Who ‘likes’ populists? Characteristics of adolescents following right-wing populist actors on Facebook

ABSTRACT

Past research has primarily focused on the positive potential of Social Network Sites (SNS) for democratic citizenship. However, SNS have also become key territories for right-wing populist actors (RWPA). Yet we almost completely lack research on adolescents’ use of SNS in the context of political populism. In this study, we draw a profile of adolescents who follow RWPA on Facebook incorporating three key predictors: political trust, personality traits as measured by the Big Five, as well as motivations to use SNS. A paper–pencil survey (N = 294) of 15–20-year-old school students was implemented. In line with hypotheses, we found that followers of RWPA scored low on political trust, informational motivations to use SNS, and the personality trait openness, but scored high on extraversion and participatory motivations to use SNS. In demographic terms, we found some evidence that lower educated and male students were more likely to follow RWPA. The important role of cognitive engagement with political information for the quality of participation on SNS is discussed.

KEYWORDS:

• Social network sites
• populism
• politicians on SNS
• online participation

Social Network Sites (SNS) have become the most important channels for young people to gather information about politics (Pew Research Center, 2016) and to participate in political action (Vissers & Stolle, 2014). On SNS, politicians and political parties disseminate their messages directly and unfiltered to their followers. While such new connections between political actors and young citizens provide a chance to re-connect adolescents to the political sphere (Heiss & Matthes, 2016), some dangers exist. In particular, SNS provide ideal communication tools for the simplistic messages of populist political actors (Bartlett, 2014; Engesser, Ernst, Esser, & Büchel, 2016). Past research indicates that young people already make up a sizeable share of populist parties’ constituencies and are hence a key target group for populist campaigning (Arzheimer & Carter, 2006). This is alarming, as adolescence is a time when young people possess little knowledge, but shape and develop their political attitudes (Eckstein, Noack, & Gniewosz, 2012). In this process of identity building, they are highly vulnerable and susceptible to RWPAs’ campaigns, which tend to simplify complex realities (Canovan, 1999), disseminate political cynicism (Rooduijn, van der Brug, & de Lange, 2016), and increase racial stereotypes (Matthes & Schmuck, 2015).

Despite the abundance of research on SNS, we have identified three major research gaps in the literature. First, research so far has almost exclusively focused on offline populist support, especially in terms of voting (e.g., Arzheimer & Carter, 2006; Lubbers & Scheepers, 2000; Rooduijn et al., 2016). However, especially young people have shifted their participatory activities to SNS and RWPA have identified SNS as new key territories (Engesser et al., 2016). Second, most studies dealing with SNS participation have focused only on positive aspects of SNS. However, reading, liking, and sharing populist messages which simplify political problems, fuel cynicism, and promote negative stereotypes may not positively influence young adolescents’ political socialization. In fact, the negative effects of such messages are well documented (e.g., Hameleers, Bos, & de Vreese, 2016). Third, and most important, we still lack studies on adolescents’ political SNS use (Ekström, Olsson, & Shehata, 2014; Östman, 2013). However, adolescents are the most active users of SNS and these sites play a key role in their early political socialization. The political activities they perform and the content they expose themselves to have great influence on their future political involvement (Eckstein et al., 2012). What groups of adolescents use SNS for what kind of political activities is hence a pressing question for the future of our democracies.

Against this background, our study attempts to characterize those adolescents that connect to RWPA on SNS. Drawing on past research, we mainly focus on three distinct factors. First of all, political trust has been found to be an important antecedent of both using the online space for political participation (Cicognani, Zani, Fournier, Gavray, & Born, 2012) and support populist parties (Rooduijn et al., 2016). According to the fueling discontent argument, policy dissatisfaction and exposure to populist messages may mutually reinforce each other over time (Rooduijn et al., 2016). As we will explain in more detail below, we therefore theorize that adolescents’ low political trust fuels the following of RWPA on SNS (Lubbers & Scheepers, 2000; Swyngedouw, 2001), and that this relationship should increase with rising age (Rooduijn et al., 2016).

Second, personality traits have been found to be important correlates of SNS use (Correa, Hinsley, & De Zuniga, 2010) and political behavior (Mondak & Halperin, 2008). They are shaped in an early period of life and represent ‘relatively stable patterns of thinking, feeling and behavior’ (Bakker, Rooduijn, & Schumacher, 2015, p. 303). Personality traits may be especially influential in adolescence, when political attitudes, such as party affiliation, are still weak (Wolak, 2009). Drawing on the Big Five personality dimensions (Rammstedt & John, 2005), we argue that there is a specific pattern of personality traits characterizing those adolescents who follow RWPA on SNS.

