.82 .25 [2.33, 3.31] 11.24 .000*** Appeal on reaction angry .31 .3 [2.33, 3.31] 1.06 .288 State anger on reaction angry .6 .36 [-.11, 1,3] 1.67 .097 State anger * appeal on reaction -.16 .5 [-1.14, .83] -.31 .756 Hope state as control on reaction -.06 .29 [-.64, .52] -.2 .842 Constant on attitudes 3.31 .16 [2.99, 3.63] 20.09 .000*** Reaction angry on attitudes .18 .04 [.1, .26] 4.47 .000*** Appeal on attitudes .06 .14 [-.23, .34] .39 .695 State hope as control on attitudes -.05 .15 [-.35, .25] -.31 .754
Notes. For Type of appeal, 0 = non-populist, 1 = populist, emotional state 1 = anger, 0 = hope, 0 = control,
emotional reaction angry 1 = low, 7 = high, populist attitudes 1 = low, 7 = high. *p < .05., **p < .01., ***p < .001.
In order to test H3, populist appeal was the independent variable, emotional state hope the moderator, emotional reaction hopeful the mediator and populist attitudes the dependent variable. Anger was not used as a control.11 The model predicting emotional reaction was not
significant, F = (3,223) = 1.57, p = .196, R2 = .02. The interaction effect of populist appeal
and emotional state hope on reaction hopeful was not significant (b = -.1, p = .824), therefore rejecting a moderated mediation.12 As a result, H3 was not approved. The results of all other
paths can be found in Table 3.
11 This was done since the manipulation of emotional state hope was not succesful.
12 As an alternative, a moderation using PROCESS model 1 with emotional state hope as a moderator and a
mediation using PROCESS model 4 with emotional response hopeful as mediator were tested. They did not yield more significant results.
Table 3
Moderated mediation for predicting emotional reaction hopeful and populist attitudes
Effects b SE 95% BCBCI t p Constant on reaction hopeful 3.43 .17 [3.1, 3.76] 20.5 .000*** Appeal on reaction hopeful -.34 .25 [-.84, .15] -1.37 .173 State hope on reaction hopeful .37 .33 [-.28, 1.02] 1.12 .262 State hope * appeal on reaction -.1 .45 [-.98, .78] -.22 .824 Constant on attitudes 3.9 .2 [3,56, 4.29] 19.79 .000*** Reaction hopeful on attitudes -.02 .05 [-.11, .08] -.35 .723 Appeal on attitudes .08 .15 [-.21, .37] .53 .595
Notes. For Type of appeal, 0 = non- populist, 1 = populist, emotional state 0 = anger, 1 = hope, 0 = control,
emotional reaction hopeful 1 = low, 7 = high, populist attitudes 1 = low, 7 = high. *p < .05., **p < .01., ***p < .001.
Moderation of Agreeableness
Hypothesis 4 stated that the effect of the populist appeal on populist attitudes was moderated by Agreeableness: resulting in higher populist attitudes for low agreeable individuals than for high agreeable ones. To test this hypothesis, PROCESS model number 113 with 5000 bootstraps samples to estimate the bias corrected bootstrap confidence intervals
(BCBCI) was used (Hayes, 2013). A hierarchical multiple regression with populist appeal as the independent variable, Agreeableness as the moderator and populist attitudes as the dependent variable was performed.
The overall model was significant, F = (3,223) = 6.81, p <.001, R2 = .08. Populist
appeal was a significant predictor of populist attitudes, b = 2.49,t (223) = 2.61, p <.01, [.61,
4.36]. The impact of Agreeableness on populist attitudes was also significant, b = .52,t (223)
13 H4 was tested using a different model than H2 and H3 as it was not possible to use a single model to test all
= 4.42, p <.001, [.29, .75]. Finally, the interaction effect of Agreeableness and populist appeal on populist attitudes was significant too, b = -.43,t (223) = -2.52, p <.05, [-.76, -.09]. While
for low levels of Agreeableness a populist appeal predicts populist attitudes (b = .46, p < .05), this is not the case for average and high agreeable individuals (see Table 9). However, even if the effect is not significant, it goes in the correct direction: as Agreeableness increases, populist attitudes decrease. Consequently, H4 was approved.
