Introduction
Participation in elections is one of the most meaningful tools available to citizens of modern democracies. Yet, a single vote rarely makes a difference in any country, irrespective of the nature of its electoral system. Anthony Downs formulated what has become known as the ‘paradox of voting’ (Downs, Reference Downs1957) along the following lines: Despite their chances of influencing the final outcome being infinitesimally small, individuals still turn out to vote. This dilemma is addressed in a large and growing body of research (Blais, Reference Blais2006, Reference Blais, Dalton and Klingemann2007). Among the explanatory factors suggested are the role of external influences (e.g., social networks: Hopmann, Reference Hopmann2012; social pressures: Gerber et al., Reference Gerber, Green and Larimer2008; media content: Schuck et al., Reference Schuck, Vliegenthart and De Vreese2016); and internal motivations (e.g., interest in politics: Miller et al., Reference Miller, Shanks and Shapiro1996).
There remain, however, major gaps in our understanding of what drives turnout. Blais (Reference Blais, Dalton and Klingemann2007) argues in this regard that the challenge is to account for the reason voters think that they have an obligation to vote, even though their individual vote is unlikely to influence the outcome. This paper addresses the issue by integrating narrative theory as a new type of explanation for turnout. In line with the narrative policy framework, which calls for ‘a quantitative, structuralist, and positivist approach to the study of policy narrative’ (Jones & McBeth, Reference Jones and McBeth2010, p. 330; Jones, Shanahan & McBeth, Reference Jones, Shanahan and McBeth2014; see also Jones & Radaelli, Reference Jones and Radaelli2015), we suggest that story components can incentivize the likelihood of voting and empirically test this hypothesis. The rationale is that embracing a group’s story components will incentivize individuals to take part in that group’s political activities.
We therefore add to previous attempts to connect between perceptions of national identity and their social and political ramifications. Previous theoretical work points to the possible political benefits of national attachments, for instance, in encouraging social solidarity among various groups within the nation (e.g., Kymlicka, Reference Kymlicka2015; Tamir, Reference Tamir2019). Such social aspects of nationalism can offset current threats to democracies, notably the rise of populist movements or internal political cleavages (Gustavsson & Miller, Reference Gustavsson and Miller2020). The political and social advantages arising from national attachments are corroborated by a large body of empirical evidence, which shows that such sentiments can increase the involvement in the democratic process (Huddy & Khatib, Reference Huddy and Khatib2007), ease inter-group hostility (Levendusky, Reference Levendusky2018), reduce tax frauds (Chan, Reference Chan2019), and in some cases, encourage social and political trust (Gustavsson & Stendahl, Reference Gustavsson and Stendahl2020) and support for the welfare state (Johnston et al., Reference Johnston, Banting, Kymlicka and Soroka2010). However, while research hitherto has focused primarily on the strength of individuals’ attachment to their nation, we explore the subjective meanings that national affinity entails for individuals and the consequences of such notions for voter turnout.
This paper includes three main parts. First, we establish the idea of story incentive theoretically and propose an empirical method for its analysis. We argue that an account of voting turnout in terms of story incentive rests on the assumption that the likelihood to vote in national elections is higher amongst individuals who embrace story components of the nation. Second, based on election surveys from four parliamentary political systems (Denmark, Israel, the Netherlands, and the UK) and one presidential system (the US), we derive evidence for the relationship between story incentive and voter turnout, thereby validating an additional explanatory parameter above and beyond known explanations. In the concluding section, we discuss the theoretical implications of our findings.
Story and politics
Human beings are essentially storytelling creatures (MacIntyre, Reference MacIntyre1981). Narrating is a natural impulse, and the narrative mode of thought and expression is a fundamental dimension in human life (e.g., MacIntyre, Reference MacIntyre1981; Carrithers, Reference Carrithers and Whiten1991; McAdams, Reference McAdams1993; Shenhav, Reference Shenhav2015). A narrative is often taken to be ‘a primary means by which individuals organize, process, and convey information’ (Jones & McBeth, Reference Jones and McBeth2010: 330), created and used, inter alia, ‘to interpret and understand the political realities around us’ (Patterson & Monroe, Reference Patterson and Monroe1998: 315).
The centrality of narratives for people can be encapsulated in the concept of ‘narrative identity’ (e.g., Somers & Gibson, Reference Somers, Gibson and Calhoun1994; Hammack, Reference Hammack2008; McAdams & McLean, Reference McAdams and McLean2013). From a psychological perspective, narrative identity is defined as ‘the accumulating knowledge that emerges from reasoning about our narrative memories’ and ‘yields a life story schema that provides causal, temporal, and thematic coherence to an overall sense of identity’ (Singer, Reference Singer2004: 442).
