Political Disinformation: A Threat to Democracy?
The Role of Emotions and Personality in Political Information Processing
Master’s Thesis
Graduate School of Communication Science Political Communication Science
Supervisor: Dr. Alessandro Nai
Janne Trauzettel 11845732
February 1, 2018 Abstract
This paper investigates the implications of the increasing dissemination of political
disinformation – that aims at discrediting political institutions and incumbents – for individuals’ political trust. To account for individual’s differences in the susceptibility to disinformation, individual’s personality as the most stable and emotions as the least stable indicators of an individual’s political profile were taken into account. Participants (N = 198) of the online survey-embedded experiment were, first of all, induced with aversion or enthusiasm, and secondly exposed to a newspaper article containing disinformation or true information. No sufficient support was found that individual’s level of political trust decreases following the exposure to political disinformation. The findings were still not significant when accounting for individual’s personality traits. Interestingly, an interaction effect between individual’s level of aversion and exposure to disinformation on their level of political trust was found.
Introduction
“Yet the value of trusted, factual journalism has never been higher, the need for it never greater.” – Stephen J. Adler, Chief-in-Editor, Reuters
With the European Parliamentary Elections 2019 coming closer, politicians fear about an increasing dissemination of disinformation – often referred to as ‘fake news’ - that could sway voters. The greater extent of this relatively new phenomenon was visible during the Brexit referendum of 2016 or the past U.S. presidential election, where online disinformation campaigns initiated by Russia influenced election results. In reaction to these events, online platforms, social media networks and the advertising industry as for instance Google and Facebook have agreed on taking on responsibility in the fight against online disinformation by signing a code of practice. Studies found that disinformation is more often shared on social media than legitimate news (‘real news’) (Silverman, Strapagiel, Shaban, & Hall, 2016). Despite that, the repetitive exposure to disinformation increases the belief in the accuracy of
disinformation (Pennycook, Cannon, & Rand, 2018). The personalized information environment consisting of filter bubbles and echo chambers might even enhance exposure to disinformation for some users (Newman, 2017). According to Wardle and Derakshan (2017), we are still at the beginning of understanding the implications of this information disorder (p. 4).
Since the media is the main source for political news, disinformation that is targeted at political institutions and incumbents might affect individual’s opinion about the government. The Edelman Trust Barometer (2019) conducted across 27 countries indicates that 73% of Europeans
worry about ‘fake news’ being used as a weapon (p. 20). To this point, there has been only one study, which analyzed the effect of ‘fake news’ on voter’s attitudes during the 2009 Israeli election campaign. Balmas (2014) found that ‘fake news’ positively affect people’s feelings of inefficacy, alienation and cynicism, which is mediated by the perceived realism of the ‘fake news’. These findings suggest that disinformation might be one of the reasons for the current ‘Vertrauenskrise’ (‘crisis of confidence’) – a term that is frequently used to describe the
increasing distrust of German institutions and the media. Although, trust amongst the informed public in Germany is rising (61%), only 44% of the general population trusts the media and the government (Edelman Trust Barometer, p. 7). According to Putnam (1993), trust in political institutions is, however, essential in a functioning democracy: ”The greater the level of trust within a community, the greater the likelihood of cooperation. And cooperation itself breeds trust” (p. 171).
The aim of this paper is, therefore, to contribute to an understanding of the implications of the dissemination of political disinformation for democratic societies. First of all, the paper looks at the influence of political disinformation on individuals’ level of political trust. According to Zaller (1992), individuals’ political awareness and predispositions influence the way they seek out and process the information they encounter. In order to understand differences in the susceptibility to political disinformation, individual’s personality as the most stable
predictor of a person’s political profile is taken into account. Taber and Lodge (2016) criticize Zaller’s model of political reasoning, however, as being too narrow as it assumes that individuals would process information solely consciously (cold cognition) (p. 74). Instead, they argue that people’s emotional judgement precedes cognitive processes, which is rather unintentional and difficult to control (hot cognition hypothesis) (Taber & Lodge, 2016, p. 74). Therefore, emotions
as the least stable indicator of an individual’s political profile are also taken into account. Following the Affective Intelligence Theory by Marcus, Neumann, & MacKuen (2000), this paper focusses on emotions of enthusiasm and aversion, as these are both triggered by the same affective subsystem, the dispositional system.
The aim of this paper is to analyze to what extent individual personality and emotions (aversion and enthusiasm) moderate the effect of exposure to political disinformation on individual’s level of political trust?
In the following, the theoretical background from which the hypotheses are derived will be explained. In the second part, the methodology of the study including the research design, selection and composition of the sample and measures of the variables is presented. The third part illustrates the findings of the survey-embedded online experiment in regard to research question and hypotheses. The final part includes a discussion and conclusion with the limitations as well as theoretical and practical implications of the study for future research.
The Importance of Political Trust for Democratic Societies
According to Warren (2017), the two most important types of trust that are essential for a functional democracy are social trust and institutional trust. A trust relationship always consists of three components: A trusts B to do X, in which A has an interest (Hardin, 1999, pp. 24–26).
The focus of this paper is on institutional trust, which comprises the ‘vertical’ trust relationships between citizens and people holding institutional positions (Cleary, & Stokes, 2009; Zmerli & Newton, 2011; Warren, 2017). Citizens generally have more confidence in impartial institutions holding a public trust (first-order institutional trust) such as the police or judiciary than in political institutions (second-order institutional trust) as for instance the national parliament, political parties or politicians (Norris, 2011, 2017; Zmerli and Newton,
2017; Marien, 2011). In line with this, Norris (2017) defines “political trust as the general belief in the performance capacity of political institutions and/or belief in the benevolent motivation and performance capacity of office-holders” (p. 24). In a healthy democracy, public trust should be high as it enables institutions to provide citizens with important public goods as for instance health care, security, transportation, or communications (Van der Meer, 2017). Political trust, on the other hand, is essential as it increases citizen’s support for government activity as for instance their support for policies (Hetherington, 2006; Hetherington & Rudolph, 2015).
In recent years, we witness a decline in political trust amongst German citizens, but how can this development be explained?
The Michigan School suggests that our opinions and our political behavior (e.g. vote choice) is determined by sociological and psychological factors. In their book The American Voter,
Campbell, Converse, Miller, & Stokes (1960) present the funnel of causality showing that voting behavior is influenced by stable-long term forces such as voters’ preferences (values) and party attachment (attitude), and unstable short-term forces, which include voters’ opinions about current issues and their image of a candidate (opinion) (Campbell et al., 1960; Hofferbert, 1937, p. 228). Hence, voters’ opinions are strongly influenced by campaign activity as well as the media coverage, which explains fluctuations in voter opinions (Campbell et al., 1960). This means that one of the reasons for a decreasing political trust might be the current media environment.