Table 4
Populist attitudes for different levels of Agreeableness
Level of
Agreeableness b SE 95% BCBCI t p
Low .46 .2 [.07, .85] 2.34 .020*
Average .03 .15 [-.25, .32] .23 .817
High -.18 .18 [-.54, .18] -.98 .330
Notes. For Type of appeal, 0 = non- populist, 1 = populist, emotional state 0 = anger, 1= hope, 0 = control,
emotional reaction angry and hopeful 1 = low, 7 = high, populist attitudes 1= low, 7 = high. *p < .05., **p < .01., ***p < .001.
Discussion & conclusion
The aim of the study was to determine whether the effect of Facebook posts on populist attitudes was mediated by the emotional reaction of participants, and whether this effect was stronger for those in an angry or hopeful emotional state. In other words, users who were in an angry or hopeful state were expected to become angrier or more hopeful after seeing populist appeals on Facebook, thus bolstering their populist attitudes. Additionally, it was also investigated whether Agreeableness moderated the effect of populist posts on populist attitudes.
Contrary to what was hypothesized, populist content did not have a direct effect on populist attitudes. This could be due to the stability of populist attitudes (Hameleers et al., 2019), making them hard to change through single exposure to a post. This is in line with
extant research arguing that it is mostly the repeated exposure to certain messages that
produces effects and changes attitudes (Boomgaarden & Vliegenthart, 2007; Hameleers et al., 2019), even though single exposure has also shown small, but significant effects (Andreadis et al., 2019). Specific dimensions of populist attitudes – anti-elitism, preference for popular sovereignty or homogeneity of the people (Wettstein et al., 2019) – might be more variable and impacted by single exposure to a populist appeal. Distinguishing between those
dimensions in future research on online populist communication might lead to more conclusive results.
Populist content was also expected to have an indirect effect: if in an angry or hopeful state while seeing the post, users were presumed to become even angrier and more hopeful, increasing the persuasiveness of the post and leading to stronger populist attitudes. The hypothesized moderated mediation effect was not found. This might be explained by the techniques used to appraise anger and hope. The first manipulation of emotion through the emotion-inducing task (Valentino et al., 2011) proved to be successful for anger: participants in the anger group were significantly angrier than those in the hope and control group. Contrastingly, the second manipulation which made use of trigger sentences (Wirz, 2018) in the post failed to provoke distinctive levels of anger and hope across emotions conditions. This suggests that the first approach is more efficient in eliciting emotions in an online environment and offers a possible explanation for the absence of moderated mediation. Further, this finding indicates that while populist communication has been found to be emotional (Ernst et al., 2019), that does not necessarily mean it has the ability to provoke emotions amongst those who are exposed to it – contrasting with Wirz’s results (2018).
The link between emotions and populism could therefore be elsewhere: namely in the likelihood individuals have to experience certain emotions and how that affects their populist attitudes. The present study found that the participants who had an angry emotional reaction
also had significantly stronger populist attitudes, and this was not due to the manipulation of emotions – meaning that there is a correlation between being angry and having stronger populist attitudes. The key explanations for populist attitudes on the individual level might, therefore, lie in the personality traits relating to emotions, rather than on an emotional reaction to external messages. Some people are more prone to experience anger due to their
personality, and are therefore also more likely to have stronger populist attitudes. In general, anger has been found to be positively linked to the trait Neuroticism and negatively linked to Agreeableness (Ode et al., 2008; Pfeiler et al., 2018) and Conscientiousness (Pease & Lewis, 2015). While scholars have devoted attention to the importance of personality to explain political preferences (see Bakker et al., 2016; Schoen & Schumann, 2007 for example), the role of emotions in this context has only been hinted on.