The concept of ‘story’ is a core element in narrative theory. Various scholars often use the terms ‘story’ and ‘narrative’ interchangeably (Spector-Mersel, Reference Spector-Mersel2010). However, in narratology, they are usually differentiated (Herman, Reference Herman2009). ‘While story refers to what is being told, or the “tale”, narrative alludes more broadly to the act of “narrating”, encompassing not only the tale but also the text and its narration’ (Shenhav, Reference Shenhav2015: 6, emphasis added). Despite differences in views regarding the conceptual definition of ‘story’, an accepted narratological approach requires a story to contain at least two chronologically related events, either real or fictive (Prince, Reference Prince1980; Rimmon-Kenan, Reference Rimmon-Kenan2002 [1983]). While stories usually involve more than two events, the above basic requirement allows temporal continuity, inherent to the notion of ‘sequence’. In keeping with this perspective, which highlights a succession of events (e.g., Jaworski & Coupland, Reference Jaworski, Coupland, Jaworski and Coupland1999; Herman, Reference Herman2009), the operational definition of ‘story’ used here is as ‘the chronological sequence of events derived from a narrative, as well as the characters involved in them’ (Shenhav, Reference Shenhav2015: 16). One of the reasons for including characters into the above minimal definition of ‘story’ is that they carry considerable weight in political stories (e.g., Shanahan, Jones & McBeth, Reference Shanahan, Jones and McBeth2011).
Since we are interested in voter turnout, we chose to focus on a special type of stories, national stories – the ones that are about a nation and are also embraced or shared by its people and/or elites. Accordingly, and in light of our operational definition of ‘story’ as a proxy emphasizing key elements of the national story espoused by each respondent and in line with previous research (Sheafer et al., Reference Sheafer, Shenhav and Goldstein2011; Shenhav et al., Reference Shenhav, Oshri, Ofek and Sheafer2014), we use the concatenation of the most important past event, future event, and national character as our tested story components. To define the national story components prevalent in a given society, we aggregate the components of individuals’ national stories (past, future, and character). We argue that such collective stories have a substantial impact on individuals’ political behavior, and specifically, on their willingness to engage in the fundamental political act of voting. It is this kind of impact that is termed here story incentive. While for the use of surveys we adopt a ‘thin level’ of narrative analysis (Shenhav, Reference Shenhav2015), it is fair to assume that the stories components we capture are proxies of narratives respondents were exposed to in the social domain in the broader and richer sense of the concept.
The present study contributes to research on voter turnout by developing this line of reasoning while utilizing narrative approaches, as part of a growing trend of investigating narratives in the political domain (e.g., Patterson & Monroe, Reference Patterson and Monroe1998). This interest is evidenced in recent attempts to incorporate narrative approaches into empirical political studies, such as policy studies (Jones, et al., Reference Jones, Shanahan and McBeth2014); national security (Krebs, Reference Krebs2015a); voting behavior (Sheafer, et al., Reference Sheafer, Shenhav and Goldstein2011); coalition formation (Shenhav, et al., Reference Shenhav, Oshri, Ofek and Sheafer2014); environment policy (Lejano, et al., Reference Lejano, Ingram and Ingram2013); rhetoric (Hammer, Reference Hammer2010), and conflicts and their resolutions (Bar-Tal, Reference Bar-Tal2000).
Story incentive
The idea that stories provide an incentive for voting and thus affect turnout rests upon narratives’ communal function, in the sense that they help us ‘make sense of our reality and understand ourselves as political beings’ (Patterson & Monroe, Reference Patterson and Monroe1998: 315). Stories can connect people across generations and places (Carrithers, Reference Carrithers and Whiten1991). They serve to synchronize the ‘individual’s life cycle with that of a group’ (Shenhav, Reference Shenhav2015: 2; see also Carrithers, Reference Carrithers and Whiten1991) by creating a sense of familiarity and emotional attachments between people who live in different places and even at different times (e.g., Zerubavel, Reference Zerubavel1995; Hammack & Pilecki, Reference Hammack and Pilecki2012). If stories play such an important role in the creation and consolidation of groups, it stands to reason that those who embrace a group’s story will be more likely to participate in that group’s collective activities – in our case, voting on Election Day – than those who do not.