The Information Disorder and Implications of Political Disinformation
According to Wardle and Derakshan (2017), the current media environment can be described as an information disorder, which refers to the increasing dissemination of dis-, mis-, and mal-information online (p. 20). The latter refers to information that is based on reality, but
used to inflict harm on a person, organization or country, whereas mis-information is false information that is created without the intention to cause harm (Wardle & Derakshan, 2017, p. 20). Most problematic is, however, dis-information – often referred to as ‘fake news’- as it “includes all forms of false, inaccurate, or misleading information designed, presented and promoted to intentionally cause public harm or for profit” (2018, p. 5). Scholars acknowledge that the term ‘fake news’ is highly politicized and thus misleading as it is increasingly used by powerful actors to criticize media coverage that is against their opinion (European Union, 2018, p. 5). Responsible for spreading disinformation can be political actors, for-profit actors, media organizations, or citizens, individually or in groups (European Union, 2018, p. 5), who typically have political, financial, psychological, or social motivations (Wardle & Derakshan, 2017, p. 26). In order to distribute disinformation online on a large scale, many agents make use of automation software, primarily social bots that imitate human behavior (Shao, Ciampaglia, Varol, Flammini, & Menczer, 2017). Therefore, it is important to understand the implications of this relatively new phenomenon for democratic societies.
The analysis presented in this paper focuses on disinformation that is targeted at political incumbents and institutions as it might affect individuals’ political trust. This paper defines this type of disinformation as political disinformation. So far, there has been only one study by Balmas (2014) about this topic, showing that exposure to ‘fake news’ during the 2009 Israeli election campaign positively affected people’s feelings of inefficacy, alienation and cynicism, which is mediated by the perceived realism of ‘fake news’. Accordingly, people with high exposure to ‘fake news’ and low exposure to hard news are more likely to perceive
Therefore, it can be assumed that exposure to political disinformation might decrease individual’s political trust. Hence, it is assumed that individuals exposed to political
disinformation will have lower levels of trust in political institutions than individuals exposed to
true political information (H1). As there are no studies about the implications of political
disinformation on individual’s political attitudes focusing on a European context, this study aims at filling this gap in research.
Zaller’s Reception-Acceptance-Model (1992) implies that the effect of exposure to disinformation on political trust might differ amongst people due to basic differences in the way individuals seek out and process the information they encounter. According to his model, the opinion one holds reflects the information a person received, accepted and sampled from, which depends on their political awareness and their political predispositions (Zaller, 1992, p. 21). Political awareness refers to “the extent to which an individual pays attention to politics and understands what he or she has encountered” (p. 21). Accordingly, individuals with higher levels of political awareness are more likely to be exposed to, understand, and thus receive political messages (reception axiom) (Zaller, 1992, p. 42). In accordance with this, dual-models of attitude change show that individuals with low motivation (e.g. information is personally irrelevant) and/or ability (e.g. political awareness) mainly process information heuristically, whereas individuals with high motivation and ability engage in systematic, more cognitive effortful information processing (Chaiken, 1980; Eagly & Chaiken, 1993; Lau & Redlawsk, 2001; Petty & Cacioppo, 1979). According to Sniderman, Brody, & Tetlock (1991):
“heursitics judgmental shortcuts that organize and simplify […] choices, efficient in the double sense of requiring relatively little information to execute, yet yielding dependable
answers even to complex problems of choice.” (p. 19.)
A recent study by Pennycook and Rand (2018) found that individuals who are willing to think analytically are less susceptible to misinformation and are able to distinguish between ‘true news’ and ‘false news’ (pp. 8-9). Therefore, individuals engaging in heuristic information-processing are more susceptible to misinformation due to their cognitive limitations in processing information (p. 2). In political reality, most individuals – also the ones with high political awareness – are limited information processors that rely on cognitive heuristics (e.g. candiwhen forming political judgments and making decisions (Kahneman, Slovic, & Tversky, 1982, Nisbett & Ross, 1980; Kahneman, & Tversky; 1974). However, individuals with higher political awareness are shown to be better in using heuristics to their advantage (Lau & Redlawsk, 2006). In the literature, other concepts such as political sophistication are used to describe individuals’ level of political awareness (Zaller, 1992, p. 21). Following Luskin (1990), political sophistication depends on individuals’ motivation to engage with political information
(political interest) as well as on their intelligence (political knowledge) and can therefore be
considered as an appropriate indicator of individuals’ political awareness. Big Five Personality Traits and Political Information Processing
In order to make inferences about individuals’ level of political sophistication, individuals’ personality will be taken into account. Previous research shows that individual’s personality traits are in part heritable (Bouchard, 1997, Plomin, van Gestel & van Broeckhoven, 2003) and stable through the whole life cycle (Bouchard & Loehlin, 2001; Caspi, Roberts, & Shiner, 2005; Srivastava, John, Gosling, & Potter, 2003). Therefore, a person’s personality is seen as a very good predictor – often even better than demographic variables such as age and
income – of a person’s political profile including their political sophistication (Gerber, Huber, Doherty and Dowling; 2011b; Mondak, Hibbing, Canache, Seligson, & Anderson, 2010; Mondak & Halperin, 2008; Vecchione & Caprara, 2009). To my knowledge, there has been no study that analyzed individual differences in the receptiveness to political disinformation by taking individuals’ personality into account.
Following Gerber et al. (2011b), the most dominant theoretical framework for measuring personality traits are the ‘Big Five’ personality dimensions by Saucier and Goldberg (1998), which correspond to the personality dimensions of the parallel model, the five-factor model (Costa & McCrae; 1988). The ‘Big Five’ personality traits are extraversion, agreeableness, conscientiousness, emotional stability (often referred to by its inverse—neuroticism), and openness to experience.
Open-minded persons are described as being very creative, curious, culturally interested, and nonconforming (McCrae & Costa, 2003; Mondak & Halperin, 2008). They are interested in challenging information and the exchange of political ideas resulting in higher levels of political interest, political knowledge and opinionation and news consumption (Mondak & Halperin, 2008; Mondak, 2010; Gerber et al. 2011b). Therefore, it is expected that the effect of exposure to
political disinformation on individuals’ level of political trust will be stronger for individuals with lower levels of openness to experience than for individuals with higher levels of openness to
experience (H2a).
Extraversion is regarded as an “energetic approach toward the social and material world
and includes traits such as sociability, activity, assertiveness, and positive emotionality” (John et al., 2008, p. 120). As extraverted people value sociability and sharing their views with others, they actively participate in political discussion, and are very opinionated (Mondak & Halperin,
2008; Mondak, 2010; Gerber et al. 2011). Interestingly, extraverted people “talk (and) talk and talk”, but do not pay a lot attention to politics nor are they well informed (Mondak, 2010, p. 108). Therefore, it can be assumed that the effect of exposure to political disinformation on
individuals’ level of political trust will be stronger for individuals with higher levels of extraversion than for individuals with lower levels of extraversion (H2b).
Conscious people are very organized, productive, dutiful, ambitious, responsible, and reliable (McCrae & Costa, 2003; Mondak, 2010; Mondak & Halperin, 2008; Schoen &
Steinbrecher, 2013). In contrast to extroverted people, people scoring high on conscientiousness are found to be less opinionated (Mondak, 2010). In general, conscious individuals are politically unengaged, not very attentive to politics and therefore show lower levels of political knowledge (Mondak & Halperin, 2008; Mondak, 2010). Thus, one can assume that the effect of exposure to
political disinformation on individuals’ level of political trust will be stronger for people with higher levels of conscientiousness than for people with lower levels of conscientiousness (H2c).