The importance of psychological traits is also reflected in the results, supporting the moderating role of Agreeableness on populist attitudes. This finding is consistent with literature linking low Agreeableness to stronger populist attitudes (Ackermann et al., 2018; Bakker et al., 2016, 2020). It adds to the literature by showing that levels of Agreeableness also influence the effect of populist messages: these resonate particularly well amongst low agreeable people, bringing about stronger populist attitudes in this group.
Future research would benefit from building a bridge between personality and emotions to understand how these contribute to the formation of political attitudes. Further, including even more interactive features in experimental designs investigating the impact of populist communication on social media, as well as testing the impact of repeated exposure to populist messages, would increase the validity of these studies.
This research had some notable limitations, principally relating to the sample. Its characteristics could have played a negative role in the persuasiveness of the stimuli for two reasons: first, the large proportion of highly educated individuals, as existing research shows
that individuals with lower levels of education are more likely to support populist parties (Arzheimer, 2009; Bornschier, 2010) – of which there were very few in the sample. Second, the high proportion of people who positioned themselves on the left side of the political spectrum and did not vote for populist parties. While the subject of increasing healthcare costs used in the stimuli is not explicitly right-wing populist as immigration would be, it is possible that participants still classified the stance taken in the stimuli as right-wing, given the
predominance of right-wing populism in Switzerland, and therefore considered the stimuli to be in opposition to their own political orientation. Additionally, the complexity and length of the experiment led to a large number of drop-outs and consequently, a small sample size. Lastly, the reduced number of items used to measure Agreeableness did not capture the full complexity of the trait and ultimately led to a low scale reliability. However, even with a limited measure, exposure to populist appeals proved to impact populist attitudes, which speaks to the strength of this relationship.
Despite these limitations, the current study contributed to the existing literature by proposing an innovative research design to advance the knowledge on populist
communication in an online setting. Unlike other studies in the field, it tested the impact of Facebook posts by accounting for the fact that users can pick what they want to see on social media and added the possibility to select the Facebook friend whose post participants saw. This is an attempt to include interactivity in an experimental setting and had the added advantage of reducing possible bias towards the source of the stimuli. The study also tested the impact of populist appeals using a topic, increasing healthcare costs, that is more natural in the context of right-wing populism than equal wages, which was done in a previous study (Wirz, 2018) – leading to higher external validity. Further, different techniques to provoke emotions were tested in an online setting: an emotion-inducing task (Valentino et al., 2011) and the use of trigger sentences (Wirz, 2018).
In conclusion, this study demonstrated that emotions have a role to play in populist communication. However, not as expected through the use of emotional populist appeals on social media, but at the individual level most likely related to users’ personalities. In line with this, populist appeals were shown to resonate amongst low agreeable individuals. These findings suggest that the success of populist parties in Switzerland lies on their ability to suit the personality of the electorate.
References
Ackermann, K., Zampieri, E., & Freitag, M. (2018). Personality and voting for a right-wing populist party – evidence from Switzerland. Swiss Political Science Review, 24(4), 545–564. https://doi.org/10.1111/spsr.12330
Akkerman, A., Mudde, C., & Zaslove, A. (2014). How populist are the people? Measuring populist attitudes in voters. Comparative Political Studies, 47(9), 1324–1353. https://doi.org/10.1177/0010414013512600
Albertazzi, D. (2008). Switzerland: Yet another populist paradise. In D. Albertazzi & D. McDonnell (Eds.), Twenty-First Century Populism (pp. 100–118). London: Palgrave Macmillan. https://doi.org/10.1057/9780230592100_7
Andreadis, I., Cremonesi, C., Kartsounidou, E., Kasprowicz, D., & Hess, A. (2019). Attitudinal and behavioral responses to populist communication: The impact of populist message elements on populist attitudes and voting intentions. In C. Reinemann, J. Stanyer, T. Aalberg, F. Esser, & C. H. de Vreese (Eds.),
Communicating Populism: Comparing Actor Perceptions, Media Coverage, and Effects on Citizens in Europe (pp. 207–232). New York: Routledge.
