Social media as an exit strategy? The role of attitudes of discontent in explaining non-electoral political participation among Belgian young adults

Abstract

Feelings of dissatisfaction with the political status quo are believed to mobilize citizens into non-institutional political action, such as protest. Still, little is known about whether and how political participation through social media provide an alternative voicing route for discontented citizens. Guided by grievance theory, this article assesses how both electoral exit behaviour (e.g., abstaining) and attitudes of political discontent (political and media trust, political hopelessness and populism) are associated with three modes of non-electoral political participation: institutional, protest and social media participation. An online survey was administered to 720 young adults between 18 and 30 years old in Belgium. A hierarchical regression analysis showed no association between electoral exit and non-institutional participation. Furthermore, attitudes of discontent were found to not uniformly push young citizens away from institutional politics. Our results show that social media provide an important, additional political outlet for young citizens and lend support to the notion of political participation as complementary acts, rather than exclusive ones.

Introduction

One of the main shifts in the research field of political participation is the decline of its traditional manifestations and the simultaneous rise of creative, non-institutional forms of political action (Theocharis and de Moor 2021). Indeed, citizens’ participation patterns have become more diverse, as institutional political participation (e.g., working for a political party) is increasingly complemented by behaviours operating outside the sphere of political institutions (e.g., protesting or boycotting certain products) (Hooghe and Marien 2013; Marien et al. 2010; Oser 2021; Sloam 2014). The current paper aims to analyse this participatory turn towards non-institutional participation by relying on grievance theory and the processes of political discontent as a theoretical framework.

Grievance theory assumes that particular modes of political participation are stimulated by feelings of dissatisfaction or grievances (Craig 1980; Gamson 1968; Gurr 1970). A prominent part of research applying grievance theory focusses on the role of economic grievances, such as deprivation or unemployment (Caren et al. 2017; Ejrnæs 2017; Kern et al. 2015), while less attention has been paid to grievances targeted at the political system (Koopmans 2005). Therefore, we focus on attitudes of political discontent, which broadly refer to negative evaluations of the functioning of democracy and politics and are understood to push citizens away from institutional forms of politics (Kriesi et al. 2020; Portos 2021). Some citizens who experience these feelings of dissatisfaction with the democratic ‘status-quo’, might choose to exit the political game, by abstaining from voting and any other form of political participation. Others however, might choose to employ alternative voicing routes in the form of less common or non-institutional participatory modes (Hirschman 1970; Portos et al. 2019; Wauters 2018). Previous research has established links between various attitudes of political discontent and electoral exit behaviour (e.g., abstaining) (Hooghe and Dassonneville 2018), populist and extremist voting (Hooghe et al. 2011), as well as engagement in non-institutional modes of participation such as protesting (van Stekelenburg and Klandermans 2018).

An important gap in the literature, however, concerns the role of social networking sites within these alternative voicing routes. Over the past decade, it has become clear that social media provide an additional arena for political participation in which citizens can not only express themselves on social and political issues, but also coordinate collective action and exert pressure on public actors and the government (Boulianne 2015; Ekström and Shehata 2018; Theocharis 2015). Moreover, the low-cost, low-threshold and immediate nature of these platforms is often assumed to attract a new audience and to provide a voicing route for an otherwise disengaged public (Keating and Melis 2017). While various authors have pointed to the importance and independent nature of online and social media participation (Gibson and Cantijoch 2013; Theocharis and van Deth 2018b; Waeterloos et al. 2021a, b, c), studies exploring the pathways underlying this particular mode of action are scarce. Moreover, no prior studies have assessed the link between distinct attitudes of political discontent and social media participation. As such, it remains unclear why citizens choose to employ social media as a political outlet and more specifically, whether social media participation is similar to protest as a non-institutional mode of action in providing an alternative voicing route in the context of discontent.

The current study therefore asks whether and how electoral exit behaviour (RQ1), as well as different attitudes of political discontent (political and media trust, political hopelessness and populist attitudes) (RQ2) are related to three types of non-electoral political participation (institutional, protest and social media participation). To answer these questions, we rely on data from an online survey distributed to Flemish young adults between 18 and 30 years old (n = 720).

Given that we do not rely on a representative sample of the Flemish population, our work should be considered as a first exploration of the associations between established indicators of discontent and novel forms of political participation in a new media context. Still, the study of youth participation in this context is of crucial importance, as assumed democratic crises are often brought back to their ‘problematic position’ towards politics. Youth are often disregarded as apathetic of politics and disengaged while in fact, they have always been at the forefront of reimagining political engagement and have been shown to take part in a variety of political activities, often fuelled by disillusionment about established and institutional political processes (Cammaerts et al. 2013; Dalton 2008; Farthing 2010). Youth are therefore believed to be especially drawn to alternative, elite-challenging modes of participation (Ekström and Sveningsson 2019). Moreover, as political participation patterns have been shown to develop at a young age and change over time, detecting shifts in youth participation form an ideal way of taking the temperature of both the current and the future state of participatory democracies (Kirbiš et al. 2017; Ohme et al. 2022). Combined with an overall lack in participation studies focussing on youth (Keating & Melis 2017), our study will therefore assess previously unexplored relationships among a population prone to both attitudinal as well as behavioural shifts in political engagement.

Theoretical framework

Social media within a changing participatory democracy

Political participation plays a central role in all normative conceptions of liberal democracies (Strömbäck 2005). But, scholars differ in their views about which forms of participation are favoured in order to establish a healthy democracy. Within a representative model of democracy, such evaluations are mainly related to the institutional sphere of politics (Tormey 2014), and citizen participation is linked almost exclusively to turnout at elections or party membership (Andersen et al. 2020; Ercan and Gagnon 2014).

However, when only a select few institutional activities are considered to be indicators of a good, active citizenry, this is likely to cause biased evaluations of the functioning of democratic society (Loader et al. 2014; Theocharis and de Moor 2021). As the well-being of a democracy is determined not only by what happens during elections, but also by what happens between them, scholars have drawn attention to political behaviours ‘beyond the ballot box’ (Andersen et al. 2020; Barber 2014; Hilmer 2010; Portos et al. 2019). Indeed, in contemporary democracies, citizens increasingly engage in a range of activities that can be considered political, but are not necessarily located within, nor are they facilitated by the state (Dalton 2008; Vromen 2017). Rather, these non-institutional modes of action actively challenge the institutional status quo (Theocharis and van Deth 2018a; van Deth 2014). Protest participation in particular, through demonstrations and petitions, corresponds with this particular notion of contentious politics (Oser 2021; Tilly and Tarrow 2015).

These abovementioned shifts have been explicitly recognized in the work of Theocharis and van Deth (2018b) as the ‘continuous expansion of the repertoire of political participation’. This study therefore explicitly employs their and van Deth’s (2014) operational definition of political participation as voluntary behaviours done by citizens that are either located in or targeted at the sphere of government, state or politics; aimed at solving collective or community problems; used to express political aims or intentions; or are occurring in a political context. This definition explicitly recognizes the existence of different types of participation and therefore corresponds with our aim to evaluate the role of discontent in the choice between such forms of engagement.