We suggest two mechanisms to account for the effect of national stories on voter turnout. These mechanisms are related to intrinsic and extrinsic benefits of voting (see e.g., Gerber et al., Reference Gerber, Green and Larimer2008). The intrinsic mechanism conceives of voting as an expressive act that represents citizens’ concern for their community (see Coleman, Reference Coleman2013; Blais & Achen, Reference Blais and Achen2019) and allows them to communicate their sense of group belonging to others (Schatz & Lavine, Reference Schatz and Lavine2007). As a symbolic depiction of a nation, our story components can ‘essentialize the group as a transcendent psychological entity, one that connects the individual to larger meaning and purpose’ (Schatz & Lavine, Reference Schatz and Lavine2007, p. 333). Embracing national story components, in and of itself, enhances one’s commitment to the nation, and ‘uniquely accentuate citizens’ identification as national members’ (Schatz & Lavine, Reference Schatz and Lavine2007, p. 332; see Carrithers, Reference Carrithers and Whiten1991). It is the presence of a national story in a person’s mind that fosters meaningful group belonging. Since social stories generally arouse the feeling of ‘groupness’ (Shenhav, Reference Shenhav2015), those who embrace a group’s story are more likely to care for that group, connect to its history, and feel responsible for its future (see also Miller, Reference Miller1995). Therefore, those who embrace components of national stories may use the act of voting to express their commitment and care for the group.Footnote 1
The extrinsic mechanism relates to the social pressures associated with not voting. Embracing group stories not only has the potential to enhance the emotional benefits an individual derives from being part of a wider community but may also increase the fear of social disapproval. Previous studies find that, when individuals believe their social networks will disapprove of the act of not voting, they will be more likely to vote in order to meet the norms of their political community and thus to avoid social isolation (Gerber et al., Reference Gerber, Green and Larimer2008; Panagopoulos, Reference Panagopoulos2010; Abrams et al., Reference Abrams, Iversen and Soskice2011). We expect these social costs to be perceived as more severe among those embracing national stories.Footnote 2 This is because voting is widely acknowledged as an important civic duty (Blais & Achen, Reference Blais and Achen2019). People are expected to project this positive social norm to other community members, as individuals attribute positive characteristics to their fellow in-group members (e.g., Mummendey et al., Reference Mummendey, Otten, Berger and Kessler2000; Iyengar et al., Reference Iyengar, Sood and Lelkes2012; Rothschild et al., Reference Rothschild, Howat, Shafranek and Busby2019). As community attachment increases conformity with its norms (Huddy & Khatib, Reference Huddy and Khatib2007; Turner et al., Reference Turner, Hogg, Oakes, Reicher and Wetherell1987; Jetten et al., Reference Jetten, Spears and Manstead1997), we expect that citizens who embrace national stories will be more deterred by group isolation (Nachmias, Reference Nachmias1974; Marques et al., Reference Marques, Yzerbyt and Leyens1988; Panagopoulos, Reference Panagopoulos2010), and thus will be more likely to vote.
The two mechanisms suggested above for the impact of story incentive are anchored in the assumption that voting is crucial to maintaining an individual’s standing in a national community. In this context, a central parameter that is likely to affect voter turnout is the completeness of the national story. We take an individual’s national story to be ‘complete’ if, when asked directly about her nation, she incorporates three story components: past, future, and a prominent national character. We posit that embracing a more complete story endows an individual with a narrative identity and incentivizes her to participate in that nation’s political activities. For example, we consider a person who depicts the nation only through a major past event (e.g., September 11 in the US context), but omits a prospect for its future and a central character in the nation’s history, as one whose national story is incomplete (this person has only one component). In contrast, a person whose view of the nation incorporates all three components (e.g., September 11 as the past event; better economic climate as a desired future for the nation; and Abraham Lincoln as the most important character in the nation’s history) is regarded as embracing a complete national story. We therefore predict that such a person is more likely to vote.
If narrative identity does indeed play an important role in shaping one’s life trajectory, the lack of this dimension must necessarily divest one’s political identity of an important facet. A prominent advocate of this approach is political philosopher Michael Sandel (Reference Sandel1996), who expressed concern about what he called a ‘storyless condition’: ‘[T]he loss of the capacity for narrative would amount to the ultimate disempowering of the human subject, for without narrative there is no continuity between present and past, and therefore no possibility of acting together to govern ourselves’ (p. 351). To apply Sandel’s argument to the case in hand, in default of a national story, the act of voting may lose some of its symbolic, community-derived meanings. The upshot is that individuals who do not embrace a complete national story may be demotivated to participate in national elections. Put simply, they will be less likely to vote.
Story incentive, then, can be seen as a force that propels citizens to take part in national narration, in line with Coleman’s (Reference Coleman2013) thesis that voting is not solely an expression of preference for certain policies or ideologies but also a performative act that affords voters a symbolic opportunity to become part of their nation. If, by voting, citizens actually participate in the creation of their national story, the act of voting can be perceived as a symbolic juncture between past and future. From this perspective, for individuals lacking important components of national stories – such as national past, future, and characters – the act of voting could be less meaningful, in that it is essentially a crossroads of missing elements.
To sum, our main hypothesis predicts that embracing a national story will be positively correlated with the likelihood of voting. Those who have adopted a more complete national story are expected to feel a greater responsibility for the nation and appreciate more deeply the symbolic significance of the act of voting.
HYPOTHESIS 1 (The National Story Completeness Hypothesis): The more components of a national story a voter embraces, the more likely he or she is to vote.