The fourth trait, emotional stability, “describes even-temperedness and contrasts with negative emotionality, such as feeling anxious, nervous, sad, tense” (John et al., 2008, p. 120). Gerber et al. (2011b) find evidence that emotionally stable people are politically interested and
sophisticated, but are - because of their calm demeanour - not very opinionated (Mondak & Halperin, 2008, p. 345). Due to the contentiousness of political information, emotionally unstable or neurotic individuals tend to avoid political news that is upsetting (Mondak & Halperin, 2008; Gerber et al., 2011b). Accordingly, it is expected that the effect of exposure to
political disinformation on individuals’ level of political trust will be stronger for individuals with lower levels of emotional stability than for individuals with higher levels of emotional
Finally, agreeableness is defined as a trait that “contrasts a prosocial and communal orientation toward others with antagonism and includes traits such as altruism,
tender-mindedness, trust and modesty” (John et al., 2008, p. 120). Similar to neurotic people, agreeable people are less attentive to political information due to its conflictual potential as they prefer harmonious relationships. That is why people scoring low on agreeableness are less politically sophisticated and opinionated than agreeable people (Mondak & Halperin, 2008, p. 357; Mondak, 2010, p. 119). Hence, it can be assumed that the effect of exposure to political
disinformation on individuals’ level of political trust will be stronger for individuals with lower levels of agreeableness than for individuals with higher levels of agreeableness (H2e).
Taber and Lodge (2016) criticize Zaller’s model of political reasoning, however, as being too narrow as it assumes that individuals would process information solely consciously (cold
cognition) (p. 74). Instead, they argue that people’s emotional judgement precedes cognitive
processes, which is rather unintentional and difficult to control (hot cognition hypothesis) (Taber & Lodge, 2016, p. 74). Similarly, the Affective Intelligence Theory (AIT) by Marcus et al. (2000) shows that a person’s emotional judgment influences their information seeking and processing, which in turn determines their opinion. These studies suggest that individuals current emotional state, which changes constantly and is therefore very unstable, also affects their political information processing and judgement. Therefore, it is assumed that emotions of aversion and enthusiasm might influence the effect of exposure to political disinformation on individuals’ level of political trust.
The Role of Emotions in Political Information Processing
Emotions are “internal, mental states representing evaluative, valenced reactions to events, agents, or objects that vary in intensity…[t]hey are generally short-lived, intense, and
directed at some external stimuli” (Nabi, 1999, p. 295). The three-dimensional model by
Steenbergen and Ellis (2006) shows that different emotions can be captured with one dimension of positive affect, enthusiasm, and two of negative affects, fear and anger (p. 111). Following functional emotion theory, Steenbergen and Ellis (2006) state that fear and anger are two distinct classes of emotions that have different effects on an individual’s behavior and decision making, which is also reflected in the Affective Intelligence Theory (Steenbergen & Ellis, 2006, p. 111).
Drawing upon neuroscience, the dual-model of Affective Intelligence by Marcus et al. (2000) suggests that our emotional states result from two affective subsystems, the dispositional
system and the surveillance system, each activated under different conditions. When confronted
with familiar liked or disliked stimuli, enthusiasm as well as aversion are generated by the dispositional system (Marcus et al., 2000, p. 47). In novel situations (e.g. epidemic, natural catastrophe), however, the surveillance system is activated triggering emotions of anxiety (Marcus et al., 2000, p. 57). Lau and Redlawsk (2001) point out that the political environment is usually not one of high emotional anxiety, which is why this paper focusses on the role of enthusiasm and aversion triggered by the dispositional system. In general, the dispositional system encourages the pursuit of individual political goals, and increases political interest and participation (Brader, 2006, MacKuen et al., 2010).
Enthusiasm is associated with other emotions such as ‘hope’, ‘pride’. ‘joy’, and ‘happiness’ (Brader, 2006; Marcus et al., 2000; Valentino et al., 2011), eliciting “state of
excitement and expectation for what’s happening and what’s ahead” (Brader & Marcus, 2013, p. 8). Enthusiasm is associated with a heuristic processing style, which strengthens the loyalty and reliance on prior beliefs (Brader, 2006; Marcus et al., 2000, Isabell, Ottati, & Burns, 2006).
The effects of emotions of aversion on individual’s behavior and decision-making are, in contrast to enthusiasm and anxiety, under-researched. Aversion comprises a cluster of feelings including anger, disgust, contempt, and hatred (Brader & Marcus, 2013, p. 11). Steenbergen and Ellis (2006) state that aversion “reinforces a desire to react against the causes of a negative stimulus or to punish those who could control or moderate the stimulus” (p. 111). Angry individuals engage in heuristic information processing relying on their habits as for instance prior opinions and stereotypes (Bodenhausen, Sheppard, and Kramer, 1994; Tiedens & Linton, 2001; Isabell et al., 2006).
Studies suggest that both, aversion and enthusiasm, are high-arousal emotions and therefore have a stronger effect on opinions and policy support (Lecheler, Schuck, & De Vreese, 2013). Despite that, aversion and enthusiasm increase individuals’ reliance on heuristic cues, which makes them more susceptible to political disinformation. Therefore, I assume that the
effect of political disinformation on individuals’ trust in political institutions will be stronger for people experiencing higher levels of aversion, than for people experiencing lower levels of
aversion (H7). And secondly, I expect that the effect of political disinformation on individuals’
political trust will be stronger for people with higher levels of enthusiasm than for people
experiencing higher levels of enthusiasm (H8).
The Experimental Design
According to Wilson and Eckel (2008), researchers increasingly use survey experiments to measure attitude change and more specifically political trust. Therefore, the study employed a true-experimental between-subjects design. Experiments enable researchers to manipulate the content of information in a controlled setting. Through random assignment of participants to the
experimental group and the control group one can make inferences about whether differences in attitudes are due to the manipulation or not. To test the hypotheses, a survey-embedded online experiment was conducted.
Participants
The online experiment was conducted from the 12th until the 20th of December 2018. A total of 249 subjects participated in the survey, of which 49 cases had to be excluded from the sample as they did not complete the survey. As there were not enough participants that indicated to identify with a gender other than male or female (‘other’), I had to exclude these cases from the analysis. This resulted in a total of 198 participants (81 = male, 117 = female). The sample constitutes a convenient sample as most participants were friends, family members and
colleagues that shared the study with their network, which resulted in a snowball effect. Participants were between 19 and 68 years old (M = 32.97, SD = 12.82) with most participants having a higher educational level (tertiary education = 160) and only a few having an upper
secondary education (= 34) and a lower secondary education (= 3). The selection criteria for
participants was their country of origin and age. As it was a prerequisite to have sufficient
knowledge about the German political system, only German citizens or people living in Germany for more than ten years who were at least 18 years old were allowed to participate in the study.