https://doi.org/10.4324/9780429402067
1980–2002. American Journal of Political Science, 53(2), 259–275. https://doi.org/10.1111/j.1540-5907.2009.00369.x
Bakker, B. N., & Lelkes, Y. (2018). Selling ourselves short? How abbreviated measures of personality change the way we think about personality and politics. The Journal of
Politics, 80(4), 1311–1325. https://doi.org/10.1086/698928
Bakker, B. N., Rooduijn, M., & Schumacher, G. (2016). The psychological roots of populist voting: Evidence from the United States, the Netherlands and Germany. European
Journal of Political Research, 55(2), 302–320.
https://doi.org/10.1111/1475-6765.12121
Bakker, B. N., Schumacher, G., & Rooduijn, M. (2020). The populist appeal: Personality and anti-establishment communication. Retrieved from https://osf.io/wq8um/
Banks, A. J., & Valentino, N. A. (2012). Emotional substrates of white racial attitudes.
American Journal of Political Science, 56(2), 286–297.
https://doi.org/10.1111/j.1540-5907.2011.00561.x
Bernhard, L. (2017). Three faces of populism in current Switzerland: Comparing the populist communication of the Swiss People’s Party, the Ticino League, and the Geneva Citizens’ Movement. Swiss Political Science Review, 23(4), 509–525.
https://doi.org/10.1111/spsr.12279
Boomgaarden, H. G., & Vliegenthart, R. (2007). Explaining the rise of anti-immigrant parties: The role of news media content. Electoral Studies, 26(2), 404–417.
https://doi.org/10.1016/j.electstud.2006.10.018
Bornschier, S. (2010). The new cultural divide and the two-dimensional political space in Western Europe. West European Politics, 33(3), 419–444.
https://doi.org/10.1080/01402381003654387
leaders. Communications: The European Journal of Communication Research, 35(2), 141–163. https://doi.org/10.1515/COMM.2010.008
Bos, L., van der Brug, W., & de Vreese, C. (2011). How the media shape perceptions of right-wing populist leaders. Political Communication, 28(2), 182–206.
https://doi.org/10.1080/10584609.2011.564605
Bos, L., van der Brug, W., & de Vreese, C. H. (2013). An experimental test of the impact of style and rhetoric on the perception of right-wing populist and mainstream party leaders. Acta Politica, 48(2), 192–208. https://doi.org/10.1057/ap.2012.27 Brader, T. (2005). Striking a responsive chord: How political ads motivate and persuade
voters by appealing to emotions. American Journal of Political Science, 49(2), 388– 405. https://doi.org/10.1111/j.0092-5853.2005.00130.x
Chadwick, A. E. (2015). Toward a theory of persuasive hope: Effects of cognitive appraisals, Hope Appeals, and Hope in the Context of Climate Change. Health Communication,
30(6), 598–611. https://doi.org/10.1080/10410236.2014.916777
Conseil fédéral. (2019). Coûts de l’assurance-maladie: La prime moyenne augmentera de 0,2% en 2020. Retrieved from
https://www.admin.ch/gov/fr/accueil/documentation/communiques.msg-id-76507.html Curato, N. (2016). Politics of anxiety, politics of hope: Penal populism and Duterte’s rise to
power. Journal of Current Southeast Asian Affairs, 35(3), 91–109. https://doi.org/10.1177/186810341603500305
de Vreese, C. H., Esser, F., Aalberg, T., Reinemann, C., & Stanyer, J. (2018). Populism as an expression of political communication content and style: A new perspective. The
International Journal of Press/Politics, 23(4), 423–438.
https://doi.org/10.1177/1940161218790035
politicians spread a fragmented ideology. Information, Communication & Society,
20(8), 1109–1126. https://doi.org/10.1080/1369118X.2016.1207697
Ernst, N., Blassnig, S., Engesser, S., Büchel, F., & Esser, F. (2019). Populists prefer social media over talk shows: An analysis of populist messages and stylistic elements across six countries. Social Media + Society, 5(1).