Third, some adolescents may already have developed political motivations to use SNS. Past research has shown that young people may either use SNS passively to seek information, or may be more motivated to use SNS actively, for participatory activities (Bakker & De Vreese, 2011). Whereas informational motivations are strongly related to cognitive elaboration (Eveland, 2001), participatory motivations may also be driven by non-political user activity (Östman, 2013) and current emotional states (Valentino, Brader, Groenendyk, Gregorowicz, & Hutchings, 2011). Since RWPA often appeal to emotions rather than offering channels for cognitive elaboration, we test the notion that participatory motivations are positively and informational motivations negatively related to adolescents’ following of RWPA.

RWPA on SNS

There are various forms of populism, but all share the same view of an antagonistic relationship between the ‘evil’ elite and the ‘pure’ people. The ‘pure’ people are perceived as a homogeneous entity with unique interests, all exclusively represented by the populist party (Rooduijn et al., 2016). In Europe, right-wing populism expresses these characteristics especially by campaigning against the European Union and immigrants, who are not part of the ‘pure’ people. In demographic terms, there is strong evidence that certain groups are more susceptible to right-wing populist campaigns: younger and older individuals, men, and less-educated people (Arzheimer & Carter, 2006). The increasing share of young people supporting populist parties is of special concern. For populist actors, young people are important targets. They tend to lack a clear party identification (Wolak, 2009) because they have comparatively low levels of political knowledge and are still in the process of shaping their political identity and attitudes (Eckstein et al., 2012). Moreover, young people score lower on political trust, making them more susceptible to anti-establishment campaigning (Henn & Foard, 2012; Rooduijn et al., 2016).

Research on the role of SNS in the current upswing of populist movements is still scarce, and almost non-existent when it comes to adolescents. Some evidence point in the direction that populist actors are especially successful in utilizing SNS for their purposes. Bartlett (2014) argues that SNS provide ideal platforms for their campaigns, because they are not controlled by ‘the elites’ and content is largely generated by ordinary citizens. Engesser et al. (2016) find empirical support that the network logic of SNS gives populist actor more freedom when attacking the elites and ostracizing others, such as in the use of strong language. They hence conclude that populism may even ‘thrive’ on the logic of connective action. This is also in line with recent findings in content analysis, indicating that some communicative features typical for populist communication are important divers for the virality of politicians’ Facebook posts, such as negativity (Bene, 2016) and political attack (Xenos, Macafee, & Pole, 2016). Other, more positive content characteristics, such as opinion expressions and accounts, are scarce and have no or even negative effects on dissemination (Bene, 2016).

These findings are alarming, as young people have shifted their informational and participative activities to SNS (Pew Research Center, 2016; Vissers & Stolle, 2014). While young people have traditionally been among the key constituencies of populist parties (Lubbers & Scheepers, 2000), this targeting might become even more intensive on SNS. Furthermore, young people are often incidentally exposed to political information on SNS. While this might increase their level of online participation (Valeriani & Vaccari, 2016), the quality of their participatory actions might be questionable. Many online political activities might remain superficial and may not contribute to individual knowledge levels or the overall quality of democratic decision-making (Baumgartner & Morris, 2009; Vitak et al., 2011). For example, when young people engage in uninformed participatory online activities, they might ‘share’ conspiracy posts or ‘like’ posts which ostracize certain ethnic groups. Such uniformed activities may contribute to the success of RWPA to spread their messages on SNS.

In the following sections, we will derive hypotheses and research questions about the predictors of following RWPA. We can only assess the unique characteristics of followers of RWPA if we compare them to the characteristics of individuals who follow non-populist actors (NPA) only and those who do not follow political actors at all. We thus follow a multinomial logit regression logic assessing how our independent variables predict to which group of followers participants belong: The right-wing populist followers (RWPA group), the exclusively non-populist followers (NPA group), or the non-followers (NF group).

Political trust

Past research has shown that political trust is an important predictor of adolescents’ online participation (Cicognani et al., 2012) and general RWPA support (Lubbers & Scheepers, 2000; Rooduijn et al., 2016; Swyngedouw, 2001). Young people who have low trust in the political system may be less likely to engage through traditional channels and more likely to engage in non-institutionalized ‘connective action’ on SNS. However, RWPA may also use SNS to bypass traditional gatekeepers and disseminate strong anti-establishment messages (Engesser et al., 2016). Young people with low political trust may be highly susceptible to such anti-establishment messages of RWPA (Lubbers & Scheepers, 2000; Rooduijn et al., 2016; Swyngedouw, 2001).