Procedure
As mentioned before, the online experiment created on Qualtrics was distributed through traditional media (e.g. email) and social media channels (e.g. Facebook, WhatsApp and
LinkedIn). After indicating an interest in the study, the participants were sent a survey link. The survey took approximately ten minutes to complete. First, participants read the informed consent introducing the structure and topic of the survey. The actual topic of the study was not
mentioned, instead the participants were informed that the survey would be about “Political Communication and Political Attitudes” to prevent biased answers. After providing consent, the participants were asked two control questions about their country of origin and age to ensure their eligibility. After that, a few questions about participant’s political demographics and personality traits followed. The experiment consists of two different treatments. The first treatment aimed at inducing the participants with the emotions enthusiasm or aversion by exposing them to an image. The participants were randomly assigned to the aversion condition and the enthusiasm condition. In order to control whether the emotion induction was successful, participants were asked to self-report their affective state following the exposure to one of the two images. As part of the second treatment, individuals were exposed to a news article. Participants induced with aversion were randomly assigned to the experimental condition (political disinformation), whereas the other participants induced with enthusiasm were
randomly assigned to the control condition (true political information). The same was done with participants induced with enthusiasm. After being exposed to one two conditions, participants were asked to report the credibility of the content of the message to see whether the manipulation of the second treatment was successful. Finally, the participants had to answer a question about their level of political trust and their socio-demographics. In a debrief, they were informed about the research purpose of the study and the manipulation (See Figure 3).
Stimuli
Treatment I. Images were used to induce participants with emotions of aversion and enthusiasm, which is a widely used and standardized method (Uhrig et al., 2016). The images were obtained from the open-access online stimulus set, the Open Affective Standardized Image Set (OASIS) (Kurdi, Lozano, & Banaji, 2017) as well as from the open-source database
pixabay.com. Most databases indicate the normative ratings including the valence (the quality of the affective response) and arousal (the intensity of the affective response), but not the emotion that is evoked by the image. Therefore, the images were pre-tested in order to ensure that these would elicit emotions of aversion and enthusiasm (See Figure 4 and 5). Based on two
manipulation checks, one image for each condition was selected. As the second treatment was a political stimulus, the chosen images were non-political to prevent biased answers. In total, 39 people participated in the pre-test that were excluded from the main study as they would have become sensitized to the experimental treatment.
Treatment II. Individuals were exposed to an extract of a newspaper article consisting of a headline, an image and a short paragraph. The control groups received the true information claim, which was an original newspaper article by the British quality newspaper The Telegraph dating from December 2, 2018 (See Figure 7). The news article was about the German politician Friedrich Merz questioning the NATO’s spending targets. Only the source of the newspaper article and the author was manipulated to prevent biased answers. The length of the article was shortened without changing the meaning of the article. The experimental groups of each
condition received a newspaper article including disinformation (See Figure 6). The article was the same one as the one of the control group, but included one fabricated paragraph aiming at discrediting the Friedrich Merz since he is an incumbent official of the governing Christian Democratic Union (CDU) party and was seen as the potential successor of the current chancellor Angel Merkel. The topic ‘tax evasion’ was chosen as it implies the immoral behavior of a
politician and might affect the public’s attitude toward people holding positions of power and the government.
Independent variable
Information environment. Dummy variables were created to distinguish between true information (= 0) and disinformation (= 1). To see whether the manipulation was successful, the control variable message credibility was included for which I used a scale tested by Appelman and Sundar (2016). According to them, message credibility is defined as an “individual’s
judgement of the veracity of the content of communication” (Appelman & Sundar, 2016, p. 63). Participants were asked to indicate how well the three adjectives (a) accurate (b) believable and (c) authentic describe the content of the article they have read, using a seven-point Likert scale ranging from 1 (describes very poorly) to 7 (describes very well). The three items were combined to one ‘message credibility scale’, which is highly reliable, Cronbach’s alpha = .87 (Appelman & Sundar, 2016, p. 73). To check if participants perceived the experimental condition (political disinformation) as intended, two independent samples t-tests were conducted with information environment (political disinformation vs. true political information) as independent variable, and message credibility as dependent variable. A significant effect of message credibility was found on information environment (dummy: political disinformation) t (99) = p < .05. Participants perceived political disinformation as significantly less credible (M = 3.99, SD = 1.43) than true political information (M = 4.03, SD = 1.14). This indicates that the manipulation of information environment was successful.
Emotions. Dummy variables were created to represent the induced emotions aversion
condition (= 1) and enthusiasm condition (= 0). Although the images were pre-tested, the control
variable emotion induction was included to see whether the manipulation was successful. The participants were asked to indicate on a 5-point Likert scale ranging from 1 (disagree strongly) to 5 (agree strongly) including the option ‘don’t know’ (= 6) to what extent the image made
them feel different dimensions of aversion and enthusiasm. Marcus et al. (2000) recommend to measure the enthusiasm dimension using the three indicators enthusiastic, hopeful and proud and for the aversion dimension the indicators angry, disgusted and frustrated. Instead of frustrated, I included outraged’ as a third dimension (Weeks, 2015). In order to see whether the items
measure two separate concepts, I conducted a principal axis factor analysis (PAF). As I used separate scales to measure the emotional arousal after each condition, I combined the two scales to one, created six new variables and specified 6 (‘don’t know’) as a missing value. Then I conducted the principal axis factor analyses (PAF) using the six items proud, enthusiastic,
hopeful, angry, disgusted and outraged. The analysis shows that the 6 items form two
uni-dimensional scales. The first component (eigenvalue 3.99) and the second component (eigenvalue 1.03) have an eigenvalue above 1 and there is a clear point of inflexion after the second component in the scree plot. Together, these factors explain 83.71% of the variance in the original variables (items). After an oblimin rotation, all items correlate positively with the first factor, the variable ‘outraged’ has the strongest association (factor loading is .95). Therefore, it appears that proud, hopeful and enthusiastic - form the enthusiasm scale and the last three items – angry, disgusted, and outraged - form the aversion scale (See Table 1). Reliability of both scales, the enthusiasm scale (Cronbach's alpha =.88) and of the aversion scale (Cronbach’s
alpha =.91) is good.
Table 1
Factor Loadings for Principal Axis Factor Analysis (PAF) With Oblimin Rotation
Enthusiasm scale Aversion scale
Hopeful -.89
Enthusiastic -.75
Outraged .95
Angry .77
Disgusted .76
To check if participants perceived the emotion induction as intended, two independent samples t-test were conducted with emotion induction (aversion condition vs. enthusiasm condition) as independent variable, and enthusiasm scale and aversion scale as dependent variables. A significant effect of enthusiasm was found on emotion induction t (91) = p < .001. Enthusiastic participants perceived emotion induction as significantly more enthusiastic (M = 3.03, SD = 1.15) than aversive (M = 1.25, SD = 1.15). This indicates that the manipulation of emotion induction was successful. Aversive participants perceived emotion induction as more aversive (M = 4.12, SD = 0.90) than enthusiastic (M = 1.99, SD = 1.09). However, the effect of aversion on emotion induction t (89) = p < .064 was not significant, which indicates that the manipulation was not successful.
Personality. To avoid survey fatigue, participant’s personality was measured using the brief Ten Item Personality Measure (TIPI) (Gosling, Rentfrow, & Swan, 2003). Each of the Big Five personality traits are measured with two dimensions. For extraversion, participants had to indicate the extent to which they agree with the statement “I see myself as extraverted or
enthusiastic” (standard item) and “I see myself as reserved or quiet” (reverse-coded) on a 7-point Likert scale ranging from 1 (disagree strongly) to 7 (agree strongly). The dimensions for
agreeableness are (a) sympathetic, warm and (b) critical, quarrelsome, for conscientiousness (c)
dependable, self-disciplined and (d) disorganized, careless, for emotional stability (e) calm,
new experiences, complex and (h) conventional, uncreative. First, the reverse-scored items were recoded and formed into a new scale (i.e. TIPI measure) for each personality trait.