Social media provide an additional, virtual sphere where participation can occur, outside the realm of institutional politics. These platforms enable specific forms of fluid, personalized and creative citizenship (Bennett and Segerberg 2012), and allow citizens to not only mobilize their networks, but also raise awareness about or exert pressure for solving certain political or social issues (Theocharis 2015; Vromen et al. 2016). Consequently, participation through social media has been argued to constitute a distinct form of political participation and hence, an important additional indicator of democratic health (Earl 2014; Waeterloos et al. 2021a, b, c). Indeed, these actions often align with targeted definitions of political participation, as they might be aimed towards decision-making processes by the state or at community issues from within the digital sphere. Moreover, seemingly non-political acts within these networks, such as sharing a personal experience of sexual assault on Twitter, can turn out to be a specimen of political participation based on the underlying motivations and circumstances. While in general, identifying the motivations behind human behaviour is hard, digital media provide important cues about the context and motivations of these behaviours (Theocharis 2015; Waeterloos 2022). In this regard, Theocharis (2015, p. 6) notes that “if we accept these acts as political participation, then it is likely to see not only a greater number of people participating in politics but see among them citizens who have traditionally been disengaged from politics and for whom this type of participation has come to be the only repertoire”.

In this sense, social media have been argued to constitute a virtual ‘third’ or ‘invented’ space, that provides a counterweight against top-down, institutional structures. Moreover, the expressive and demonstrative nature of these online acts shows important parallels with protest participation (Kersting 2013; Wright 2012). Because of their social affordances, social media add a sense of ‘everydayness’ to political participation, centred around one’s own peer networks. As such, they allow for political action that merges ‘the personal’ with ‘the political’, and might keep seemingly uninterested citizens from opting-out of politics (Papacharissi 2010; Wright et al. 2015; Yamamoto et al. 2017).

While some contradicting findings exist about whether social media are indeed applying to a new public (Keating and Melis 2017), it is clear that especially young citizens are drawn to more flexible outings of citizenship outside institutional politics. As their social relations are increasingly enacted through social media networks, their democratic engagement is more likely to occur within these same spaces (Loader et al. 2014). According to Farthing (2010), some youth have even become ‘radically unpolitical’: they are fleeing certain established political spheres, while simultaneously navigating new, often virtual ones. Indeed, some studies have found evidence for exclusive engagement in online participation, both among youth and adult samples, supporting the notion that social media are potentially mobilizing otherwise disaffected citizens (Leyva 2016; Portos et al. 2019). Still other work has shown how youth tend to engage in a variety of activities within hybrid public spaces and media systems, both virtual and physical (Sloam 2014; Waeterloos et al. 2021a, b, c).

In short, the shifting and expanding political practices towards more individualized and often virtual forms of action can be considered a response to the shortcomings of, and disenchantment with, institutional politics (Loader et al. 2014; Theocharis and de Moor 2021; Tormey 2014). By considering these particular forms of political participation, alternative routes towards ‘active citizenship’ become apparent. As such, some young citizens who have given up on traditional politics, and would otherwise remain disengaged, might be attracted to these expressive and creative acts and thus explore new ways of engaging with politics (Portos et al. 2019; Theocharis and de Moor 2021).

The choice to exit the electoral arena is an important behavioural indicator of discontent towards traditional, institutionalized politics. An exit in this regard does not only entail abstaining from voting (Kemmers 2017), but can also include intentional invalid or blank voting, motivated by dissatisfaction with the political system or circumstances (Aron and Superti 2021; Driscoll and Nelson 2014; Hooghe et al. 2011; Kouba and Lysek 2019). Therefore, the following hypotheses are developed regarding the relationship between electoral exit behaviour and non-electoral modes of political participation:

H1

Past electoral exit behaviour will be negatively associated with institutional participation.

H2

Past electoral exit behaviour will be positively associated with protest participation (H2a) and social media participation (H2b).

Attitudes of political discontent

As a way to understand citizen’s pathways towards political participation, several authors have called for more research that examines psychological characteristics as drivers of participation (Miller and Saunders 2016). A specific set of cognitions to be considered, stems from grievance theory, which proposes particular feelings of dissatisfaction as central mobilizing forces for non-institutional participation (Gurr 1970; Kern et al. 2015; van Stekelenburg and Klandermans 2013).

As Parvin (2018) remarks, many current liberal democracies are characterized by a profound disconnection between citizens and the democratic system. Increasingly, citizens feel cut off from traditional political processes, and feel resentful towards it. Such feelings are assumed to be especially pronounced among young people. Therefore, discontent should not only be considered in relation with the economic system, but also applied to democracy and governance (Wauters 2018). This sense of ‘political discontent’ broadly captures citizens’ perception about the functioning of democracy and can therefore be understood as negative evaluations of political actors and institutions (Kriesi et al. 2020; Portos 2021). As the concept taps into different dimensions relating to both political support as well as feelings of political (dis)empowerment (Christensen 2016), the current study includes three specific attitudinal indicators of political discontent: trust, political hopelessness and populist attitudes. Each of these indicators have long been recognized as different, though central dimensions of political discontent (Craig 1980; Craig and Maggiotto 1981). Moreover, empirical research tends to employ these concepts as operationalizations of political discontent, but often in separate ways (e.g., Krouwel and Abts 2007; Rooduijn et al. 2016). By incorporating these three indicators together in our study design, we explicitly build on recent work by Geurkink et al. (2020). The authors note that these three attitudes are commonly associated with behavioural indicators of discontent, but they also found that they are in fact conceptually different and measure empirically distinct phenomena. That is, while each of the concepts are related and tap partly into some form of anti-elitism, they still represent different dimensions of political discontent.

Political and media trust

Strongly related to the concept of political support is political trust. While its conceptualization remains complex, political trust concerns a personal evaluation of the political world in terms of the anticipated quality of government outputs (Craig 1979; Newton 2001). More specifically, Norris (2017) defines political trust as a “general belief in the performance capacity of political institutions and/or belief in the benevolent motivations and performance of office holders”.

The relationship between political trust and political participation is strongly debated in the literature. Broadly, two competing claims exist regarding this relationship (Levi and Stoker 2000). The first proposes a political exit from citizens in the case of political distrust: here, political trust is understood as a necessary precondition for any mode of political participation (Almond and Verba 1989; Verba et al. 1995). A second viewpoint, however, assumes that a lack of political trust pushes citizens towards alternative, non-institutionalized modes of political action. Thus, low levels of trust do not necessarily translate into low overall participation, but rather create a shift towards alternative voicing options, located outside the sphere of institutional politics (Craig 1980; Craig and Maggiotto 1981; Gamson 1971; Kaase 1999; Norris 1999).