While story incentive is assumed to be contingent on the extent to which a person embraces any one national story, we also assume that electoral participation can be accounted for by the type of the story adopted. More specifically, we posit that adherence to the nation’s central story components – those consisting of the most popular components of past, future, and character (see Online Appendix B) – will be associated with a higher likelihood of voting, compared to that among individuals who embrace uncommon stories or lack components of national stories. Accordingly, a closer association with one of the most popular stories (e.g., in the US, September 11 as the past event; better economy as the desired future, and Abraham Lincoln as the character), rather than with uncommon ones (e.g., landing on the moon as the past event; changes in energy policy as the future prospect; and Thomas Edison as the character), might indicate that the person perceives herself as standing at the nation’s pivotal crossroads between past and future, and therefore feels a stronger duty to impact its future – a commitment which is symbolized by the act of voting. This expectation is further reinforced by the notion of dominant stories, defined as stories that are ‘embraced by a group and which its members consider normal or desirable or perceive as compulsory’ (Shenhav, Reference Shenhav2015). Indeed, central stories are apt to be taken as par for the course, or as a commonsensical way of perceiving political reality (Krebs, Reference Krebs2015a, Reference Krebs2015b). While central stories are not necessarily dominant in the theoretical sense described above, it stands to reason that their centrality in a given society warrants at least some expectations formulated theoretically in regard to the notion of dominant stories or the idea of ‘master narratives’ (Hammack, Reference Hammack2008; McLean & Syed, Reference McLean and Syed2015). This rationale is also in line with the ‘compulsory nature’ and ‘ubiquity’ of master narratives (McLean & Syed, Reference McLean and Syed2015) and with MacKinnon’s argument (Reference MacKinnon, Brooks and Gewirtz1996: 235) that dominant narratives ‘are not called stories’ but rather ‘reality’. Consequently, those who embrace central stories are more likely to see the nation as a ‘natural’ group of belonging that is worth one’s time and effort investment. Since, as already stated, voting is widely accepted as a citizen’s duty as well as a social norm (Gerber et al., Reference Gerber, Green and Larimer2008; Blais & Achen, Reference Blais and Achen2019), by following this rule, the person will avoid social disapproval and increase her standing in the group. From this perspective, embracing mainstream stories of a political community is tantamount to adhering to its mainstream norms and dominant goals and will thus decrease political isolation (Nachmias, Reference Nachmias1974). We note that eschewing any type of national story component will have a more dramatic effect on the likelihood of voting than embracing uncommon stories, as the former could be an indication of apathy or even alienation, while the latter may point to recalcitrance or reluctance to march in step. One’s offstream position vis-à-vis the central national narrative could detract from one’s motivation to be politically active, albeit not as drastically as total indifference.
However, along with the expectation that adherence to the nation’s central story components will be associated with a higher likelihood of voting, narrative theory can also offer an opposite prediction. In the case in point, those embracing uncommon stories will tend to vote more than those embracing common ones. This rationale follows the idea of ‘counter-narratives’ (Bamberg & Andrews, Reference Bamberg and Andrews2004) or ‘competing stories’, which emphasizes the ‘struggle between stories for hegemony or control’ (Shenhav, Reference Shenhav2015, p. 78). Accordingly, as part of their effort to legitimize the alternative story or even in an attempt to change the group’s master narrative, individuals that hold uncommon stories will be more motivated to vote.
To sum up, our second hypothesis states that the more components of an individual’s national story are shared with the central national stories, the greater the likelihood that she or he will vote. We must keep in mind, however, that the coordinates of this expectation within narrative theory are not clearly defined due to the counter-narrative effect.
HYPOTHESIS 2 (The Story Centrality Hypothesis): The greater the proximity of a citizen’s national story to the respective society’s central stories, the greater his or her likelihood of voting.
At this point, it is important to address the relationship between story incentive and national identity in connection to voter turnout. While scholars emphasize the multi-dimensional nature of national identity, the focus is usually put on the strength of national attachment, be it in terms of national cohesion, national pride, or chauvinism (Huddy & del Ponte, Reference Huddy, Del Ponte, Gustavsson and Miller2019; Gustavsson & Stendahl, Reference Gustavsson and Stendahl2020). These dimensions are employed to gauge self-reported strength of relatedness with the nation (e.g., the extent to which individuals are emotionally close to their nation). Crucially, however, national stories constitute another distinctive facet of the general concept of national identity, tapping perceptions of what the nation is, in terms of key events and characters. In other words, at issue is not the extent to which people are attached to their nation, but rather the presence or the lack of an array of contents, or ‘scripts’ (Singer, Reference Singer2004; see also Schank & Abelson, Reference Schank and Abelson1977), individuals might conceive of when thinking about the nation (see also Schatz & Lavine, Reference Schatz and Lavine2007). It is important to separate the strength of people’s sense of national identity and its subjective meaning to citizens (Gustavsson & Stendahl, Reference Gustavsson and Stendahl2020). This distinction is essential since it may entail different political consequences such as policy support and inter-group relations (Huddy, Reference Huddy2001; Transue, Reference Transue2007).
Specifically, in our context, individuals with higher levels of national attachment do not automatically hold a ‘complete’ national story (one that includes all three components: past, future, and actor). Their strong attachments to the nation can be a result of other factors, such as satisfaction with the country’s economy (Meier-Pesti & Kirchler, Reference Meier-Pesti and Kirchler2003) or age (Huddy & Khatib, Reference Huddy and Khatib2007). Therefore, the effect of the completeness of a national story on turnout is expected to be above and beyond the effect of national attachment (for empirical support for this expectation, see the Results section).