Dependent variable
Political trust. Participants were asked to rate their level of trust in a number of
institutions on an 11-point Likert scale ranging from 0 (no trust at all) to 10 (complete trust). The list includes political institutions namely political parties, the national parliament (Bundestag) and politicians as well as impartial institutions namely the legal system and the police. As explained before, political and impartial institutions have to be distinguished theoretically. However, both are linked to another in the mind of people and thus perceived as political (Marien, 2011). Therefore, the five items form one political trust scale. The items are based on the political trust scale used within the European Social Survey (ESS), which originally included more than five items. The items ‘European Union’ and ‘United Nations’ were however excluded from the scale as the focus of this study is on citizens level of trust in German political
institutions. The reliability of the scale is good (Cronbach’s alpha = .87).
Control variables
Socio-demographics. To control my models, I included gender (0 = male, 1 = female, 2 = other) as well as year of birth as an open-ended question. Gender was recoded into a dummy variable (0 = male, 1 = female). The item ‘other’ had to be excluded from the analysis. The variable year of birth was recoded into the continuous numerical variable age in numbers of years. As education is relevant to cognitive processes and thus opinion formation, participants were asked to indicate the highest educational level they have successfully completed. The variable was coded through 0 = Primary education (end of compulsory education), 1 = Lower
secondary education (e.g. Hauptschule, Realschule), 2 = Upper secondary education (e.g. Ausbildung, Berufsfachschule, Gymnasiale Oberstufe), 3 = Tertiary education (e.g. Berufsakademie, Fachhochschule, Universität). The variable had to be excluded from the analysis as it correlated too strongly with the dependent variable.
Political demographics. Partisanship is an important heuristic for cognitive processing and as previous research suggests related to political misperceptions (Nyhan & Reifler, 2010; Weeks & Garrett, 2014). Thus, participants were asked to self-assess their left-right placement using a commonly used 10-point Likert scale ranging from 1 (far left) to 10 (far right). The question wording was taken from the European Social Survey 2016. Despite that, a measure of
political sophistication as an indicator of individual’s level of political awareness was included
as it influences how individuals engage with political information (Zaller, 1992).
Political interest, is measured through a variable coded 1 = very interested, 2 = somewhat
interested, 3 = not much interested, 4 = not at all interested and 5 = don’t know. In order to
include the variable in the analysis, I reverse coded it so that the direction of the scale
corresponds to the direction of the other scales and specified 5 as a missing value. The scale was treated as continuous as distances between the items are equal and to avoid creating dummy variables, which are difficult to interpret. Politically interested people consume more political information, consider more of the political information they engage with, and think more seriously about it (Chaiken, 1980; Luskin, 1990; Petty & Cacioppo, 1979).
Political knowledge was measured with three items asking about the German political
system including an item about the main function of the Federal Council, the person currently holding the position of the Federal President and the name of the institution holding the
‘false answer’, 1 = ‘true answer’ and then added to one ‘political knowledge scale’. The variable had to be excluded from the analysis as it was too strongly correlated with the dependent
variable.
Randomization check
To check if participants’ age and political interest was comparable across conditions with different types of information, an independent samples t-test was conducted. This t-test had information environment (political disinformation vs. true political information) as independent variable, and age as dependent variable. The independent samples t-test showed that participants’ mean age was not significantly different across the conditions with different types of
information, t (98) = 0.50, p = .551. The second t-test shows that participants’ mean political interest was not significantly different across the conditions, t (99) = -1.41, p = .991. This means that randomization of participants across conditions was successful in terms of participants’ age and political interest.
Robustness check
There might be a reason why the results displayed in Table 1 are not representing political trust adequately. As the target of the created political disinformation was the
conservative politician Friedrich Merz, who is part of the Christian Democratic Union (CDU) of Germany, participant’s partisanship might have influenced the extent to which the participants were affected by the stimuli. As mentioned before, people evaluate the veracity of a news story based on their political ideology, and tend to reject stories conflicting with it (Flynn, Nyhan, & Reifler, 2017; Kahan, et al., 2012; Kahan, 2013). As individual’s partisanship strongly correlates with the dependent variable, the variable was not included as a control variable in the regression analysis. Instead, a robustness check was conducted to see whether individuals’ partisanship (M
= 4.15, SD = 1.52) changes the effects discussed above. In all models, the interaction effects are not significant indicating that the regression analysis can be considered robust.
Results
Multiple linear regression analyses were conducted to test if exposure to political
disinformation, in conjunction with personality and emotions, significantly predicted individuals' political. On average, participants scored 5.68 on the political trust scale (SD = 1.44) indicating a moderate level of trust in political institutions.
Table 2 presents the results of a linear regression analysis testing the direct effect of the experimental treatment on political trust (Model 1). Secondly, it includes the results of multiple linear regression analyses testing the direct as well as the interaction effects between the
moderating variables agreeableness (M = 4.62, SD = 0.87), emotional stability (M = 4.90, SD = 1.19), conscientiousness (M = 5.36, SD = 1.12), openness to experience (M = 5.30, SD = 1.03), and extraversion (M = 4.62, SD =1.27) and the experimental treatment (Model 2 and 3). And thirdly, the results of the direct effects of aversion and enthusiasm on political trust as well as their interaction with the experimental treatment are displayed (Model 4 and 5). In these
regression models, the control variables age, gender (dummy: female) and political interest (M = 3.30, SD = 0.66) were added to control for spurious effects and account for possible group differences caused by the randomization of participants to the different treatments.
Before conducting the analyses, I tested whether the models meet the assumptions of multicollinearity, homoscedasticity, and normally distributed residuals. Model 1 meets the assumptions for multicollinearity as the VIF value is not substantially greater than 10 and the Tolerance is above the threshold of 0.2. The bell-shaped curve in the histogram shows that the
residuals are normally distributed. Lastly, the scatterplot indicated that there is no lack in
homoscedasticity. Therefore, Model 1 as well as the other regression models meet the necessary assumptions to conduct the analyses. The results of the first analysis testing the direct effect of the experimental treatment on individual’s political trust indicate that Model 1 is not significant
F (1, 194) = 0.25, p = .621. Therefore, the assumption stated in the beginning is not supported
(H1).
Model 2, to which personality and the control variables were added, is significant, F (9, 194) = 3.92, p < .001, and can be used to predict individual’s political trust. 16 percent of the variation in the dependent variable can be predicted on the basis of the experimental treatment, the five personality traits and the control variables (adj. R2 = .12). There was still no significant direct effect of the experimental treatment on political trust found, b = -0.01, b*= -0.00, t = 2.54,
p = .832, 95% CI [-0.46, 6.07]. The results indicate that conscientiousness is the most influential
and only personality trait that has a significant, but weak association with trust in institutions, b* = 0.17, t = 2.54, p < .05, 95% CI [0.05, 0.40]. For every unit increase in conscientiousness, trust in political institutions increases by 0.22 while controlling for age, gender (dummy: female) and political interest.