Furthermore, there is reason to believe that the relationship between political trust and following right-wing populism will become stronger with age. Rooduijn et al. (2016) have shown in a six-wave panel study that political distrust is both a cause and a consequence of populist party support. According to their fuel discontent argument, ‘citizens become more dissatisfied as a result of being exposed to the populist message and being more susceptible to that message’ (Rooduijn et al., 2016, p. 38). Based on this fueling discontent logic, we argue that the correlation between political trust and following RWPA may become stronger with increasing age. Following RWPA on Facebook is not only a statement of support, but also a predictor of selective exposure to subsequent populist messages. These messages contain anti-establishment information and may hence further fuel an individual’s initial level of distrust.

H1 Compared to NF and followers of NPA, followers of RWPA score lower on political trust.

H2 The negative relationship between political trust and following RWPA will become stronger with age.

Personality traits

Personality traits were found to be related to young people’s SNS use (Correa et al., 2010), different dimensions of political involvement (Mondak & Halperin, 2008), and electoral support for RWPA (Bakker et al., 2015). The most prominent personality concept is the ‘Big Five’ or ‘Five Factor’ model (Quintelier, 2014). This model assumes that there are five key personality dimensions, extraversion, openness, agreeableness, conscientiousness, and neuroticism. Past research suggests that personality traits are important predictors of political behavior in adolescence (Quintelier, 2014). In this period, young people’s political attitudes are still in their infancy, while their personality traits were shaped since childhood (Roberts, Walton, & Viechtbauer, 2006). Personality traits may hence represent comparably stable and reliable predictors of political behavior in adolescence.

In terms of SNS use, extraversion has been identified as a key predictor for the quantity of young people’s expressive activities (Correa et al., 2010). As they engage in multiple activities, they may also be less selective in their choice of political actors to follow. Furthermore, following or ‘liking’ a political actor is also strongly related to expression, as other people in one’s network may see such activities (Marder, Slade, Houghton, & Archer-Brown, 2016). As extraverts tend to not shy away from conflict (Wood & Bell, 2008), they may also be less afraid from actively defending their behavior when following a controversial, populist politician. In the political realm, extraversion was found to be unrelated to political interest or trust, negatively related to political knowledge, but positively related to political discussion and opinionation (Mondak & Halperin, 2008). Hence, individuals high in extraversion may be more likely to participate especially in expressive types of political participation on SNS. However, the finding that extraverts possess less political knowledge may affect the quality of their participatory actions. Strong extraverts may base political decisions to follow or not follow a political actor on impulse and current emotional states, and less on cognitive elaboration (Verplanken & Herabadi, 2001). This may make extraverts more likely to follow RWPA, who tend to disseminate emotionally loaded and simplistic political messages (Hameleers et al., 2016). To sum up, the lack of conflict aversion, high expressive motivation in the context of little political knowledge, and impulsive decision-making may make adolescents high in extraversion more likely to follow RWPA.

H3 Compared to NF and followers of NPA, followers of RWPA score higher on extraversion.

Openness has generally been found to be a positive predictor SNS use (Correa et al., 2010) and informational Facebook use (Hughes, Rowe, Batey, & Lee, 2012). However, these findings may be less applicable for the younger generation, who have grown up with new technologies (Correa et al., 2010). More importantly, openness to experience has been generally linked to higher levels of cognitive orientations. In the political realm, openness was not only found to be positively related to political discussion and opinionation, but individuals high in this trait were also more politically interested, paid more attention to politics and had greater political knowledge (Mondak & Halperin, 2008). Adolescents high in openness may hence start to seek political information from early on, elaborate on this information, and question superficial political solutions provided by populist parties. Furthermore, the trait openness is strongly related to liberal political attitudes (Gerber, Huber, Doherty, & Dowling, 2011; Mondak & Halperin, 2008), which may shield adolescents from fear messages regarding a cultural identity loss (e.g., through immigration) and the erosion of national borders (e.g., Europeanization). We thus hypothesize:

H4 Compared to NF and followers of NPA, followers of RWPA score lower on openness.

What is more, there is reason to believe that less agreeable individuals may be more likely to follow RWPA. Agreeable individuals tend to be co-operative, sympathetic, and altruistic. In the political realm, agreeableness was found to be positively related to political trust and external efficacy and encourages non-conflictual political participation, such as contacting political officials, but is negatively related to political discussion (Mondak & Halperin, 2008). There is hence some evidence to assume that young people high in agreeableness may be more likely to follow political actors on SNS in general, as simply following a political actor is per se a rather passive type of political participation. However, there is also some evidence that young people may be less likely to follow RWPA. Bakker et al. (2015) have found convincing evidence that low levels of agreeableness significantly predict populist voting based on four samples in three countries. They explain their finding arguing that low agreeable individuals tend to be socially and politically distrusting, egoistic, intolerant as well as uncooperative, and that these traits fit the anti-establishment and exclusionist messages of RWPA. In line with these findings, we hypothesize that adolescents low in agreeableness may be also more likely to support RWPA by following them on Facebook.