Moreover, the concept of central national stories incorporates a substantively different aspect of one’s relationship with the nation. For example, two compatriots who are strongly attached with their nation might assign this group two completely different meanings (Huddy, Reference Huddy2001), and therefore embrace different national stories: one will regard September 11 (a more central past event in the US context) as the defining past event in the history of the American nation, while the other will consider it to be the landing on the moon (a less central event).
Method
Data were obtained from election surveys in Denmark, Israel, the Netherlands, the UK, and the US. These countries were chosen because they are all developed democracies, but differ in their political systems (presidential and parliamentary democracies) as well as their electoral systems and district magnitude (e.g., extreme proportionality in Israel and the Netherlands or single-member plurality systems in the UK and the US). This variation across the countries sampled makes it possible to extrapolate the results to other developed democracies and draw general inferences regarding the relationship between national stories and voter turnout.
All surveys were administered in the respondents’ native language and incorporated items on a wide set of issues, including story-related questions.Footnote 3 The Danish survey (N = 1000) was conducted during the 2015 parliamentary elections; the Israeli survey (N = 1292) – during the 2013 parliamentary elections; the Dutch survey (N = 1455) – during the 2012 parliamentary elections; the UK survey (N = 1002) – during the 2015 parliamentary elections; the US survey (N = 1001) – during the 2016 presidential elections.
The dependent variable: voter turnout
We measured voter turnout by asking respondents if they (1) will definitely vote, (2) will probably vote, (3) do not think they will vote, or (4) will definitely not vote. A definite intention to vote was coded as ‘1’ and all other responses as ‘0’. In the Netherlands, the response options were (1) Yes, I will cast a vote; (2) No, I will not cast a vote; (3) I do not know yet; and (4) I am not eligible to vote (these respondents were excluded from the analysis). The final turnout variable in the Netherlands was coded such that 1 represented ‘I will cast a vote’ and 0 represented all other responses. Online Appendix A provides the descriptive statistics for the main dependent variable and the independent variables.
The independent variable: national story
The independent variable was constructed in two separate steps: first, we identified the full list of events and characters in each country, and then we constructed the story variables.
Identifying the full list of events and characters
National stories have been previously used in investigations of voting behavior and coalition formation by Sheafer et al. (Reference Sheafer, Shenhav and Goldstein2011), as well as by Shenhav and colleagues (Reference Shenhav, Oshri, Ofek and Sheafer2014). These studies presented respondents with a closed list of major past and possible future national events and asked them to choose the most important item in each category.Footnote 4 Such an approach placed certain constraints on respondents, who had to choose from a limited set of possibilities. By contrast, in the present study, our proxy for national stories is based on open-ended questions, including an item targeting a major historical character. Specifically, participants answered three open-ended questions:
Past Event: When you think about the history of [country] and [the designation for its people, that is, American, British, Danish, Dutch, and Israeli], which past event do you consider to be the most important?
Future Event or Prospect: When you think about the prospects of [country] and [designation for its people], which future event do you wish to see?
Character: If you have to name one person who is most significant for [country] and the [designation for its people] from early history until today, who would that person be?
The above three open-ended questions directly and explicitly target the nation’s past, future, and major character, respectively, and thus tap by proxy respondents’ national stories as defined here. Due to many cases in which respondents referred to the same events and characters in the same way, or very similarly, we labeled these references using a single description. We determined the categories inductively based on respondents’ answers – for both voters and nonvoters – by unifying references to the same, or very similar, events and characters. For example, respondents referred to the same past event, ‘the establishment of Israel’, in multiple ways (‘the establishment of the state’, ‘the country’s independence’, ‘the 1948 declaration of Israel’s independence’); similarly, World War II was mentioned using a variety of designations (WW2, Second World War, etc.). Two independent coders assigned categories (past, future, and character) to 300 answers, based on a list of all coding categories. Intercoder reliability was no lower than alpha Krippendorff 0.83 for the past category, 0.7 for the future category, and 0.87 for the character category. The coding protocol and the most central stories in each country are presented in Online Appendix B.
Construction of the independent variables
Our main independent variable, Completeness of National Story, is measured on a four-level scale: (0) no story: a respondent does not hold any component of a national story (i.e., did not mention any story components in the open-ended questions); (1) an incomplete story with two missing components (e.g., the respondent mentioned only a past event, but no future prospect or a central national character); (2) an incomplete story with one missing component (e.g., no future prospect was mentioned, but both past event and a character were); (3) a complete national story, which incorporates all three essential components stipulated above (i.e., the respondent mentioned past event, future prospect, and a national character). For example, if a Danish respondent mentioned that the most important past event for the Danish people was the introduction of the Basic Law but did not mention a future prospect and an actor, she scored 1 in this measure as her story incorporates only one component. If she mentioned three story components (e.g., the introduction of Basic Law as the past event, a better welfare system as the future prospect, and Thorvald Stauning as the important character), she scored 3 in this measure. Online Appendix A presents descriptive statistics for the completeness variable (ranging from 0 to 3) and for the other independent variables for all five countries in our sample.