Model 3 testing for the interaction effects is significant, F (14, 194) = 2.95, p < .001 and can be used to predict 19 percent of the variance in political trust (adj. R2 = .12). Compared to the previous model, conscientiousness is still significant and weakly associated with political trust, b
= 0.27, t = 2.19, p < .05, 95% CI [0.03, 0.50]. For every unit increase in conscientiousness, trust in institutions increases by 0.27. Despite that, emotional stability has a significant, but weak association with political trust b = 0.24, t = 1.98, p = .05, 95% CI [0.001, 0.49]. For every unit increase in emotional stability, political trust increases by 0.24.
Individuals scoring high on conscientiousness and emotional stability have higher levels of political trust. However, no significant interaction effects were found. Hence, it can be assumed that the relationship between the experimental treatment and political trust does not depend on individual’s personality, which is against the expectations stated in the beginning (H2a-H2e). For all these effects, the other independent variables are assumed to be held constant.
Model 4 with the experimental treatment, aversion (M = 3.30, SD = 1.40) and enthusiasm (M = 2.12, SD = 1.25) as independent variables and political trust as the dependent variable is significant, F (6, 165) = 2.75, p < .01, explaining 9 percent of the variance in political trust (adj. R2 = .06). The analysis revealed that there is neither a direct effect of aversion,
b* = 0.04, t = 0.37, p = .711, 95% CI [-0,15, 0.23], nor of enthusiasm b* = -0.01, t = -1.11,
p = .914, 95% CI [-0.24, 0.21], on political trust.
Lastly, Model 5 testing for an interaction effect between aversion and enthusiasm and the experimental treatment on political trust is significant, F (8, 165) = 2.68, p < .01 and explains 12 percent of the variance in the dependent variable (adj. R2 = .08). Compared to the previous model, a significant negative relationship between aversion and political trust was found.
A significant interaction effect between the experimental treatment and aversion was found, b = 0.07, t = 2.45, p < .05, 95% CI [0.01, 0.13]. The positive effect of the interaction between the experimental treatment and aversion (b = 0.07) signals that the effect of the
experimental treatment is more strongly positive or less negative at higher levels of aversion (See Figure 1).
Figure 1. Regression line for the interaction effect between the experimental treatment and
aversion on political trust.
Compared to aversion, there was no interaction effect between enthusiasm and exposure to political disinformation on political trust found, b = -0.10, t = -1.50, p = .137, 95% CI [-0.23, 0.03]. For all these effects, the other independent variables are assumed to be held constant.
Discussion
The aim of this paper was to contribute to the understanding of the implications of the increasing dissemination of political disinformation for democratic societies. This paper tested for the direct effect of exposure to political disinformation on political trust. To account for individual differences in the susceptibility to political disinformation, individual’s personality was added to the analysis as it is the most stable predictor of a person’s political profile. Therefore, the paper tested for the interaction effects between personality and exposure to political disinformation on individual’s level of political trust. However, the theory of Affective
Intelligence by Marcus et al. (2000) implies that an individuals’ emotional judgment precedes
their cognitive processes. Therefore, emotions as the least stable indicator of individuals’
political information processing and thus political judgment were also taken into account. Hence, this study tested for the interaction effect between emotions and exposure to political
disinformation on political trust.
To answer the hypotheses, a survey-embedded online experiment with two treatments was conducted. First of all, the participants were exposed to images inducing emotions of aversion and enthusiasm. Secondly, the participants were exposed to a newspaper article that either contained political disinformation or true political information.
First, the direct effect of exposure to political disinformation on political trust was not significant. This finding is not in line with the results in Balmas’ study (2014) stated at the beginning. However, the results come as no surprise as other intervening factors influencing individuals’ political information processing and therefore their political judgement have to be taken into consideration.
The analysis regarding individual’s personality has shown that openness to experience and extraversion are not related to political trust. In regard to emotional stability, the results reveal that neurotic individuals have lower levels of political trust which is in concordance with the findings by Freitag and Ackermann (2016). Their study focusing on cantonal institutional trust in Switzerland shows that neuroticism negatively correlates with cantonal institutional trust (p. 718), whereas emotionally stable individuals place especially high trust in national
governments (Freitag, & Ackermann, 2016, p. 719). Although no significant result for
agreeableness on trust was found, other studies found agreeableness to strongly correlates with higher levels of political trust (Freitag, & Ackermann, 2016; Mondak & Halperin, 2008). Gabriel and Völkl (2005) explain that conscientious people have higher levels of political trust as they are conservative, adhere to norms and rules and therefore accept authorities. Their findings are supported by this paper, which shows that conscientious individuals have higher levels of political trust. Therefore, the results contribute to an understanding of the relationship between individual’s personality and their level of political trust.
Against the expectations stated in the beginning, the interaction effects between
personality and political disinformation on a person’s level of political trust were not significant. One explanation for these findings could be that the majority of participants had a higher
educational level and was moderately to highly interested in politics. Thus, most participants had higher cognitive abilities to process the information they were exposed to and therefore might have been less susceptible to political disinformation, which is in line with the findings of Pennycook and Rand (2018) mentioned in the beginning. Moreover, most participants did not believe the article containing disinformation was credible, which might explain why they have not been affected by it. Following Zaller’s (1992), individuals’ change their attitude if they
accept the information they encounter. And this depends on their political awareness, but also on their predispositions. Following theories of motivated reasoning, recent studies have shown that individuals that were exposed to misinformation concordant with their political ideology are more likely to believe in it and therefore more likely to change their attitude (Calvert, 2017; Kahan, 2017; Nyhan & Reifler, 2010; Weeks & Garett, 2014, Weeks, 2015). However, this paper suggests that, in line with Pennycook and Rand (2018), individual’s susceptibility to political disinformation most likely depends on their cognitive abilities to process political information, but not necessarily on their political ideology. Another reason could have been that all participants were induced with emotions, which influences their information processing and thus might have affected the results.
In regard to the direct effect of aversion on political trust, results are inconsistent. The data suggests that there is no significant direct effect for aversion and enthusiasm on political trust. However, when interaction effects between both emotions and exposure to political disinformation are added to the analysis, aversion and political trust correlate negatively. This is in line with prior findings, which suggest that negative affective states (e.g. anxiety, anger, fear, or threat) negatively affect political attitudes and behavior (Huddy, Feldman, and Cassese, 2007; Valentino et. al, 2011). Other studies found that negative affective states lead for instance to a more negative evaluation of a candidate (e.g. Ottati, Steenbergen, & Riggle, 1992). In line with this, Webster (2018) found that anger negatively affects trust in government and the American electorate. In contrast to this, studies found that positive affective states also lead to a positive evaluation of a political candidate (Isabell et al., 2006, Ottati, Steenbergen, & Riggle, 1992). However, the direct effect of enthusiasm on political trust was not significant.
Interestingly, the findings show that individual’s political trust increases following the exposure to political disinformation, when experiencing higher levels of aversion than lower levels of aversion, although this difference is relatively small. This finding contradicts the expectations from the beginning and is not in line with the studies discussed above.
Conclusion
To conclude, the study could not find sufficient support for the assumption that exposure to political disinformation decreases political trust. Individuals’ personality did not moderate the relationship between exposure to disinformation and political trust. In regard to emotions, enthusiasm did not have an influence on the relationship, but interestingly aversive individuals were positively affected by political disinformation, which resulted in higher political trust.