https://doi.org/10.1177/2056305118823358
Ernst, N., Engesser, S., Büchel, F., Blassnig, S., & Esser, F. (2017). Extreme parties and populism: An analysis of Facebook and Twitter across six countries. Information,
Communication & Society, 20(9), 1347–1364.
https://doi.org/10.1080/1369118X.2017.1329333
Fletcher, R., & Nielsen, R. K. (2018). Are people incidentally exposed to news on social media? A comparative analysis. New Media & Society, 20(7), 2450–2468. https://doi.org/10.1177/1461444817724170
Gallego, A., & Oberski, D. (2012). Personality and political participation: The mediation hypothesis. Political Behavior, 34(3), 425–451. https://doi.org/10.1007/s11109-011-9168-7
Gerbaudo, P. (2018). Social media and populism: An elective affinity? Media, Culture &
Society, 40(5), 745–753. https://doi.org/10.1177/0163443718772192
Hameleers, M. (in press). Augmenting polarization via social media? A comparative analysis of Trump’s and Wilders’ online populist communication and the electorate’s
interpretations surrounding the elections. Acta Politica. https://doi.org/10.1057/s41269-018-0119-8
Hameleers, M., Andreadis, I., & Reinemann, C. (2019). Investigating the effects of populist communication: Design and measurement of the comparative experimental study. In C. Reinemann, J. Stanyer, T. Aalberg, F. Esser, & C. H. de Vreese (Eds.),
Communicating Populism: Comparing Actor Perceptions, Media Coverage, and Effects on Citizens in Europe (pp. 168–182). New York: Routledge.
https://doi.org/10.4324/9780429402067
Hameleers, M., Bos, L., & de Vreese, C. H. (2017). “ They did it”: The effects of emotionalized blame attribution in populist communication. Communication
Research, 44(6), 870–900. https://doi.org/10.1177/0093650216644026
Hameleers, M., & de Vreese, C. H. (2018). To whom are “the people” opposed?
Conceptualizing and measuring citizens’ populist attitudes as a multidimensional construct. Journal of Elections, Public Opinion and Parties, 1–20.
https://doi.org/10.1080/17457289.2018.1532434
Hameleers, M., & Schmuck, D. (2017). It’s us against them: A comparative experiment on the effects of populist messages communicated via social media. Information,
Communication & Society, 20(9), 1425–1444.
https://doi.org/10.1080/1369118X.2017.1328523
Hayes, A. F. (2013). Introduction to mediation, moderation, and conditional process
analysis: A regression-based approach. New York: The Guilford Press. Retrieved
from http://www.personal.psu.edu/jxb14/M554/specreg/templates.pdf
Housholder, E. E., & LaMarre, H. L. (2014). Facebook politics: Toward a process model for achieving political source credibility through social media. Journal of Information
Technology & Politics, 11(4), 368–382.
https://doi.org/10.1080/19331681.2014.951753
Kalmoe, N. P., Gubler, J. R., & Wood, D. A. (2018). Toward conflict or compromise? How violent metaphors polarize partisan issue attitudes. Political Communication, 35(3), 333–352. https://doi.org/10.1080/10584609.2017.1341965
Public Opinion Quarterly, 76(4), 747–760. https://doi.org/10.1093/poq/nfs055
Karpf, D. (2017). Digital politics after Trump. Annals of the International Communication
Association, 1–9. https://doi.org/10.1080/23808985.2017.1316675
Krämer, B. (2014). Media populism: A conceptual clarification and some theses on its effects.