While the literature on the matter is far from settled, current empirical research seems to favor the second viewpoint. For instance, trust has been found to boost institutional participation (and voting in particular) in European countries, while reducing non-institutionalized participation (Hooghe et al. 2011; Hooghe and Marien 2013). In a recent study, Ellison et al. (2020) confirmed this relationship in a youth sample. Simultaneously however, several studies failed to identify any relation between non-electoral participation and political trust, or report inconsistent results (Ehsan 2018; Zhang and Lin 2018). Furthermore, van Stekelenburg and Klandermans (2018) found that both trusting and distrusting citizens engaged in protest behaviour.

The current study considers social media participation to be an important form of non-institutional political participation. Based on the literature regarding alternative voicing routes in the case of discontent, low levels of political trust would likely be associated with engagement in this particular mode of action. Again, however, results are mixed. In a local community context for instance, political trust has been found to stimulate online participation, whereas it was negatively associated with offline action (Kwon et al. 2020). On the contrary, according to a study by Theocharis and de Moor (2021), those who engage in digitally networked participation through social media are generally younger, ‘critical citizens’ who are more distrustful of institutions of representative democracy. In a similar vein, a recent study found how low political trust proved to be a crucial stimulating factor for online political participation if citizens found conventional participation through voting unattractive (Koivula et al. 2021).

Bearing in mind these mixed empirical findings in the literature, the following hypotheses are formulated:

H3

Political trust will be positively associated with institutional political participation.

H4

Political trust will be negatively associated with protest participation (H4a) and social media participation (H4b).

Apart from trust in the political system, this study also considers diminishing trust in news media as a relevant indicator of discontent. As is the case with political trust, media trust involves the relationship between a trustee (i.e., the citizen) and a trustor (i.e., news media), and entails a certain risk, as our knowledge and understanding of the world largely depends on the information conveyed by such media (Prochazka and Schweiger 2019). While trust in news media can be conceptualized and studied on several levels (e.g., media content, type and brand) (Strömbäck et al. 2020), research has also shown that it is possible to discern a generalized (dis)trusting attitude towards news media as an institution (Prochazka and Schweiger 2019). Here, people are assumed to respond to ‘the media’ as a collective entity representing legacy journalistic mainstream media that are part of corporations and have large audiences (Tsfati and Cappella 2003, 2005).

Given our interest in the role of diminishing trust in explaining political participation, it is important to point to news media as a key institution in democracies. In this regard, Bennett and Livingston (2018) identify a breakdown of trust in democratic institutions that occurs parallel across press and politics. Trust in media is not isolated from perceptions of other public and democratic bodies but is tied to a general disenchantment with and disdain for social and especially, political institutions. This interrelatedness is what Hanitzsch et al. (2018) label the ‘trust nexus’ and, according to the authors, can be explained by growing feelings of anti-elitism and an assumed gap between the elite (i.e., media corporations, journalists, politicians) versus the people. Whereas, empirical research has found evidence for the existence of this nexus (Ariely 2015; Knudsen et al. 2021), there has been rather limited attention to the ways low trust in news media might affect citizens’ choices regarding their political participation (Zimmermann and Kohring 2020). Still, building on grievance theory and the literature on the trust nexus, we expect the following:

H5

Trust in news media will be positively associated with institutional political participation.

H6

Trust in news media will be negatively associated with protest participation (H6a) and social media participation (H6b).

Political hopelessness

This study also considers feelings of political hopelessness, which has been defined as the subjective feeling of lacking external political efficacy (EPE; Spruyt et al. 2016; Keppens et al. 2016). Whereas a sense of EPE includes the belief that the government is responsive to citizen demands (Niemi et al. 1991), political hopelessness refers to its inverse as a subjective experienced vulnerability (Spruyt et al. 2016). In other words, within this manuscript, we refer to political hopelessness as the perceived lack of EPE and hence, the belief that one has little influence on political processes and specific frustrations about the inability to have your voice heard by political elites (Geurkink et al. 2020).

When citizens do not believe their government to be responsive to their expressed demands, this affects the perceived utility of taking action and hence, participation (Bandura 1982; van Stekelenburg and Klandermans 2013). However, the role of EPE in explaining participation seems to differ depending on the political act considered. As EPE is conceptualized in direct relation to particular political institutions (i.e., the government, parties and politicians), its presence is considered mainly important in explaining forms of political action facilitated by or targeted towards these institutions (e.g., writing letters to civil servants or politicians). In other words, young citizens who feel sceptical about the willingness of the state to bring about change, are likely to abstain from these particular acts (De Moor 2016).

H7

Political hopelessness will be negatively associated with institutional political participation.

For non-institutional modes of participation, an opposite relationship can be assumed (Craig and Maggiotto 1982). These acts aim to bring about social or political change in an indirect manner, by keeping distance from the political system or by circumventing it altogether (Marien et al. 2010). However, while these modes of action, such as protesting, are not necessarily located within the sphere of the state, some authors argue how they often still target or pressure institutional actors to act (van Deth 2014). Prior inconsistent results in this regard might be due to this often unclear relationship between the need for responsiveness and offline non-institutional action (Chan 2016; Diemer and Rapa 2016). Still, drawing from grievance theory and assumptions on democratic discontent, we expect feelings of irresponsiveness to drive young citizens towards offline alternative voicing opportunities.

This reasoning can be extended when considering political participation occurring on social media. Recent forms of creative political participation, such as consumerism or digitally networked participation, do not necessarily count on the state to facilitate social change. Instead, they are politically motivated and expressive in nature (van Deth 2014). As social media in particular offer an alternative space where such individualized activities can emerge (Theocharis 2015), this is likely to impact the relevance of EPE (De Moor 2016). In this sense, social media participation can be considered an additional alternative voicing route apart from offline protest, driven by a perceived lack of institutional responsiveness.

While empirical evidence in this regard is scarce (Yang and DeHart 2016), we hypothesize the following:

H8

Political hopelessness will be positively associated with protest participation (H8a) and social media participation (H8b).

Populist attitudes

Populism has been defined as “an ideology that considers society to be ultimately separated into two homogeneous and antagonistic groups, ‘the pure people’ versus ‘the corrupt elite’, and which argues that politics should be an expression of the volonté générale (general will) of the people” (Mudde 2004). Related to other concepts such as trust and efficacy, populism constitutes a distinct ideology which is often assumed to both pose a democratic threat and function as a democratic corrective, fuelled by feelings of dissatisfaction (Geurkink et al. 2020; Mudde and Kaltwasser 2017; Zaslove et al. 2021).

Only recently have scholars begin to explore how populist attitudes among citizens influence the ways they participate in politics (Akkerman et al. 2014). It has been argued that populism moves beyond political apathy, and even entails a possibility for political mobilization (Keppens et al. 2016; Pirro and Portos 2021). Still, it remains unclear how the element of discontent, which is assumed to push citizens away from certain modes of political action, interacts with the complex mechanisms of populism (Spruyt et al. 2016).