H5 Compared to NF and followers of NPA, followers of RWPA score lower on agreeableness.

There is less evidence how neuroticism and conscientiousness might be related to following RWPA. Generally, individuals high in conscientiousness tend to be less politically informed and efficacious (Mondak & Halperin, 2008) and use Facebook generally less for political information (Hughes et al., 2012). Neuroticism was found to be unrelated to SNS use among young people (Correa et al., 2010). As individuals high in this trait tend to be anxious and distressed, they may generally avoid negative political messages (Riding & Wigley, 1997). Nevertheless, there is also little evidence that conscientiousness and neuroticism are related to populist support (Bakker et al., 2015). Hence, we pose a research question:

RQ1 How are conscientiousness and neuroticism related to following RWPA?

Political motivations

Past research has shown that SNS may serve two political functions. First, they play a key role in the provision of political information in a new hybrid media system (Chadwick, 2013). Second, they provide platforms for new types of political participation. This is especially true for political actors’ SNS accounts: They provide information on the policy standpoints of political candidates and parties, but they also allow followers to support certain political causes by liking, sharing, and commenting political posts. Past research has convincingly shown that political information on SNS is positively related to political participation (e.g., Ekström et al., 2014). However, Östman (2013) also found evidence that not all participatory activity on SNS may be driven by information. In a second path, participatory activities can also be driven by engagement in non-political content creation, fostering the behavioral dimension of political participation. Such participatory activities, which are not based on elaboration, may be more easily triggered by emotional cues or simple language, which are often used by populist politicians (Hameleers et al., 2016).

Based on these findings, there is reason to believe that informational and mere participatory user motivations may exert distinct effects on the quality of SNS participation. If young people engage in active information seeking they may be more likely to critically process political content and increase their knowledge (Eveland, 2001). In-depth processing and knowledge acquisition may make young people less susceptible to RWPA messages (Matthes & Schmuck, 2015). That is, those who look for detailed information in order to elaborate, may not be able to satisfy their needs when following RWPA. By contrast, individuals with low informational motivations may be more susceptible to simple cues (Eagly & Chaiken, 1993), such as strong negative emotions.

H6 Compared to NF and followers of NPA, followers of RWPA score lower on informational motivation.

Second, participatory motivations are related to the behavioral dimension of political SNS use. As compared to informational activities, participatory SNS activities, such as liking and sharing political content, do not necessarily involve a high amount of cognitive elaboration (Östman, 2013). Hence, some authors have argued that many political SNS activities can be described as ‘slacktivist’ or ‘feeld good’ activities (Vitak et al., 2011), which hardly contribute to young people’s political involvement (Baumgartner & Morris, 2009). As such activities are not necessarily driven by cognitive involvement, they may be more sensitive to current affective conditions (Valentino et al., 2011), such as feelings of dissatisfaction or anger. Especially anger is a key driver of participatory action. In fact, RWPA aim to provoke and canalize feelings such as anger on their SNS. Thus, participatory motivations may resonate with the content offered by RWPA on SNS (i.e., promising simple solutions, spurring protest against mainstream politics, fueling anger). Young people with strong participatory motivations may hence be more likely to engage with RWPA on SNS.

H7 Compared to NF and followers of NPA, followers of RWPA score higher on participatory motivation.

Method

A total of 294 pupils between 15 and 20 years (M_Age = 17.35, SD = 1.17; 33.7% male; 32.3% non-academic high schools) from different Austrian high schools completed the paper and pencil questionnaire. The study was part of a research course and students helped to collect the data. We selected schools representing all three major school types in Austria, including college-bound academic schools, vocational schools, and part-time vocational schools. Participants were recruited by contacting representatives of the target schools who were fully briefed about the study. Data collection was implemented at the schools. The questionnaires were handed out in the classroom and completed under teacher observation. Even though participation was voluntary and we did not provide any incentive, the response rate was high (close to 100 percent). Eight returned questionnaires were largely incomplete and hence removed from the sample. The final sample size was 294. However, the number of observations was reduced in the different models we ran, as cases with single missings were excluded from the statistical analysis.