Communication Theory, 24(1), 42–60. https://doi.org/10.1111/comt.12029
Krämer, B. (2017). Populist online practices: The function of the internet in right-wing populism. Information, Communication & Society, 20(9), 1293–1309.
https://doi.org/10.1080/1369118X.2017.1328520
Metzger, M. J., Flanagin, A. J., & Medders, R. B. (2010). Social and heuristic approaches to credibility evaluation online. Journal of Communication, 60(3), 413–439.
https://doi.org/10.1111/j.1460-2466.2010.01488.x
Mondak, J. J., & Halperin, K. D. (2008). A Framework for the study of personality and political behaviour. British Journal of Political Science, 38(2), 335–362.
https://doi.org/10.1017/S0007123408000173
Mudde, C. (2004). The populist Zeitgeist. Government and Opposition, 39(4), 541–563. https://doi.org/10.1111/j.1477-7053.2004.00135.x
Mudde, C. (2019). The far right today. Cambridge, UK: Polity Press.
Müller, P., Schemer, C., Wettstein, M., Schulz, A., Wirz, D. S., Engesser, S., & Wirth, W. (2017). The polarizing impact of news coverage on populist attitudes in the public: Evidence from a panel study in four European democracies. Journal of
Communication, 67(6), 968–992. https://doi.org/10.1111/jcom.12337
Nabi, R. L. (1999). A cognitive-functional model for the effects of discrete negative emotions on information processing, attitude change, and recall. Communication Theory, 9(3), 292–320. https://doi.org/10.1111/j.1468-2885.1999.tb00172.x
for Agreeableness in moderating Neuroticism’s influence on anger and aggression.
Journal of Research in Personality, 42(2), 295–311.
https://doi.org/10.1016/j.jrp.2007.05.007
Office fédéral de la Statistique. (2019a). Année des élections fédérales—Pages web de l’OFS. Retrieved from https://www.elections.admin.ch/fr/ch/
Office fédéral de la Statistique. (2019b). Niveau de formation de la population—Données de l’indicateur—1996-2018. Retrieved from
https://www.bfs.admin.ch/bfs/fr/home/statistiques/education-science.assetdetail.7886034.html
Pease, C. R., & Lewis, G. J. (2015). Personality links to anger: Evidence for trait interaction and differentiation across expression style. Personality and Individual Differences, 74, 159–164. https://doi.org/10.1016/j.paid.2014.10.018
Petty, R. E., & Briñol, P. (2015). Emotion and persuasion: Cognitive and meta-cognitive processes impact attitudes. Cognition and Emotion, 29(1), 1–26.
https://doi.org/10.1080/02699931.2014.967183
Pfeiler, T. M., Weber, H., & Kubiak, T. (2018). Experiencing anger in a social interaction: The role of personality. Personality and Individual Differences, 132, 45–51. https://doi.org/10.1016/j.paid.2018.04.045
Plaisant, O., Courtois, R., Réveillère, C., Mendelsohn, G. A., & John, O. P. (2010). Validation par analyse factorielle du Big Five Inventory français (BFI-Fr). Analyse convergente avec le NEO-PI-R. Annales Médico-Psychologiques, Revue Psychiatrique, 168(2), 97–106. https://doi.org/10.1016/j.amp.2009.09.003
Rico, G., Guinjoan, M., & Anduiza, E. (2017). The emotional underpinnings of populism: How anger and fear affect populist attitudes. Swiss Political Science Review, 23(4). https://doi.org/10.1111/spsr.12261
Schoen, H., & Schumann, S. (2007). Personality traits, partisan attitudes, and voting behavior. evidence from Germany. Political Psychology, 28(4), 471–498.
https://doi.org/10.1111/j.1467-9221.2007.00582.x
Schulz, A., Müller, P., Schemer, C., Wirz, D. S., Wettstein, M., & Wirth, W. (2018). Measuring populist attitudes on three dimensions. International Journal of Public
Opinion Research, 30(2), 316–326. https://doi.org/10.1093/ijpor/edw037
Sheets, P., Bos, L., & Boomgaarden, H. G. (2015). Media cues and citizen support for right-wing populist parties. International Journal of Public Opinion Research, 28(3), 307– 330. https://doi.org/10.1093/ijpor/edv014
Stavrakakis, Y., Andreadis, I., & Katsambekis, G. (2017). A new populism index at work: Identifying populist candidates and parties in the contemporary Greek context.