The few available studies in this context present inconsistent results. Zaslove et al. (2021) found no relation between voting intentions and populism, while citizens with stronger populist attitudes were found to engage less in protest. Keppens et al. (2016) report almost opposite findings: populism increased voting intentions, but was not associated with alternative modes of participation. Again a different body of research presents findings in line with grievance theory, where populist attitudes (as an indicator of discontent) seem to drive non-electoral and -institutional modes of engagement (Anduiza et al. 2019; Pirro and Portos 2021). Online modes of action in particular have been found to attract populist citizens, as they avoid intermediation from institutional elites, providing a low-cost opportunity to connect with like-minded communities (Anduiza et al. 2019; Boulianne et al. 2020). Following these arguments, we expect the following:

H9

Populist attitudes will be negatively associated with institutional participation.

H10

Populist attitudes will be positively associated with protest participation (H10a) and social media participation (H10b).

Methods

Procedure and sample

The data were collected through an online survey in Qualtrics among Belgian young adults. It must be noted that Belgium is one of the countries with the highest turnout rate of all established democracies, which is attributed to its compulsory voting system that requires citizens to vote for a party, candidate or cast a blank vote (Caluwaerts and Reuchamps 2022). Given the purpose of this study, this context should be considered.

Eligibility criteria for participation were (a) being between 18 and 30 years old, (b) speaking Dutch and (c) having the Belgian nationality, because of the questions relating to voting behaviour. Both students, working and non-active respondents were eligible to participate. In total, 907 respondents participated in the survey. After omitting incomplete surveys, a final sample of 720 respondents was obtained to be used for subsequent analyses. Table 1 provides a description of the final study sample.

Table 1.

Characteristics of the study sample

	Study sample (n = 720)
Sex (n/%)
Male	293 (40.7%)
Female	427 (59.3%)
Age in years (M/SD)	23.99 (3.15)
Activity
Higher education student	411 (57.1%)
Working (employee/self-employed)	284 (39.4%)
Non-active	25 (3.5%)
Education (n/%)
Primary or lower secondary education diploma	9 (1.3%)
Secondary education diploma	107 (14.9%)
Higher education diploma (bachelor degree)	338 (46.9%)
Higher education diploma (master degree)	257 (35.7%)
Higher education diploma (PhD)	9 (1.3%)
Disposable income in euros (M/SD)	1935.41 (724.39)

Disposable income was measured using the OECD equivalence scale (Hagenaars et al. 1994) and only measured among non-students. 15 People chose not to disclose their income. As such, sample for disposable income: n = 294

Respondents were recruited using a two-stage strategy with assistance of a university student. During this first stage, which ran from March 11th until April 2nd 2021, the survey was distributed through various social media channels, such as Facebook, Twitter, Instagram and TikTok. All posts included a short statement on the purpose of the study, an invitation to participate and a link to the survey. Recruitment was done via the networks of the student involved, as well as by relying on social media influencers. These influencers1 were contacted because of their reach with our target respondents and because they had been outspoken on social and political issues on their profiles.2 In addition, in this first stage, the researchers relied on several political opinion leaders, such as members of city councils, political party members and members of local (youth) organizations, who were all contacted through mail or social media. Employing a snowball approach, these political opinion leaders were asked to spread the questionnaire among their network and followers. Lastly, the survey was shared within various online groups (such as community Facebook groups) as to reach more diverse subgroups within the age category.

The second recruitment stage started on March 25th during which a new pool of respondents was recruited. First, the survey was shared on the online and social media channels of the research institution, as well as through the networks of the principal investigators (Facebook, Instagram, Twitter and LinkedIn). In addition, the survey was spread among students of Ghent University and the University of Antwerp, through the channels of several courses of the principal investigators. Again, several political opinion leaders were contacted through mail and social media and the survey was shared by different representatives of Flemish youth parties,3 student councils as well as additional youth organizations. Lastly, because of a possible underrepresentation of male respondents following the first recruitment stage, Facebook ads were used to promote an invitation to participate in the study. A Facebook page dedicated to the study was created and a recruitment message was posted on the page, accompanied by an image and the survey link. This message was then promoted to two different audiences: the first was targeted towards Belgian citizens between 18 and 30 years old and the second towards graduated Belgian males within the same age category. On April 12th, the survey was closed.

Respondents who agreed to participate in the study, were redirected to an online Qualtrics survey environment. Prior to filling out the questionnaire, respondents were provided information on the purpose of the study as well as contact details of the researchers. The participants were assured that their responses would remain confidential, that their participation was voluntary and that they were free to withdraw their participation at any time. Furthermore, each respondent was asked for their informed consent by opt-in. The study received a positive advice from the Ethical Committee of Ghent University.

Measures

“Appendix” contains a detailed overview of the survey items for the attitudes of discontent and the control variables, as well as the range, mean and standard deviation per item. An overview of the participation and electoral exit variables, their prevalence and descriptives can be found in Table 2.

Table 2.

Descriptives and prevalence of participation and electoral exit

Participation items	Prevalence (% of sample that engaged in behaviour)	Mean	SD
Institutional participation
I worked for a party or candidate	28.5	1.57	1.09
I contacted a politician or public servant to draw attention to a social or political issue	27.8	1.49	0.93
I attended a (digital) meeting organized by a politician, party or governing body	24.9	1.54	1.09
I donated money to a political party or politician	7.9	1.15	0.56
Protest participation
I worked for a political action group (e.g., Youth for Climate)	34	1.62	1.03
I signed a petition	82.8	2.71	1.11
I participated in a demonstration, strike, blockade or occupation	28.6	1.50	0.91
I contacted a media outlet to draw attention to a social or political issue	18.2	1.28	0.69
Social media participation
I posted or shared something (status, meme, link, …) on a social or political issue on social media	72.2	2.67	1.36
I commented on something on a social or political issue on social media	63.9	2.32	1.29
I changed my profile picture on social media to express myself on a social or political issue (e.g., by using Facebook Frames)	33.1	1.59	0.99
I created a group, page or event in the context of a social or political issue on social media	8.2	1.15	0.57
I indicated that I would attend an event in the context of a social or political issue on social media	52.5	1.99	1.16
Electoral exit behaviour (past elections)
Did not vote	1	–	–
Cast blank vote	1.7	–	–
Purposely cast invalid vote	0.6	–	–

Dependent variables

Drawing from the taxonomy of Theocharis and van Deth (2018b), three different types of political participation were included in the study: institutional participation, protest participation and social media participation. Institutional participation was measured using four items (Cronbach’s α = 0.83; M = 1.44; SD = 0.76). An example item is ‘I contacted a politician or public servant to address a social or political issue’. For protest participation, respondents rated four items (Cronbach’s α = 0.74; M = 1.78; SD = 0.71). An example item is ‘I worked for a political action group (e.g., Youth for Climate, Sound of Silence,…)’. Finally, social media participation was measured using five items (Cronbach’s α = 0.79; M = 1.94; SD = 0.82). To measure this construct, a shortened and adapted version of the Social Media Political Participation Scale (Waeterloos et al. 2021a) was employed. An example item is ‘I posted or shared something related to a social or political issue on social media’. The respondents were asked how often in the past 12 months they had engaged in any of these activities [never (1)–very often (5)]. The three theoretical constructs were confirmed by conducting an exploratory factor analysis (EFA) and confirmatory factor analysis (CFA).