In Austria, there are six parties with seats in the parliament and two of them can be defined as right-wing populist: the Austrian Freedom Party (FPÖ) and the Team Stronach (Schmuck, Matthes, & Boomgaarden, 2016). The FPÖ is well established in the Austrian political landscape and is well-known for its attacks against minorities and its calls to expel immigrants from Austria (Schmuck & Matthes, 2015). The Team Stronach, a smaller, euro-skeptic party, can be defined as a neo-liberal, anti-establishment populist party with less of a focus on nationalism (Schmuck et al., 2016). However, single members of the Team Stronach (including their chairman) have constantly gained media attention with anti-immigrant, homophobe, and sexist expressions. There are two non-populist oppositional parties: The Austrian Green Party and the NEOS (The New Austria). Furthermore, the Social Democratic Party (SPÖ) and the Austrian People’s Party (ÖVP) are the two big traditional parties which currently form the national coalition government.

Measures

Following politicians: The dependent variables assessed whether the pupils followed certain types of political actors. Table 1 provides an overview how many young people followed what kind of political actors (candidate or organization for each party). This was assessed using a total of 12 items. For each political party which is represented in the Austrian parliament (six in total), we asked whether the pupils (1) followed a politician of that party and (2) whether they followed the party organization’s or its respective youth organization’s Facebook page. To utilize these 12 items for multinomial logit regression, we created a categorical variable which included 3 distinct groups of people. The first group was composed of adolescents who followed RWPA on Facebook (30%), the second group followed only NPA (21%), and the third group did not follow any kind of political actors (NF, 49%).

Table 1. Percentage of adolescents in our sample following different political actors on Facebook.

Dependent variable	^aPolitical party	% Following (per party)	% Following (groups)
Right-Wing Populist Actors (RWPA)	FPÖ	25.96	^b29.72 (at least one RWPA)
	Team Stronach	10.84
Non-populist actors (NPA)	The Greens	21.05	^b20.56 (exclusively NPA)
	NEOS	18.47
	SPÖ	28.92
	ÖVP	26.67

^aRefers to the political party a followed political actor (candidate and/or party organization) belongs to.

^bNumbers refer to the groups we used as dependent variables in our regression analysis.

Political trust (α = .88; M = 3.03; SD = 1.28) was measured with four items obtained from Craig, Niemi, and Silver (1990). The items include: (1) I trust that politicians make decisions in the interest of the people, (2) Politicians in Austria keep the promises they have made, (3) Politicians are honest to their voters, and (4) Politicians can be trusted to do what is right, even without being checked by the public.

The Big Five were measured with 15 items obtained from the short version Big Five scale by Rammstedt and John (2005). Extraversion: (1) I tend to be loud and talk a lot, (2) I am outgoing and sociable, (3) I am enthusiastic and easily entertain others (α = .70; M = 4.55; SD = 1.40). Conscientiousness: (1) I fulfill assignments thoroughly, (2) I make plans and implement them, (3) I see myself as a diligent and fast worker (α = .68; M = 4.93; SD = 1.27). Neuroticism: (1) I often feel nervous or unsecure (2) I often feel depressed or down (3) I am worried a lot (α = .76; M = 3.73; SD = 1.62). Openness: (1) I have an active imagination, (2) I am interested in many different things, (3) I am profound and think about things a lot (α = .64; M = 5.06; SD = 1.32). Agreeableness was measured with two items, since one item was excluded due to insufficient reliability: (1) I tend to act cold and reserved, (2) I can act dismissive or harsh (α = .75; M = 3.53; SD = 1.76).

To assess whether participants were motivated to use SNS for information or for participatory activities, participants were asked seven questions (see Table 2). A principal component analysis (PCA) was conducted with oblique rotation. An initial test was run (no rotation) to obtain eigenvalues for each component in the data. Two components had eigenvalues higher than Kaiser’s criterion of 1 and explained 79% of variance (KMO = .86, Barlett’s test of sphericity, χ²(21) = 1499.94, p < .001). Factor loadings, eigenvalues (after rotation), and reliabilities are shown in Table 2. We identified two components, informational and participatory motivations.

Table 2. Summary of PCA (oblique rotation) for political motivations to use SNS (factor loadings smaller than .1 suppressed).

Reasons for SNS use related to politics	Factor 1	Factor 1
Informational motivation
To get political information	.85
To follow current political events	.96
To learn about interesting political perspectives	.91
To see what people in my network think about political issues	.76
Participatory motivation
I use SNS to make a difference in politics		.77
Political SNS engagement allows me to make a difference		.91
Liking and sharing allow me to increase awareness about important political issues		.76
Eigenvalue	3.10	2.40
Cumulative variance	.44	.79
Reliability	α = .90	α = .87
Mean (SD)	3.36 (1.77)	2.62 (1.64)

Note: All items were measured on a 7-point scale.