European Politics and Society, 18(4), 446–464.
https://doi.org/10.1080/23745118.2016.1261434
Stieglitz, S., & Dang-Xuan, L. (2013). Emotions and information diffusion in social media-sentiment of microblogs and sharing behavior. Journal of Management Information
Systems, 29(4), 217–248. https://doi.org/10.2753/MIS0742-1222290408
Thiele, D. (2019). Emotional framing in right-wing populist communication on Twitter: A
German dictionary-based measurement. Retrieved from
https://ecpr.eu/Events/PaperDetails.aspx?PaperID=47111&EventID=123
Turner, R. N., Dhont, K., Hewstone, M., Prestwich, A., & Vonofakou, C. (2014). The role of personality factors in the reduction of intergroup anxiety and amelioration of outgroup attitudes via intergroup contact. European Journal of Personality, 28(2), 180–192. https://doi.org/10.1002/per.1927
Valentino, N. A., Brader, T., Groenendyk, E. W., Gregorowicz, K., & Hutchings, V. L. (2011). Election night’s alright for fighting: The role of emotions in political
participation. The Journal of Politics, 73(1), 156–170. https://doi.org/10.1017/S0022381610000939
Van Hauwaert, S. M., Schimpf, C. H., & Azevedo, F. (2019). The measurement of populist attitudes: Testing cross-national scales using item response theory. Politics.
https://doi.org/10.1177/0263395719859306
Van Hauwaert, S. M., & Van Kessel, S. (2018). Beyond protest and discontent: A cross-national analysis of the effect of populist attitudes and issue positions on populist party support. European Journal of Political Research, 57(1), 68–92.
https://doi.org/10.1111/1475-6765.12216
Wettstein, M., Schulz, A., Steenbergen, M., Schemer, C., Müller, P., Wirz, D. S., & Wirth, W. (2019). Measuring populism across nations: Testing for measurement invariance of an inventory of populist attitudes. International Journal of Public Opinion Research,
edz018. https://doi.org/10.1093/ijpor/edz018
Wirz, D. (2018). Persuasion through emotion? An experimental test of the emotion-eliciting nature of populist communication. International Journal of Communication, 12, 1114–1138. https://doi.org/10.5167/uzh-149959
Appendix A Pictures used for the Facebook posts
Prompt: please select the picture of someone who could be your Facebook friend.14
Figure A1. Man 1 [Photography]. (n.d.). Figure A2. Man 2 [Photography]. (n.d.).
Retrieved from Canva.com Retrieved from Canva.com
Figure A3. Woman 1 [Photography]. (n.d.). Figure A4. Woman 2 [Photography]. (n.d.). Retrieved from https://foter.com/ffff/photo/ Retrieved from Canva.com
9524771956/3abe7b82b1/
Figure A5. Man 3 [Photography]. (n.d.). Figure A6. Man 4 [Photography]. (n.d.).
Retrieved from Canva.com Retrieved from Canva.com
Figure A7. Woman 3 [Photography]. (n.d.). Figure A8. Woman 4 [Photography]. (n.d.). Retrieved from https://foter.com/ffff/photo/ Retrieved from Canva.com
Appendix B Example of stimuli for the appeal in French
Note. The source of the post varied according to the person the participant selected as possible
Facebook friend.
Appendix C Table C1
Agreeableness scale in English
N° Item Wording
1 kind Is kind to others
2 rude Is sometimes a bit rude to others
3 forgiving Is forgiving
4 critical Has a tendency to criticize others
5 trusting Generally trusts others
6 helpful Is helpful and is not selfish
7 contemptuous Is sometimes contemptuous
8 conflictual Looks for conflict with others
9 quarrelling Easily starts to quarrel with others
10 cooperative Likes to cooperate with others