However, it is plausible that a number of limitations might have influenced the results obtained. First of all, the findings have no external validity as the sample constitutes a
convenient sample that is not representative for the population. Although the study employed a highly controlled true-experimental design, some factors might have reduced internal validity of the experiment as for instance survey fatigue (maturation) or answer biases due to social
desirability. In order to avoid survey fatigue, the TIPI scales with two items per trait were included. However, the TIPI scales have lower internal reliability as longer batteries (Gerber et al., 2011b, p. 268), which is why future studies could include the NEO-PI-R or the IPIP
measures. Furthermore, measures of political sophistication, political interest and political knowledge, should be revised as the scales were not reliable.
Although the images were pre-tested, emotions of aversion were not successfully induced. One reason might have been that participants were limited in their personal
introspection, which might have decreased participants objectivity of self-reported emotions. As an alternative to images one could use video clips or ask participants to write about an event that made them angry in the past, or even combine multiple methods to induce participants with emotions. Further data collection is needed to examine how political disinformation affects individual’s political attitudes.
Future studies should take individuals reception and belief in disinformation as this influences their susceptibility to disinformation. Moreover, it would be interesting to include anxiety into the analysis, which is triggered the second affective subsystems, the surveillance system. Studies indicates that anxiety drives a person’s conscious awareness to a problem,
decreases the reliance on predispositions and thus promotes learning (Marcus et al., 2000), which might affect individual’s attitudes even stronger.
References
Aarts, K., Fladmoe, A., & Strömbeck, J. (2012). Media, political trust and political knowledge: A comparative perspective. In T. Aalberg, & J. Curran (Eds.), How media inform
democracy: A comparative approach (pp. 98-118). United Kingdom: Routledge.
Albertson, B., & Gadarian, S. K. (2016). Did that scare you? tips on creating emotion in experimental subjects. Political Analysis, 24(4), 485-491.
Anderson, R. (2014). Happy family. Retrieved from https://pixabay.com/en/water-fight-children-water-play-442257/
Appelman, A. & Sundar, S. S. (2016). Measuring message credibility: Construction and validation of an exclusive scale. Journalism and Mass Communication Quarterly, 93(1), 59-79.
Avery, J. M. (2009). Videomalaise or Virtuous Circle? The Influence of the News Media on Political Trust, The International Journal of Press/Politics, 14(4), 410-433.
Balmas, M. (2014). When fake news becomes real: Combined exposure to multiple news sources and political attitudes of inefficacy, alienation, and cynicism, Communication Research,
41(3), 430-454.
Barbaranelli, C., Caprara, G. V., Vecchione, M., & Fraley, C. R. (2007). Voters’ personality traits in presidential elections. Personality and Individual Differences, 42(7), 1199-1208.
Bodenhausen, G. V., Sheppard, L. A., & Kramer, G. P. (1994). Negative affect and social judgement: the differential impact of anger and sadness. European Journal of Social
Psychology, 24(1), 45–62. doi:10.1002/ejsp.2420240104
Bouchard, T. J. (1997). The genetics of personality. In K. Blum, E.P. Noble (Ed.), Handbook of
Psychiatric Genetics (pp. 273–96). Boca Raton, FL: CRC Press
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. Brader, T. (2006). Campaigning for hearts and minds: How emotional appeals in
political ads work. Chicago: University of Chicago Press.
Campbell, A., Gurin, G., & Miller, W. E. (1954). The voter decides. Evanston: Row, Peterson & Company.
Caspi, A., & Roberts, B. W. (2001). Personality development across the life course: The argument for change and continuity. Psychological Inquiry, 12(2), 49-66.
Caspi, A., Roberts, B. W., & Shiner, R. L. (2005). Personality Development: Stability and Change. Annual Review of Psychology, 56(1), 453–484. doi:
10.1146/annurev.psych.55.090902.141913
Cawvey, M., Hayes, M., Canache, D. & Mondak, J. J., 2017. Personality and Political Behavior. Retrieved from http://politics.oxfordre.com/view/10.1093/acrefore/ 9780190228637 .001.0001/acrefore-9780190228637-e-221
Ceron, A. (2015). Internet, News, and Political Trust: The Difference Between Social Media and Online Media Outlets, Journal of Computer-Mediated Communication, 20(5), 487-503.
Chaiken, S. (1980). Heuristic versus systematic information processing and the use of source versus message cues in persuasion. Journal of Personality and Social Psychology, 39(5), 752-766.
Chipuer, H. M., Rovine, M. J., & Plomin, R. (1990). LISREL modeling: Genetic and
environmental influences on IQ revisited. Intelligence, 14(1), 11–29. doi:10.1016/0160-2896(90)90011-H
Citrin, J., & Stoker, L. (2018). Political Trust in a Cynical Age. SSRN. doi:10.1146/annurev-polisci-050316-092550
Cleary, M. R. & Stokes, S. (2009). Democracy and the Culture of Skepticism: The Politics of Trust in Argentina and Mexico. New York: Russell Foundation.
Converse, P. E. (1972). Change in the American electorate. In A. Campbell & P. E. Converse (Eds.), The human meaning of social change. New York: Russell Sage Foundation. Craig, S. C., Niemi, R. G., & Silver, G. E. (1990). Political efficacy and trust: A report on the
NES pilot study items, Political behavior, 12(3), 289-314.
Easton, D. (1965). A systems analysis of political life. U.S.: John Wiley & Sons. European Commission. (2018). Code of Practice on Disinformation. Retrieved from
https://ec.europa.eu/digital-single-market/en/news/code-practice-disinformation European Union. (2018). A multi-dimensional approach to disinformation. Report of the
independent High level Group on fake news and online disinformation Luxembourg: Publications Office of the European Union.
Flynn, D., Nyhan, B., & Reifler, J. (2017). The nature and origins of misperceptions: Understanding false and unsupported beliefs about politics. Advances in Political
Freitag, M., & Ackermann, K. (2016). Direct democracy and institutional trust: Relationships and differences across personality traits. Political Psychology, 37(5), 707-723.
Gerber, A. S., Huber, G. A., Doherty, D., & Dowling, C. M. (2011a). Personality traits and the consumption of political information. American Politics Research, 39(1), 32–84. doi:10.1177/1532673X10381466
Gerber, A., Huber, G., Doherty, D., & Dowling, C. (2011b). The Big Five Personality Traits in the Political Arena. SSRN. doi:10.1146/annurev-polisci-051010-111659
Hardin, R (1999). Do we want trust in government? In M. E. Warren (Ed.), Democracy and
Trust, (pp. 22–41). Cambridge University Press.
Hetherington, M. (2015). Why polarized trust matters. Forum (Germany), 13(3), 445–458. doi: 10.1515/for-2015-0030
Hetherington, M. J. (2006). Why Trust Matters: Declining Political Trust and the Demise of
American Liberalism. Perspectives on Politics. Princeton University Press, Princeton,
N.J.; Woodstock. doi: 10.1017/S1537592706490149
Hofferbert, R. I. (1937). The study of Public Policy. Michigan: Bobbs-Merill. Huddy, L., Feldman, S., & Cassese, E. (2007). On the distinct political effects of
anxiety and anger. The affect effect: Dynamics of emotion in political thinking
and behavior, 202-230.