Age was measured as a continuous variable and gender as a dummy variable (male = 1). To measure education, we created two dummies for academic schools and full-time vocational schools (reference group = part-time vocational schools). Facebook use (M = 5.13, SD = 2.21) was measured as the frequency of Facebook use on a 7-point scale.

Results

Table 3 shows the correlations among the key independent variables. They show weak to moderate correlations, except the one between the two political motivational variables (r = .63, p < .001). We ran multinomial logit regressions to assess how the independent variables predict to which group of followers participants belonged to: The right-wing populist followers (RWPA group), the exclusively non-populist followers (NPA group), or the non-followers (NF group). Multinomial regression allows to directly compare the effects for these three groups of individuals. Results are reported in Table 4. We ran one model for each set of our key independent variables (Models 1–4) and also ran a full model, including all our independent variables simultaneously (Model 5).

Table 3. Intercorrelations (Pearson) among continuous independent variables.

	1	2	3	4	5	6	7	8	9
1 Political trust	–
2 Facebook use	−.05	–
3 Extraversion	.03	.08	–
4 Openness	.09	.04	.29***	–
5 Conscientiousness	.08	.11	.26***	.33***	–
6 Agreeableness	.13*	−.08	−.16**	−.31***	−.21***	–
7 Neuroticism	−.05	.01	−.08	.35***	.04	−.31***	–
8 Informational motivation	.24***	.26***	.12*	.14*	−.06	.05	−.07	–
9 Participatory motivation	.34***	.17***	.18**	.12^+	.13*	.06	−.07	.63***	–

 *p < .05.

 **p < .01.

***p < .001.

 +p < .10.

Table 4. Multinomial logit regressions predicting following RWPA as opposed to following NPA and NF. Odds Ratios reported.

Demographics	Model 1		Model 2		Model 3		Model 4		Model 5
	RWPA vs. NPA	RWPA vs.NF	RWPA vs. NPA	RWPA vs. NF	RWPA vs. NPA	RWPA vs.NF	RWPA vs. NPA	RWPA vs. NF	RWPA vs. NPA	RWPA vs. NF
Age	1.05	0.93	1.04	0.88	1.16	1.03	1.18	0.95	1.16	1.06
Male	1.22	2.37**	1.26	2.45**	1.03	1.96*	1.33	1.96*	1.33	1.79^+
^aVocational school	0.83	0.56	0.70	0.45	0.92	0.60	0.74	0.56	0.77	0.59
^aAcademic school	0.56	0.57	0.53	0.53^+	0.49	0.43*	0.50	0.59	0.35*	0.36*
Facebook use	0.99	1.15*	0.98	1.15*	0.98	1.17*	1.06	1.11	1.03	1.10
Political trust	0.66**	0.94	0.66**	0.95					0.61**	0.71*
^bAge * trust			0.81^+	0.80*
Personality
Extraversion					1.54**	1.39**			1.54**	1.31*
Openness					0.62**	0.92			0.66*	0.89
Conscientiousness					0.97	0.82			0.85	0.77^+
Agreeableness					0.97	1.33**			1.02	1.35**
Neuroticism					1.02	1.01			0.97	1.00
User motivations
Informational motivation							0.73*	1.08	0.74*	1.06
Participatory motivation							1.10	1.36**	1.26	1.54***
Log likelihood	−278.118		−275.68		−266.70		−264.97		−242.13
Observations	283		283		283		282		280

^aReference group = part-time vocational schools.

^bVariables are mean-centered. RWPA = right-wing populist actors, NPA = non-populist actors only, NF = not following any actor.

 *p < .05.

 **p < .01.

***p < .001.

 +p < .01.

A first look at the demographic variables provides some evidence that school type is a significant predictor of following RWPA. Even though the effects are pointing in the expected direction throughout all models, they become stronger when controlling for personality and motivations. In our final model (Model 5), students from academic schools were significantly less likely to belong to the RWPA as opposed to the NPA group and as opposed to the NF group. Furthermore, models 1–4 indicate that male adolescents were more likely to belong to the RWPA group as opposed to the NF group. This relationship is becoming weaker when including more variables, but is still close to significant in the final model.