To measure past electoral exit behaviour, respondents were asked whether, during the last federal elections, they had done any of the following: ‘Did not vote, while eligible to do so’, ‘Purposely cast a blank vote’ or ‘Purposely cast an invalid vote’. Answers were coded in a dichotomous manner (0 = no; 1 = yes). Afterwards, responses were recoded to create a new dichotomous variable where a value of ‘1’ indicated that respondents had engaged in any of the three exit behaviours. In the questionnaire, we also controlled for the fact that some of the respondents might have been underage during the last federal election and that they did not vote due to this particular reason. This was done by including the following item in our survey: “I did not vote because I was not called to vote (I was still underage)” (0 = no; 1 = yes; 100 respondents indicated ‘yes’). Each of these items were mutually exclusive if answered positively.

Independent variables: attitudes of discontent

Two different measures for political trust were included in the questionnaire. The first measured trust in political institutions, while the second referred to trust in the political outcomes produced by these institutions (Geurkink et al. 2020). Trust in institutions was measured using four items, asking respondents how much trust they have in (a) the government, (b) political parties, (c) the parliament, and (d) politicians (Geurkink et al. 2020; Hooghe and Dassonneville 2018). Responses were rated on a scale from 1 to 10 and averaged into a sum scale (Cronbach’s α = 0.81; M = 4.78; SD = 1.72). Trust in outcomes was measured using the political trust scale as employed by Zimmermann and Kohring (2020). Respondents rated four statements using a five-point Likert scale [disagree (1)–agree (5), e.g., ‘Usually, parties and politicians search for appropriate solutions for social problems’] which were averaged into a sum scale (Cronbach’s α = 0.81; M = 2.58; SD = 0.79).

Furthermore, trust in news media was measured using a single statement (‘I think you can trust most news most of the time’) that was rated on a five-point Likert scale [disagree (1)–agree (5); M = 3.16; SD = 1.13]. While we recognize the different potential levels of analysis in the context of trust in news media, a single-item measure was deemed most appropriate to capture generalized trust in news media as a democratic institute as an indicator of discontent (Fletcher and Park 2017; Strömbäck et al. 2020).

To measure political hopelessness, six items were adapted from prior research (e.g., ‘Political parties are only interested in my vote, not my opinion’) (Craig et al. 1990; Geurkink et al. 2020; Keppens et al. 2016; Spruyt et al. 2016). Specifically, we aimed to create a more sophisticated and comprehensive measure of political hopelessness by combining reversed items of the established ‘external political efficacy’ scale by Craig et al. (1990) and the measures of Spruyt et al. (2016). Respondents were asked to evaluate the statements on a five-point Likert scale [disagree (1)–agree (5)]. In addition, an EFA confirmed the one-factor structure of the items. As such, the items were averaged into a sum scale (Cronbach’s α = 0.80; M = 2.97; SD = 0.79).

The Populist Scale by Akkerman et al. (2014) was included to assess populist attitudes among the respondents. Six items were rated on a five-point Likert scale [disagree (1)–agree (5)], and averaged into a sum scale (Cronbach’s α = 0.76; M = 3.28; SD = 0.73). An example item is ‘Elected politicians need to follow the will of the people’.

Socio-demographic variables and controls

Apart from socio-demographic variables (age and sex), additional covariates were included in the analysis based on their possible confounding effects. A subjective measure of financial stress was used to capture feelings of financial need (Ponnet 2014). Three items were rated on a five-point Likert scale [disagree (1)–agree (5)]. An example item is ‘With our current income, it is difficult to make ends meet’. The items were averaged in a sum scale (Cronbach’s α = 0.78; M = 1.75; SD = 0.84). Overall social media use was measured by asking respondents how often, on an average day, they use social media [never (1)–multiple times a day (7); M = 6.86; SD = 0.62]. To measure political ideology, respondents were asked the following: “Most people use ‘left’ and ‘right’ to distinguish between various political attitudes. Where would you place yourself on the political spectrum ranging from ‘entirely left oriented’ (1) to ‘entirely right oriented’ (11)?” (M = 5.15; SD = 2.54). Furthermore, respondents were asked about their political interest (‘Overall, I am interested in political and social issues’), which was rated on a five-point Likert scale [disagree (1)–agree (5); M = 3.99; SD = 1.07]. Internal political efficacy was measured using three items on a five-point Likert scale ranging from disagree (1) to agree (5) (e.g., ‘I consider myself to be well qualified to participate in politics’) and averaged into a sum scale (Cronbach’s α = 0.83; M = 3.57; SD = 0.96) (Niemi et al. 1991). Lastly, respondents were asked about their overall satisfaction with the functioning of the Belgian democracy [not at all satisfied (1)–completely satisfied (10), M = 5.14; SD = 2.09, Portos et al. 2019].

Analytical strategy

All analyses were conducted in SPSS 26. Incomplete surveys were omitted from the dataset, as well as respondents who gave a wrong answer to an attention check question in the survey. As a first step in our analyses, we assessed the correlations between our study variables (i.e., the attitudes of discontent and the participation variables). The main analyses consisted of a series of hierarchical regression analyses aimed at assessing which indicators of political discontent are significantly associated with the different participation variables. Specifically, three blocks of variables were included in the regression: the first consisted of socio-demographic variables, the second included controls and the third included our variables of political discontent. Prior to conducting the regressions, we ensured that assumptions regarding normality, linearity, multicollinearity and homoscedasticity were met.

Results

Table 3 presents the bivariate correlations among the study variables. Each of the attitudes of discontent showed to be moderately to highly correlated to each other, albeit in different ways. The only exception was the association between populist attitudes and trust in outcomes. Past electoral exit behaviour was only significantly and positively correlated to political hopelessness. Furthermore, the different attitudes of discontent were not consistently correlated to the different participation variables.

Table 3.

Correlations among study variables

Variables		1	2	3	4	5	6	7	8	9
1	Past electoral exit		− 0.02	− 0.05	− 0.05	0.13**	0.01	− 0.04	− 0.06	− 0.01
2	Political trust (outcomes)			0.66**	0.41**	− 0.55**	− 0.47	− 0.13**	0.08*	− 0.10**
3	Political trust (institutions)				0.46**	− 0.56**	− 0.44**	− 0.01	0.10**	− 0.03
4	Trust in news media					− 0.38**	− 0.28**	0.003	− 0.02	− 0.03
5	Political hopelessness						0.55**	− 0.07	− 0.23**	− 0.12**
6	Populist attitudes							0.10**	− 0.06	0.05
7	Protest								0.46**	0.69**
8	Institutional									0.55**
9	SNS participation

^*p < 0.05; **p < 0.01

Table 4 presents the results from the regression analyses, respectively, with institutional, protest and social media participation as dependent variables. The table presents the final models from the hierarchical regression analysis, with all variable blocks included and with standardized β coefficients reported.