Isbell, L.M., Ottati, V.C., & Burns, K.C. (2006). Affect and politics: Effects on judgment, processing, and information seeking. In D. Redlawsk (Ed.) Feeling Politics (pp. 57-86). United Kingdom: Palgrave Publishing Company.
Iyer, A., Leach, C. W., & Crosby, F. J. (2003). White guilt and racial compensation: The benefits and limits of self-focus. Personality and Social Psychology Bulletin, 29(1), 117-129.
John, O. P., & Srivastava, S. (1999). The Big Five trait taxonomy: History, measurement, and theoretical perspectives. Handbook of personality: Theory and research, 2(1999), 102-138.
John, O., Naumann, L., & Soto, C. (2008). Paradigm shift to the integrative Big Five Trait taxonomy. Handbook of Personality: Theory and Research, 114–158. doi:10.1016/S0191-8869(97)81000-8
Kahan, D. M. (2013). Ideology, motivated reasoning, and cognitive reflection. Judgment and
Decision Making, 8(4), 407–424. doi:10.2139/ssrn.2182588.
Kahan, D. M. (2017). Misconceptions, misinformation, and the logic of identity-protective cognition. SSRN Electronic Journal. doi:10.2139/ssrn.2973067.
Kahan, D. M., Peters, E., Wittlin, M., Slovic, P., Ouellette, L. L., Braman, D., & Mandel, G. (2012). The polarizing impact of science literacy and numeracy on perceived climate change risks. Nature Climate Change, 2(10), 732–735. doi:10.1038/ nclimate1547. Kahneman, D., & Tversky, A. (1982). The simulation heuristic. In D. Kahneman, P. Slovic, & A.
Tversky (Eds.), Judgment under uncertainty: Heuristics and biases (pp. 201–208). Cambridge University Press. doi: 10.1093/oxfordhb/9780195376746.013.0038
Kahneman, T., & Tversky, A. (1974). Judgement Under Uncertainty. Science, New Series,185 (4157). 1124-1131.
Kurdi, B., Lozano, S., & Banaji, M. R. (2017). Introducing the Open Affective Standardized Image Set (OASIS). Behavior Research Methods, 49(2), 457-470.
Lau, R. R., & Redlawsk, D. P. (2001). Advantages and Disadvantages of Cognitive Heuristics in Political Decision Making. American Journal of Political Science, 45(4), 951.
Lecheler, S., Schuck, A. R. T., & De Vreese, C. H. (2013). Dealing with feelings: Positive and negative discrete emotions as mediators of news framing effects. Communications, 38(2), 189–209. doi:10.1515/commun-2013-0011
Lin, N. (2002). Social capital: a theory of social structure and action (structural analysis in the
social sciences). A theory of social structure and action. doi: citeulike-article-id:244822
Liu, C. & Stolle, D. (2017). Social capital, civic culture and political trust. “What Kinds of Trust Does a Democracy Need? Trust From the Perspective of Democratic Theory.” In S. Zmerli, & T.W.G. van der Meer (Ed.), Handbook on Political Trust (pp. 338-352). United Kingdom: Edward Elgar Publishing.
Luskin, R. C. (1990). Explaining Political Sophistication. Political Behavior, 12(4), 331-361. MacKuen, M. B., Marcus, G. E., Neuman, W. R., & Miller, P. R. (2010). Affective Intelligence
or Personality? State vs. Trait Influences on Citizens’ Use of Political Information. Annual Meeting of the American Political Science Association. doi:10.1016/S1130-8621(06)71222-0
Marcus, G. (2002). The sentimental citizen: Emotion in democratic politics. Pennsylvania State University Press.
Marcus, G. E., MacKuen, M., & Neuman, W. R. (2011). Parsimony and complexity: Developing and testing theories of affective intelligence. Political Psychology, 32(2), 323-336. Marcus, G., & MacKuen, M. B. (1993). Anxiety, enthusiasm, and the vote: The emotional
underpinnings of learning and involvement during presidential campaigns. American
Political Science Review, 87(3), 672–685.
Marcus, G., MacKuen, M., & Neuman, W. R. (2000). Affective intelligence and political judgment. University of Chicago Press.
Marien, S. (2011). Does political trust matter? An empirical investigation into the relation between political trust and support for law compliance. European Journal of Political
Research, 50 (2), 267-291.
Marien, S. (2017). The measurement equivalence of political trust. In S. Zmerli, & T. W.G. van der Meer (Ed.), Handbook on political trust (pp. 89-101). United Kingdom: Edward Elgar Publishing.
McCrae, R. R., & Costa, P. T. (2003). Personality in adulthood: A five-factor theory perspective. New York: Guilford Press.
Mondak, J. J., & Halperin, K. (2008). A framework for the study of personality and political behaviour. British Journal of Political Science, 38(2), 335-362. Mondak, J. J., Hibbing, M. V., Canache, D., Seligson, M. A., & Anderson, M. R. (2010).
Personality and civic engagement: An integrative framework for the study of trait effects on political behavior. American Political Science Review, 104(1), 85-110.
Mondak, J. J., Hibbing, M. V., Canache, D., Seligson, M. A., & Anderson, M. R. (2010).
Personality and civic engagement: An integrative framework for the study of trait effects on political behavior. American Political Science Review, 104(1), 85–110. doi:
10.1017/S0003055409990359
Nabi, R. L. (2003). Exploring the framing effects of emotion: Do discrete emotions differentially influence information accessibility, information seeking, and policy preference?.
Communication Research, 30(2), 224-247.
Newman, N. (2017). Overview and Key Findings: Digital News Report, Reuters Institute for the Study of Journalism. Retrieved from
Newton, K., & Zmerli, S. (2011). Three forms of trust and their association. European Political
Science Review, 3(2), 169–200. https://doi.org/10.1017/S1755773910000330
Niemi, R. G., Craig, S. C., & Mattei, F. (1991). Measuring internal political efficacy in the 1988 National Election Study. American Political Science Review, 85, 1407-1413.
Norris, P. (2011). Democratic deficit: Critical citizens revisited. Democratic Deficit: Critical
Citizens Revisited (pp. 1–335). Cambridge University Press.
doi:10.1017/CBO9780511973383
Norris, P. (2017). The conceptual framework of political support. In S. Zmerli, & T.W.G. van der Meer (Ed.), Handbook on Political Trust (pp. 19-32). United Kingdom: Edward Elgar Publishing.
Nyhan, B., & Reifler, J. (2010). When corrections fail: The persistence of political misperceptions. Political Behavior, 32(2), 303-330.
Ottati, V. C., Steenbergen, M. R., & Riggle, E. (1992). The cognitive and affective components of political attitudes: Measuring the determinants of candidate evaluations. Political
Behavior, 14(4), 423–442. doi: 10.1007/BF00992043
Petty, R. E., & Cacioppo, J. T. (1979). Issue involvement can increase or decrease persuasion by enhancing message-relevant cognitive responses. Journal of Personality and Social
Psychology, 37, 1915-1926.
Petty, Richard E., & Cacioppo, J. T. (1984). The effects of involvement on responses to argument quantity and quality: Central and peripheral routes to persuasion. Journal of
Personality and Social Psychology, 46, 69–81.
Plomin, R., & Caspi, A. (1999). Behavioral genetics and personality: Handbook of personality: Theory and research (2nd Edition). In L. A. Pervin, & O. P. John (Eds.), Handbook of