Political trust

Table 4 (Model 3) indicates that, in line with H1, a one-unit increase in political trust makes adolescents 44% more likely to belong to the NPA group as opposed to the RWPA group. Even though we did not find significant differences between the RWPA group and the NF group in the Model 1, we did find significant differences in our final model, when controlling for psychological and political motivational factors. In line with H2, moderation analysis revealed that the interaction effect between trust and age is significant within the following group and close to significant when comparing RWPA followers with NF (Model 2). We also performed the moderation analysis in our final model (not shown in the table), finding that the effects became stronger. We calculated predicted probabilities to visualize this effect (see Fox & Hong, 2009). Figure 1 shows the effect of political trust for the ages of 16, 18, and 20. The figure indicates that, as expected, the negative relationship between political trust and following RWPA became stronger with increasing age. The relationship between trust and belonging to the NPA or NF groups did not considerably change with age.

Figure 1. Predicted probabilities to follow right-wing populist actors (RWPA), only non-populist actors (NPA) or no political actor at all (NF, non-followers) for different levels of political trust (x-axes) and for different levels of age. Graphs on the left side show results for age = 16, graphs in the middle for age = 18, and graphs on the right side for 20 age = 20. Probabilities are calculated based on Table 4 (Model 2). All other variables are set to their mean values. Dotted lines indicate 95% confidence intervals.

Personality traits

Model 3 in Table 4 shows the results for personality traits. In line with H3, results show that extraversion was indeed positively related to following RWPA. As adolescents’ level of extraversion increased by one, they were 54% more likely to belong to the RWPA group as opposed to the NPA group. Furthermore, a one-unit increase in extraversion increased the likelihood by 39% that students belonged to the RWPA group as opposed to the non-following group. Hence, H3 was confirmed by our data. H4 was partly supported, such as that among followers, a one-unit increase in openness increased the likelihood to belong to the NPA versus the RWPA group by 62%. Furthermore, a one-unit increase in openness increased the likelihood to belong to the NPA group as opposed to the NF group by 49% (OR = 1.49, p > .05). We did not find significant differences between the RWPA group and NF group. Finally, we did not find support for H5. Instead, the results revealed that individuals in the RWPA group and the NPA group (OR = 1.37, p < .01) had higher levels of agreeableness as compared to individuals in the NF group. Finally, we did not find any significant results for conscientiousness and neuroticism (RQ1).

User motivations

Model 4 in Table 4 shows the results of the two motivational variables. Results indicate partial support for H6. A one-unit increase in informational motivation increased the likelihood to belong to the NPA group as opposed to the RWPA group by 38%. Furthermore, as adolescents’ informational motivations increased by one, the likelihood to belong to the NPA group as opposed to the NF group increased by 49% (OR = 1.49, p < .01). The level of informational motivation did not significantly differ between the RWPA group and the NF group. We also found partial support for H7. A one-unit increase in adolescents’ participatory motivation increased the likelihood that adolescents belonged to the RWPA group as opposed to the NF group by 54%. The difference between the RWPA and the NPA group, though pointing in the expected direction, did not reach significance. Furthermore, the difference between the NPA group and the NF group was also not significant.

Finally, one may assume that the relationship between openness and following NPA is partially mediated by informational motivations. We tested this mediation using a dichotomous outcome variable (following NPA vs. not following NPA). We calculated biased-corrected confidence intervals around the assumed indirect effect by means of 5000 bootstrap samples and found support for a partial mediation (b = .07, CI[.01, .18]). Yet the direct effect of openness on following NPA remained significant. We included demographics, Facebook use, the remaining Big Five variables and the remaining motivational variable as covariates in the model. We did not find evidence that the relationship between extraversion and following RWPA is mediated via participatory motivations.

Discussion

This study assessed the profile of young followers of RWPA on Facebook. Results indicate that followers of RWPA in our sample were not driven by informational motivations, but scored high on participatory motivations. Followers of NPA were significantly more motivated to use SNS for political information as opposed to followers of RWPA and NF. This points to the important role of cognitive involvement with political content. Only if adolescents become cognitively involved and seek political information they may actively avoid RWPA and follow NPA instead. Others, who are less involved in information seeking, may either not follow any kind of political actor or engage with RWPA. Interestingly, we found a strong relationship between participatory motivations (i.e., use SNS activities to make a political difference) and following RWPA, but not NPA. This indicates that mere participatory motivations, which are not necessarily accompanied by informational activities (Östman, 2013), may lead to uniformed SNS activities, such as following RWPA. This may also lower the potential of mere incidental exposure on SNS (Valeriani & Vaccari, 2016), as such incidental exposure might result in less qualitative political engagement activities.

Interestingly, the results for personality traits point in a similar direction. We found that extraverted young people were more likely to follow RWPA. This supports past research indicating that even though extraversion may lead to a high quantity of participatory actives on SNS (Correa et al., 2010), the quality of the activities may suffer. This is also in line with findings that extraversion is positively related to impulsive decision-making (Verplanken & Herabadi, 2001) and negatively related to the cognitive engagement with politics (Mondak & Halperin, 2008). A second explanation may be drawn from the finding that the decision to follow a politician may depend on the perceived network approval of the specific politician (Marder et al., 2016). More extraverted adolescents may be less restrained to like controversial (e.g., populist) actors since they may be less afraid of actively defending their behavior and engaging in discussion.