Table 4.

Regression analysis

	Institutional			Protest			Social media participation
	Standard β	t value	p value	Standard β	t value	p value	Standard β	t value	p value
Constant		2.02*	0.05		3.97***	0.00		1.52	.13
Block 1: Demographics
Age	0.01	0.21	0.83	− 0.11	− 3.31**	0.001	− 0.06	− 1.88	0.06
Gender (ref = female)	− 0.04	− 1.07	0.29	0.13	3.61***	0.00	0.12	3.14**	0.002
Financial stress	0.12	3.55***	0.00	0.07	2.12*	0.03	0.12	3.56***	0.00
Block 2: Controls
Political interest	0.18	3.93***	0.00	0.17	3.75***	0.00	0.25	5.59***	0.00
Social media use	0.01	0.34	0.74	0.03	0.93	0.35	0.12	3.52***	0.00
Internal political efficacy	0.16	3.66***	0.00	0.16	3.56***	0.00	0.17	3.78***	0.00
Satisfaction with democracy	− 0.24	− 5.21***	0.00	− 0.08	− 1.67	0.10	− 0.08	− 1.62	0.11
Political ideology	0.03	0.76	0.45	− 0.28	− 7.27***	0.00	− 0.13	− 3.44**	0.001
Block 3: Indicators of discontent
Past electoral exit	− 0.02	− 0.49	0.62	− 0.01	− 0.24	0.81	0.03	0.91	0.36
Political trust (outcomes)	0.02	0.39	0.70	− 0.14	− 2.86**	0.004	− 0.11	− 2.36*	0.02
Political trust (institutions)	0.15	2.99**	0.003	0.06	1.20	0.23	− 0.01	− 0.27	0.79
Trust in news media	− 0.08	− 1.88	0.06	− 0.08	− 2.08*	0.04	− 0.10	− 2.44*	0.02
Political hopelessness	− 0.29	− 6.23***	0.00	− 0.16	− 3.42**	0.001	− 0.23	− 4.94***	0.00
Populist attitudes	0.09	2.16*	0.03	0.13	3.11**	0.002	0.09	2.11*	0.04
Adjusted R2 (model 1: only block 1 variables)	0.03	0.04	0.01
Adjusted R2 (model 2: block 1 + block 2 variables)	0.15	0.20	0.20
Adjusted R2 (final model)	0.22	0.23	0.23
N (respondents)	720	720	720

^*p < 0.05; **p < 0.01; ***p < 0.001

Regarding socio-demographics, age was negatively associated with protest participation, whereas no age differences were found for the other types of participation, which was to be expected given the study’s focus on young adults. The regression did reveal gender differences, with women being significantly more likely to engage in both protest and social media participation. This corresponds with prior findings on the socio-demographic profile of citizens engaging in non-institutional participation (Marien et al. 2010; Theocharis et al. 2019). Furthermore, financial stress was positively associated with all of the participation types. In other words, young adults with a higher subjective feeling of financial need, were more likely to participate politically, both through institutionalized means, protest and social media. To some extent, this contradicts established assumptions on the role of economic grievances and political participation. Whereas feelings of economic hardship are generally assumed to drive citizens towards non-institutional modes of participation, our study shows how financial stress positively affected all participation types among young citizens, both institutional, protest and through social media (Gurr 1970; Kern et al. 2015; Kurer et al. 2019).

Our second block included several control variables. As expected, overall social media use was significantly and positively associated with social media participation. Furthermore, political interest and internal political efficacy were both positively associated with engagement in all participation types. Rather unexpectedly, satisfaction with democracy was a negative predictor of institutional participation. Finally, political ideology was significantly associated with both protest and social media participation. In both cases, more left leaning respondents were more likely to engage in these forms of political participation, which again confirms prior studies (Theocharis et al. 2019).

Moving to the study’s main variables of interest, past electoral exit behaviour was not significantly related to any of the participation variables. Consequently, our first two hypotheses are rejected. Our second set of hypotheses assumed that political trust would be positively associated with institutional participation (H3), but negatively with protest (H4a) and social media participation (H4b). Interestingly, the two included forms of political trust showed to be related to different modes of participation. First, political trust in institutions was a positive predictor of institutional participation. This is in line with H3. In contrast, trust in political outcomes was a negative predictor of both protest and social media participation, which confirms our fourth hypothesis. A similar conclusion can be drawn when looking at trust in news media, as this type of trust was again a negative predictor of protest and social media participation, hereby confirming H6.

Furthermore, political hopelessness was a negative predictor of all three participation types. This suggests that a perceived lack of political responsiveness and a sense of political disillusionment diminishes participation in democratic society, both through institutional as well as non-institutional means. While these findings confirm H7, they contradict our expectations regarding the relationship between hopelessness and non-institutional modes of participation (H8), which we assumed to be positive. As a final predictor in the model, populist attitudes were positively associated with both institutional, protest and social media participation, and were an especially important predictor in the case of protest participation. Again, these results are not fully in line with our expectations. As we expected a negative relation between populism and institutional participation, H9 could not be confirmed. On the other hand, H10 expected populism to mobilize young citizens into protest and social media participation, and could therefore be accepted.

Discussion

This study sought to assess how young citizens in Belgium are engaging in various modes of political participation and how these acts are associated with both behavioural as well as attitudinal indicators of political discontent. In particular, the study drew from grievance theory, while also building on the growing literature acknowledging new, creative forms of political participation as alternative indicators of democratic health. Youth in particular increasingly rely on social media as a political outlet, to express their voice, mobilize their peers or pressure political actors. To date, however, little research has explored how political participation through social networking sites might provide an alternative voicing route for young citizens who feel dissatisfied with the political status-quo.

Our first research question asked how past electoral exit behaviour was linked to three non-electoral participation types. Specifically, we assumed that young citizens who engaged in electoral exit behaviour would be less likely to participate through institutional means, and engage more in non-institutionalized modes of action such as protest and social media participation (Portos et al. 2019). Contrary to our expectation however, no significant associations could be identified. As such, our findings from a youth population do not lend support for the understanding that some citizens withdraw from voting as a rejection of institutional representation, while still engaging in non-institutional acts (Quaranta 2018). Given the absence of significant effects, our results do not support alternative frameworks either. For instance, Stockemer (2014) argues how voting might function as a stepping stone towards non-institutional modes of participation, while others have proposed an opposite direction, where participation through social media specifically seems to increase the likelihood to vote among citizens who were initially least likely to do so (Steinberg 2015). Nonetheless, it is possible that the absence of effects here is due to the particular compulsory voting system in Belgium, which likely explains the very low frequency of electoral exit behaviours in our study (Hooghe et al. 2011). Moreover, it might be that our measure of electoral exit behaviour did not properly capture the complexity of expressions of discontent in this voting context. That is, while purposely blank or invalid voting is often used as a way to channel discontent, abstaining in a compulsory voting system might also be due to certain constraints or costs (Katz and Levin 2018). While we filtered out respondents who indicated they did not vote because of work-related or practical reasons during the data cleaning process, we still cannot fully assume that the abstaining respondents did so out of discontent. In any case, more research is necessary to identify how voting behaviour is combined with other modes of action (including online) in individuals’ political repertoires, especially in compulsory voting contexts.