To measure the Proximity to the Central Stories variable, namely the proximity of a respondent’s personal national story to the nation’s central stories, we first identified the latter for each country. To the best of our knowledge, research to-date has not developed concrete empirical tools to recognize society’s central stories. We define central stories as ones consisting of the most popular components of past, future, and character components in each society targeted.
To assess the distance between a respondent’s personal national story and the central stories in each country sampled, we created a seven-point scale, ranging from 0 to 6. The scale reflects the extent to which respondents’ answers regarding the three story components (past, future, and character) were consistent with those of the respective countries’ central stories, defined as the five most popular answers for each component of national stories. To each of the three story components mentioned by a respondent, we assigned a score based on the following criteria: not providing a component (e.g., no past component) did not add any points to the scale; a story component that deviated from the nation’s central components (e.g., a past event that is not one of the five most central past events) added one point; and a story component that aligned with the nation’s central components added two points to a respondent’s overall score (i.e., a past event that is one of the five most central such events).
To create the final scale, we summed the scores for each of the three components. Accordingly, 0 represents the maximum distance from the nation’s central stories, while 6 represents a full overlap (i.e., no distance). For example, suppose a British respondent mentioned Winston Churchill as the most significant character in British history, World War 2, as the significant past event, but also mentioned ‘fairer welfare system’ as the nation’s desired future prospect. Since Churchill and World War 2 are among the five most central characters and past events, respectively, but the future prospect designated is not, the respondent scored 5 on the proximity scale, as she is close to the central national story in two components.
Control variables
Based on previous models of voter turnout, we divided the possible predictors into two groups. The first, sometimes called ‘the personal resources model’, includes sociological-demographic variables. As part of this group, we controlled for variables that were previously found as relevant, namely, age and education – individual-level factors that have been shown to increase turnout (Wolfinger & Rosenstone, Reference Wolfinger and Rosenstone1980; Blais, Reference Blais2000; Dalton, Reference Dalton2013; Smets, Reference Smets2016). We also controlled for gender (Carreras, Reference Carreras2018), as well as religiosity and income (Blais et al., Reference Blais, Gidengil and Nevitte2004).
The second group of covariates focuses on political attitudes and involvement – factors that are known to explain various types of political behavior, including turnout (Miller et al., Reference Miller, Shanks and Shapiro1996). This group is usually referred to as ‘the psychological model’ or ‘the political-attitude model’ (e.g., Verba, et al., Reference Verba, Schlozman and Brady1995; Dalton, Reference Dalton2013). Of this group of variables, we controlled for political trust (Bélanger & Nadeau, Reference Bélanger and Nadeau2005), political discussion (Hopmann, Reference Hopmann2012), and news consumption (Schuck et al., Reference Schuck, Vliegenthart and De Vreese2016). The operationalization of each control variable is elaborated in Online Appendix C. We also controlled for country-specific dummies in order to account for country-level differences (e.g., Blais & Rubenson, Reference Blais and Rubenson2013).
Results
According to the concept of story incentive, adherence to a national story is expected to increase the likelihood of voting. Moreover, we assumed the centrality of a person’s national story to be important as well, further enhancing such probability.
Completeness of national story
According to our main hypothesis (H1), the more complete a person’s national story, the higher the likelihood that she will vote. Figure 1 is based on a logistic regression estimating turnout (binary variable) in all five countries as a function of national story completeness while controlling for relevant alternative explanations (full results are in Table D1 in Online Appendix D). It presents the relationship between completeness of the national story and voting turnout. The results show that the more complete a national story in the respondents’ minds, the greater the probability that they will vote. For example, if a respondent mentioned only a character, but no past or future event, she scored 1 on the completeness variable, and her probability of voting was estimated as 84%. The probability of voting for respondents who mentioned all three components (character as well as past and future events) was 90%.
The relationship between the Completeness of National Story and Turnout Probability (with 95% confidence intervals). (1) Y axis presents the predicted probabilities for voting within different values of the national story variable (X axis). (2) Results are derived from Table D1 (Model 1) in Online Appendix D.
Thus, in line with the first hypothesis, completeness of the respondents’ national stories was found to have a positive and significant impact on their likelihood of voting. As in previously reported studies (see above), the psychological and personal resource variables emerged as strong predictors of voting intentions.Footnote 5
Although the effect size of the story completeness variable is relatively small, we believe that it should not be taken lightly for two reasons. First, in close races, even small changes in turnout rate can have a large effect on the final election results – both in first-past-the-post systems and in proportional systems, where parties must win a minimum vote share to get seats in parliamentFootnote 6. Second, the effect size of the story variable is quite similar to other important covariates. For example, the difference between the minimum level of education (no education or only elementary school) and the maximum level (MA degree and above) is associated with a 10% increase in the likelihood of voting. Age, another well-known covariate, increases turnout probability all in all by 8%, and the effect of political trust is even smaller (a 6% increase). In comparison, for the completeness variable, the difference between the minimum (no story) and the maximum value (complete story) is associated with an 11% increase in turnout. It should be noted that, on the whole, effects on turnout found in the existing literature are modest (e.g., Dalton, Reference Dalton2013), possibly because political participation is shaped by many factors, such as personal resources, political attitudes, interpersonal networks, and others (Verba et al., Reference Verba, Schlozman and Brady1995; Dalton, Reference Dalton2013).