By contrast, we found that, within the group which followed political actors, young people high in openness were more likely to follow mainstream actors only, whereas adolescents lower in openness tended to follow RWPA or no political actors. The relationship between openness and following NPA was partially mediated by the informational motivation to use of SNS. This is in line with past research indicating that young people high in openness are more likely to cognitively process political information and may hence engage in more qualitative political activities, such as following NPA. Furthermore, past research indicated that openness is related to liberal political attitudes (Gerber et al., 2011; Mondak & Halperin, 2008) and behavior (Quintelier, 2014), which may shield young people from the anti-immigrant and anti-establishment messages of RWPA. Against our expectations, we found that low agreeable adolescents were in general less likely to follow any kind of political actors on Facebook, that is, high agreeableness seemed to be a prerequisite to connect with both RWPA and NPA. This finding is surprising, because past research found agreeableness to be negatively related to populist voting (Bakker et al., 2015). This might be related to the different logic of SNS participation. First of all, most of those who follow RWPA do not follow them exclusively, but follow other actors, too. Moreover, whereas voting is simply about casting a vote for a party, following a political candidate is a social act, as this politician becomes part of one’s social network. Individuals who are socially distrustful may be more selective and cautious about what sources they connect to (Marder et al., 2016). Individuals high in agreeableness may be more willing to use SNS to socially connect with politicians, which is also in line with Mondak and Halperin’s (2008) finding that high agreeable individuals are more likely to directly contact political candidates.

We also found support that adolescents who follow RWPA are comparably low in political trust. Furthermore, and in line with the fueling discontent argument (Rooduijn et al., 2016), our results suggest that those low in trust may be gradually turning to RWPA with increasing age. Having liked RWPA Facebook pages once, adolescents are more likely to be exposed to the anti-immigrant and anti-elitist messages of RWPA. Such exposure may then further decrease political trust, which may either lead to demobilization or an increasing support for populist political parties. However, this finding needs to be replicated in panel analyses.

We have also found evidence that male and lower educated adolescents might be more likely to follow RWPA. This is of some concern as young, lower educated men have traditionally been among the key constituencies of RWP parties. On SNS, they may now connect more easily to RWPA and may be more intensely influenced by their campaigns, such as through stereotypical political messages (Schmuck & Matthes, 2015).

In sum, the findings indicate that individual characteristics are critical to assess how adolescents use SNS for political purposes. We found that personal characteristics which are more strongly related to cognitive orientations, including openness and informational motivations, were related to following NPA only. More behavior-oriented characteristics, including extraversion and participatory motivations were more strongly related to following RWPA. One of the reasons for these findings may lie in the specific context of SNS participation. Following a political candidate or party on SNS does not need a lot of effort in terms of time or energy. Instead, such activities can be implemented along the way. If adolescents do not reflect and elaborate on political issues, they may be more susceptible to the simplistic and the emotional language of RWPA and engage in political activities which may not necessarily contribute to their political socialization. To put it bluntly, not all informational and participatory political activities on SNS may have positive effects on adolescents. Hence, it is of critical importance to engage adolescents in deliberative political processes and provide them with the basic skills to seek political information, gain knowledge, and act upon it.

Limitations and future work

The presented study uses a sample of a highly targeted and highly relevant age group, which, to this point, has been scarcely investigated. However, some limitations should be noted. First, this research is of cross-sectional nature and needs to be replicated in panel studies. Second, our findings are based on a convenient sample and we cannot generalize to a broader population. Hence the findings should be replicated with more representative youth samples and compared to findings for older cohorts. Furthermore, findings should be replicated in other countries and for different political parties. For example, findings may critically differ for left-wing populist followers, which we were unable to investigate in the Austrian context. Finally, we have only focused on predictors of following RWPA and, of course, could not assess the actual content young people are exposed to on their profiles. Future research should also apply content analysis to investigate the messages to which young people are exposed to when visiting profiles of RWPA and should test the effects of these content characteristics in experimental designs.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes on contributors

Raffael Heiss is a PhD candidate at the Department of Communication, University of Vienna, Austria. His research interests include digital communication, media effects, and political participation [email: [email protected]].

Jörg Matthes is Full Professor of Advertising Research at the Department of Communication, University of Vienna, Austria. His research interests include advertising effects and public opinion formation and methods [email: [email protected]].