Our second research question asked how various attitudinal indicators of political discontent were associated with institutional, protest and social media participation. The main assumption tested in our study was that these attitudes drive young citizens away from institutionalized participation and towards non-institutional action. Our results however are mixed, and only partly support the main assumptions as proposed by grievance theory and the mobilizing features of discontent. First, our findings shed light on the complex ways in which political trust (which, if absent, is a crucial indicator of discontent) mobilizes young citizens into diverse forms of political action. Our study included two different measures of political trust, as a way to disentangle prior inconsistent findings in this regard. Our results show how trust in political institutions, as one aspect of political trust (Zmerli et al. 2011), is positively associated with institutional participation. That is, young adults who expressed a higher overall trust in the government, political parties, politicians and the parliament, were more likely to participate through institutional means. On the other hand, respondents who have little faith that political bodies can produce and achieve certain desired outcomes (e.g., take and implement the right decisions, search for solutions to social and political problems), are more prone to engage in protest and social media participation. As such, it seems non-institutional modes of participation are driven in particular by the evaluation of political processes, rather than the governing bodies (van der Meer and Hakhverdian 2016). In a similar vein, our results show how distrust in news media paralleled distrust in political outcomes in terms of its mobilizing features. That is, young citizens who were less trusting of news media, participated more often in both social media and protest participation. These findings therefore confirm theorizing about the so-called ‘trust nexus’ between politics and press as crucial democratic institutions (Hanitzsch et al. 2018). It would be interesting to further explore the particular nature of this trust nexus and its implications for democratic participation. In this sense, future work might want to assess whether different levels of media trust are in fact related to each other and how they influence particular forms of political participation (Strömbäck et al. 2020). Overall, our findings show how social media provide an additional outlet for young, distrusting citizens to take matters in their own hands to achieve certain desired outcomes, outside the sphere of the state and apart from offline protest (Theocharis and de Moor 2021).

As a second indicator of political discontent, we expected feelings of political hopelessness to drive young citizens towards protest and social media participation. Our findings, however, could not confirm our adopted framework on political discontent, as feelings of hopelessness seem to discourage young adults to take any type of political action, even through social media. This suggests that, contrary to the arguments of de Moor (2016), perceptions about institutional responsiveness to citizen demands remain relevant, even when modes of action are seemingly not directly targeted at or facilitated by the state. Moreover, the finding that social media political participation is also deterred by perceptions of irresponsiveness goes against claims about the motivations of citizens engaging in these forms of action. Critical voices have described political participation occurring on social media as forms of ‘slacktivism’: easy, low-threshold behaviours that are done for personal and social gain and to construct a certain online image, instead of resembling a true political commitment or being done to reach political goals (Christensen 2011; Morozov 2009, 2011). However, our results point to the opposite. Young citizens who do not expect the government to be responsive to their demands and perceive them unwilling to bring about change, will also be deterred from taking action on social media. This seems to imply that citizens employ these tools for political reasons and expecting some governmental response.

A possible explanation for these findings is the existence of a mismatch between participatory preferences of political elites on the one hand and those of younger citizens on the other. It might be the case that feelings of political hopelessness are deterring youth from participating in new, innovative acts (such as through social media) because they believe that political elites are not valuing these participatory modes (Hooghe and Marien 2014; Matthews 2020). Consequently, it seems necessary for political elites to establish signals of political responsiveness, as a way to assure a healthy and diverse participatory democracy. Because social movements, and youth in particular, are increasingly raising issues through social media platforms, acknowledging these repertoires as valuable could be a way to decrease feelings of political hopelessness.

Finally, regarding the role of populist attitudes in stimulating the pursuit of alternative voicing routes, our study again presents mixed findings. We found that populist attitudes stimulated both protest and social media participation, confirming our hypotheses as well as some previous research findings (Anduiza et al. 2019; Pirro and Portos 2021). While these findings go against the reasoning set out by Zaslove et al. (2021), we would argue that the unmediated, direct and individualized nature of non-institutional political participation fits the people-centric focus of populism. Contrary to our expectations, and grievance theory, populist attitudes were also positively associated with institutional participation. In other words, populism among young citizens does not stimulate institutional exit, but rather enhances overall political participation through various means. This corresponds with the argument that populists are not anti-democratic, but that they are critical towards the current state of it. As such, populists seek democratic reform out of dissatisfaction with the current form of representation, which is driven by ‘the corrupt elite’ rather than ‘the pure people’ (Mudde 2004; Zaslove et al. 2021). Our results indicate that this democratic correction is pursued both through institutional means, as well as protest and social media participation. Still, we recommend future research to investigate the relation between populism and diverse modes of political action further, given the current contradicting empirical studies in the field.

Limitations and conclusion

Some limitations of this study have to be considered. The most important limitation of this study concerns our sampling strategy. As we employed a convenience sampling method, we must acknowledge an overrepresentation of women and respondents with a higher education diploma in our sample. Comparing our sample with representative data for the Flemish population, it is likely that we have almost twice as many respondents with a higher education diploma compared to the young adult population (Statbel 2022). This is possibly due to self-selection bias, which may have caused already politically interested or engaged young citizens to participate. This overrepresentation likely impacted our study results, as feelings of discontent are often associated with socio-economic status and education levels (Bertsou 2019; Spruyt et al. 2016). This, together with the context of a compulsory voting system in Belgium, likely explains the low levels of electoral exit behaviour. Because our findings cannot be generalized to the general Flemish population, we strongly encourage future research to replicate our study in a representative sample.

Still, we aimed to make our study as accessible and inclusive as possible in terms of language and distributed our survey through various actors, contexts and recruiting channels as a way to cope with this possible bias. Moreover, Theocharis et al. (2019) state that a sample bias towards politically interested citizens can be useful in the study of political participation, as this gives researchers access to a part of the population that may exhibit sufficiently diverse participation patterns to explore new participatory modes (for instance, on social media) and their underlying, unexplored associations. We would therefore argue that, despite the limited representativeness of our study sample, important first steps were conducted in exploring the relations between social media political participation, offline participatory modes and attitudinal indicators of discontent.

Second, as we rely on cross-sectional survey data, we cannot make any causal claims regarding our variables of interest. For instance, it is possible that participation through institutional means enhances populist attitudes, as citizens experience certain democratic processes first-hand that might not correspond with their expectations, leaving them dissatisfied with this status quo. Similarly, it is also likely that young people engaging in political participation on social media also see their populist attitudes reinforced, due to potential echo chamber effects and increased interaction with like-minded individuals in online communities (Boulianne et al. 2020; Sunstein 2007).