Proximity to the central stories
According to the second hypothesis, the greater the proximity between an individual’s national story to his or her country’s central stories, the greater the likelihood of voting. Table 1 presents the results of a logistic regression estimating turnout as a function of the proximity to the common stories variable while controlling for relevant alternative explanations. The analysis presented in Model 1 contains all of the respondents, including those who lack several components of national stories. As expected, proximity to common stories has a significant and positive impact on voting intentions.
Story incentive: the relationship between proximity to central stories and voter turnout
† P < 0.1, *P < 0.05, **P < 0.01, ***P < 0.001.
(1) The statistical analysis is a binary logistic regression, with the dependent variable being an intention to vote. Standard errors are indicated in parentheses. (2) The reference category for the country-specific dummies is Denmark. (3) Model 1 includes the full sample; Model 2 includes respondents who had at least one story component (out of three); Model 3 includes respondents who had at least two components (out of three); and Model 4 includes respondents with all three story components.
Models 2–4 present the same analyses but when gradually excluding respondents with incomplete stories (e.g., Model 2 includes respondents who mentioned at least one story component). Results show that, while the original associations remain stable among those with at least one or two story components (Models 2 and 3, respectively), the impact of proximity to the central stories on turnout becomes statistically insignificant among those embracing a complete national story (i.e., respondents with all three components; Model 4). These results provide an initial indication that the impact of story proximity is contingent on respondents’ story completeness.
We therefore further examined the unique impact of embracing common versus uncommon stories among respondents with a complete national story. Table 2 presents the results of a logistic regression estimating the differences in voting between individuals whose story components belong to the common national story compared to those with uncommon story components. The results show a positive coefficient, meaning that embracing common stories is associated with a higher likelihood to vote compared to uncommon stories, but this relationship is statistically insignificant.
The relationship between embracing common versus uncommon stories and voter turnout
† P < 0.1, *P < 0.05, **P < 0.01, ***P < 0.001.
(1) The statistical analysis is a binary logistic regression, with the dependent variable being an intention to vote. Standard errors are indicated in parentheses. (2) The reference category for the country-specific dummies is Denmark. (3) The analysis was conducted among individuals who embraced a complete national story (all three components). (4) The story dummy variable compared between individuals whose story components all belong to the nation’s central stories and individuals whose components are uncommon stories.
Taken together, we cannot confirm the proximity hypothesis (H2) as it is mostly supported when taking into account all respondents, including those with incomplete national stories. We also did not find support for what we termed as ‘the counter-narrative effect’ of uncommon stories, which may result in higher turnout rates for those embracing uncommon stories. We believe that these inconclusive findings can provide helpful guidelines for further research that will use content-sensitive analyses, thus allowing a cross-country narrative analysis (we elaborate more on this in the Conclusion section).
Robustness tests
We empirically tested the relationship between the strength of a person’s national attachment and our main story variables. We used Huddy and Khatib’s (Reference Huddy and Khatib2007) questions, designed to capture ‘national identity’ in the US. In line with previous studies demonstrating that citizens’ ties with national symbols increase the strength of their national attachment (Schatz & Lavine, Reference Schatz and Lavine2007), we found positive but weak correlations between attachment to the nation and (1) completeness of a national story (r = 0.187) and (2) proximity to the most popular national stories (r = 0.194).Footnote 7 Moreover, we added to the regression analysis a covariate tapping the strength of a person’s attachment to her nation. Results, presented in Online Appendix E, remain stable, that is, the impact of story completeness on turnout is above and beyond the strength of national attachment (no collinearity was observed). Taken together, this robustness test suggests that our story variables tap a different facet of national identity than previous measures that capture the strength of a person’s national belonging. Our story variables tap the story people conceive of when they think about their nation and the extent to which this story is shared by their compatriots.
We also tested our models in each country separately. Despite differences in regime type and electoral systems, the results, presented in Online Appendix F, support our original conclusions and generally hold across countries.
Conclusion
The importance of voter turnout for the legitimacy and the outcomes of elections is well established in scholarship. Nonetheless, our understanding of what drives turnout remains far from comprehensive. This paper suggests a new explanation for voter turnout – one grounded in the role that social and political narratives occupy in people’s identity. It introduces the idea of ‘story incentive’, an assumption that embracing a group’s social story will incentivize individuals to take part in that group’s political activities.
Our main hypothesis (H1) was that the likelihood of voting would be higher among people who have a stronger connection with any kind of national stories, and the more so for those who espouse any complete version of a national story. Our additional hypothesis (H2) was that the greater the proximity between an individual’s national story and the central stories in the respective country, the greater the likelihood that she will vote on Election Day. Our analyses provided an initial indication that the impact of story proximity is contingent on respondents’ story completeness, and therefore this hypothesis was not confirmed.