Fourth, we would like to make a suggestion for future work to employ a platform-specific perspective within this research area. Our study focussed on generalized social media political participation and did not explicitly distinguish between social media platforms. As there is evidence that the affordances of different social media platforms facilitate different participatory behaviours, it would be valuable to further disentangle the association between discontent and participation by focussing on specific platforms such as TikTok or Twitter (Kim and Lee 2021; Valenzuela et al. 2018).

Lastly, our study was conducted during the COVID-19 pandemic in Belgium, after the second infection wave and lockdown and during the start of the vaccination campaign. During crises, protest participation is likely to increase (Grasso and Giugni 2016) and social media have played an important role in citizen participation during the pandemic (Waeterloos, De Meulenaere, et al. 2021). This crisis context has to be considered when interpreting our findings.

In conclusion, our study provides mixed support for the assumptions put forward by grievance theory. We would argue that these findings can be situated within a growing body of research addressing political participation not as a choice between exclusive, independent acts but rather as a continuum within particular repertoires of action. Our study shows that, while the ‘exit/alternative voicing’ framework applies to our sample to some extent, indicators of political discontent do not uniformly push young citizens away from institutional politics in favour of alternative ones. Indeed, in recent studies, state-oriented forms of action have been found to coexist with non-institutional participation (de Moor et al. 2017). This might explain why our findings are not in line with grievance theory, which relies on the idea of political participation as exclusive acts, instead of complementary ones. Nonetheless, our findings address an important gap in the literature, by shedding light on previously unexplored relations between established indicators of political discontent on the one hand (i.e., trust, hopelessness and populism), and innovative forms of participation through social media on the other. In short, our study points to the importance of including social media participation as a proper mode of non-institutional political participation, as it is driven by similar factors as offline protest participation and unexpectedly, even institutional forms of action. Indeed, the political implications of discontent among young adults seem to reach beyond the offline realm as these attitudes spur engagement in online spheres. This further advances the argument that, to make a correct assessment of democratic health, as well as its potential threatening factors, political participation through social media should be taken into account (Theocharis 2015).

Biographies

Cato Waeterloos

(PhD) is a Postdoctoral Researcher at imec-mict-Ghent University (Department of Communication Sciences). In her research, she investigates the role of digital media, and social media specifically, in citizens’ political and civic participation. Her interests lie with how and why citizens employ new technologies and social media as platforms for political participation. In the past, her research explored how these innovative forms of participation exist next to more established participatory acts (such as voting, protesting or signing petitions), and which factors influence the choices people make in the democratic landscape (such as news consumption and different cognitive processes).

Michel Walrave

(PhD) is a Professor at the Department of Communication Studies of the University of Antwerp and Head of the Research Group MIOS. His research is centred around online self-disclosure and privacy. He investigates individuals’ self-disclosure and privacy in interpersonal online communication. More particularly, his research focuses on self-disclosure’s motives, opportunities (e.g., sharenting) as well as risks (e.g., cyberbullying). He also investigates sensitive online disclosures (e.g., sexting). Next to interpersonal self-disclosures, he studies internet users’ entrusting of personal data to businesses and how consumers deal with interactive and personalized marketing.

Koen Ponnet

(PhD) is a Professor at the Department of Communication Sciences, imec-mict-Ghent University. He conducts research in the field of Health Psychology, Media Psychology and Social Psychology, and examines the determinants of online and offline health and risk behaviours of vulnerable people, like youngsters or those who are at risk of poverty. Ponnet holds an MA in Psychology (1999), MA in Industrial Entrepreneurship (2004) and a PhD in Psychology (2004). He also completed a 1-year Post-University Program in Business Administration (2003) and a 1-year Post-University Program in Project Management (2004).

Appendix: Survey constructs

Measure and items		Range	Mean	SD
Political trust in institutions
PTI1	The government	1–10	5.44	2.07
PTI2	Political parties	1–10	4.30	1.84
PTI3	The parliament	1–10	5.27	2.09
PTI4	Politicians	1–10	4.10	1.99
Political trust in outcomes
PTO1	The important issues are on top of the political agenda	1–5	2.67	1.01
PTO2	Usually, parties and politicians search for appropriate solutions for social problems	1–5	2.73	0.98
PTO3	In general, one can rely on politics to make the right decisions	1–5	2.38	1.03
PTO4	Most of the times, political decisions are implemented properly afterwards	1–5	2.54	0.95
Trust in news media
TNM	I think you can trust most news most of the time	1–5	3.16	1.13
Political hopelessness
PH1	There are so many people who vote at elections, that my vote will not make any difference	1–5	1.97	1.04
PH2	As soon as they are elected, politicians think they are better than people like me	1–5	3.10	1.11
PH3	Voting has no sense; the parties do what they want to do anyway	1–5	2.95	1.28
PH4	Politicians are not interested in what people like me think	1–5	2.94	1.09
PH5	Political parties are only interested in my vote, not in my opinion	1–5	3.53	1.02
PH6	People like me don’t have any say about what the government does	1–5	3.32	1.17
Populist attitudes
POP1	Elected politicians need to follow the will of the people	1–5	3.73	0.89
POP2	The people, and not politicians, should make our most important policy decisions	1–5	2.94	1.13
POP3	I would rather be represented by a citizen than by a specialized politician	1–5	2.92	1.18
POP4	The political differences between the elite and the people are larger than the differences among the people	1–5	3.31	1.16
POP5	Elected officials talk too much and take too little action	1–5	3.85	1.01
POP6	What people call ‘compromise’ in politics is really just selling out on one’s principles	1–5	2.91	1.11
Internal political efficacy
IPE1	I consider myself to be well qualified to participate in politics	1–5	3.58	1.16
IPE2	I feel that I have a pretty good understanding of the important political issues facing our country	1–5	3.73	1.02
IPE3	I think that I am better informed about politics and government than most people	1–5	3.41	1.17
Financial stress
FS1	It is difficult to afford much more than the basics with our current income	1–5	1.79	1.07
FS2	I feel that our current income allows me to maintain a desirable standard of living (reverse-coded)	1–5	4.07	1.08
FS3	With our current income, it is difficult to make ends meet	1–5	1.53	0.87

If an item was reverse coded in the analyses, we did not include the new descriptives in this table. Therefore, the descriptives displayed here are the descriptives for the original survey items as presented to the respondents

Funding

This work was supported by the Research Fund of Ghent University (Grants Code BOF.STG.2018.0002.01). The interpretation of the data, the writing of the article and the decision to submit the article were the sole responsibility of the authors and were not influenced by the funding institution.

Data availability

The data that support the findings of this study are available from the corresponding author, C.W., upon reasonable request.

Declarations

Conflict of interest

On behalf of all authors, the corresponding author states that there is no conflict of interest.