Overall, the empirical evidence obtained supports our theorizing that voting is incentivized when a national story is part of a citizen’s consciousness, as indicated in our main hypothesis. Such an association between the sense of belonging to a national community and the extent to which one’s narrative identity is shared by one’s compatriots resonates with the vast body of literature on the social and political role of sharing narratives. Moreover, our results contribute to the literature concerned with the political consequences of different dimensions of national identity (e.g., Huddy & del Ponte, Reference Huddy, Del Ponte, Gustavsson and Miller2019; Gustavsson & Stendahl, Reference Gustavsson and Stendahl2020). We have shown that national stories can be considered as a unique facet of this multidimensional construct that is beneficial to the nation as a whole, due to an association observed across countries between embracing national stories and greater responsibility to the national community manifested in the act of voting.
From a comparative perspective, previous studies found that different institution-level factors – such as power concentration and proportional electoral systems – yield different incentives for citizens to vote (Jackman, Reference Jackman1987; Blais, Reference Blais2006). While the nuances in the relation between story incentive and different electoral systems undoubtedly warrant further research, the conclusions of the study presented here generally hold across five countries which differ in their institutional settings. This, in itself, underscores the importance of national stories embraced by citizens for their decision to vote, above and beyond system-level characteristics.
Our findings also shed light on what we term a narrative meaning of voting. Voting can be perceived as a symbolic act that helps citizens feel they are taking part in establishing a link between their nation’s past and its future. Accordingly, if a person does not assign importance to her country’s past or future, the act of voting seems to lose some of its meaning. This rationale constitutes our theoretical contribution to the understanding of voter turnout.
By suggesting and validating identity-driven explanations for individuals’ likelihood to vote, our findings empirically account for voter turnout from a fresh perspective. Addressing this issue from the new angle presented here may help overcome the challenges encountered by a long line of scholarship on the voting paradox and contribute to the normatively oriented discussion on democracy and political participation. Thus, for example, our findings validate Sandel’s (Reference Sandel1996) concern over the implications of the ‘fragmented storyless condition’ into which, as he cautions, we are in danger of slipping ‘individually and collectively’ (p. 351). It appears, therefore, that studying the effects of story incentive should not be dissociated from this normative debate – especially since Sandel’s misgivings resonate with the communal function of narratives discussed above. Sandel elaborates this notion in a Webchat hosted by The Guardian:
Shared narratives are what hold communities together. The ability to tell stories is to see one’s life as part of a larger story. And so the capacity for narrative is what situates us in the world. By the drift to storylessness, I mean the tendency in modern life to think of ourselves without reference to larger frameworks of meaning and belonging. One of the great moral and civic challenges of our time is to recover our capacity to tell compelling stories that relate our identities to shared communities of value and meaning (Sandel, Reference Sandel2014).
Sandel’s words hark back to a widely known array of scholarly assumptions about the role of narratives in human society. In our view, and judging by the growing interest in empirical narrative research (e.g., Shanahan, Jones & McBeth, Reference Shanahan, Jones and McBeth2011; Jones & McBeth Reference Jones and McBeth2010; Jones & Radaelli, Reference Jones and Radaelli2015), the time has come to test these theories empirically – albeit with several important caveats, which apply to the study presented here as well.
Needless to say, it would be quite impossible for any research that undertakes this challenge to give justice to the full richness of narratives in the societal and political domains. For our part, to capitalize on the advantages afforded by a large-N cross-national research, we utilized a thin-level narrative analysis (Shenhav, Reference Shenhav2015) that rests upon proxies of stories rather than analyze stories in-depth.
As an alternative compromise, one might elect to dispense with the comparative aspect of empirical narrative research and study voters’ perceptions based on a nuanced interpretive analysis of a single case (e.g., Marin & Leo, Reference Marin and Leo2019). Another remedy, more promising particularly for cross-country analysis, is to apply the theoretical framework that hinges on the concepts of master- or meta-narrative (e.g., Bamberg & Andrews, Reference Bamberg and Andrews2004; Hammack, Reference Hammack2008, McLean & Syed, Reference McLean and Syed2015) and zooms in on the main attributes and principles that characterize a set of stories rather than on the specific contents and distinctive features of a single story. Thus, theoretically, these concepts can furnish observations that are comparable across multiple case studies. Recently, master narratives have been defined and analyzed in a number of psychology investigations (e.g., McLean & Syed, Reference McLean and Syed2015; Syed & McLean, Reference Syed and McLean2020), which can offer a conceptual path for political scientists who wish to adopt this framework to explore such key political questions as ‘the dynamic relation between the self and competing narratives’ (Syed & McLean, Reference Syed and McLean2020).
Supplementary material
To view supplementary material for this article, please visit https://doi.org/10.1017/S1755773920000399.
Acknowledgments
This research was supported by the Israel Science Foundation (Grant No. 1400/14) and Danish Council for Independent Research (FSE) (Grant No. 0602-02106B). The authors would like to thank James Stanyer, Yaron Ezrahi, Pazit Ben-Nun Blum, the editors of European Political Science Review and four anonymous reviewers for very helpful comments and suggestions.
