Electoral Studies xxx (xxxx) xxx
Contents lists available at ScienceDirect
Electoral Studies journal homepage: www.elsevier.com/locate/electstud
Notes on recent elections
What citizens learn from elections✩ The normative consequences of voter eligibility Zohreh Khoban Department of Government, Uppsala University, Box 514, 75120 Uppsala, Sweden
ABSTRACT Advocates of innovations for ways to expand citizens’ political participation claim that institutions that require more in-depth participation than voting are required to attain a democratic system with profound civic engagement. They often base this claim on the assumption that elections fail to encourage citizens to be socially and politically involved. In this paper I challenge this assumption by exploring whether voter eligibility reinforces the notion that a good citizen participates in politics. Applying a regression discontinuity design that uses the voting-age restriction as the threshold, I find that voter eligibility boosts support for extensive (non-electoral) forms of political participation. Hence, I argue that the importance of elections should not be underestimated in the quest for societies with civic-minded citizens.
1. Introduction Civic engagement is of fundamental importance to the vitality of democracy.1 Due to growing concerns about citizens’ civic disengagement and distrust in government, the last two decades have seen a growing interest for institutional innovations that could re-engage a disillusioned and disenchanted citizenry (Newton, 2012; Smith, 2009). In these discussions, there is a tendency to categorize elections as an institution that supports a minimal model of democracy. Scholars argue that because elections only require citizens to cast a ballot, they do not facilitate citizens’ active involvement (Fung and Wright, 2003; Goodin, 2008). In-depth engagement, they assert, requires citizens’ direct participation in institutions such as citizens’ assemblies, citizens’ juries, and 21st century town meetings. The underlying logic is that voting encourages minimal political commitment, while intense forms of political participation allow for extensive engagement. To visit a polling station and cast a vote is indeed one of the least time-consuming forms of political participation. However, only focusing on what citizens are asked to do on election day could provide an incomplete picture of what elections truly mean for citizens, and for democratic societies. As deliberative democrats have recently pointed out, it is necessary to study the impact of political institutions on the overall democratic system (Mansbridge et al., 2012), and to recognize that successful citizen participation cannot be artificially engineered, but is ‘‘an emergent property of the norms, expectations and ethos that together make up the political culture of a society’’ (Böker, 2017). Thus, when evaluating the significance of elections for democracy, it is
relevant to ask whether they affect citizens’ beliefs and behavior in a way that ultimately increases their civic engagement. In this paper I study the normative consequences of elections by examining whether voter eligibility affects citizens’ perceptions of how a good citizen should participate in politics. I use voter eligibility as a measure of voter-related experiences that range from the insight that one is eligible to vote to the act of voting. The analysis is based on data from ten democratic countries that participated in the Civic Education Study (CivEd) carried out by the International Association for the Evaluation of Educational Achievement (IEA) in 1999–2000. Theoretically, I lean on a tradition that can be traced back to the work of de Tocqueville ([1835-40] 1969), and in which scholars have claimed that the power exercised through the right to and practice of political participation generates a sense of political ownership and responsibility. This process has been described as an ‘‘educative effect’’ of political participation (Mansbridge, 1995). Based on this set of theories, I argue that the educative effect of political participation is likely to involve an increased support for the idea that a good citizen is politically active. Building on the mapping of norms in previous research (Theiss-Morse, 1993), I test whether voter eligibility enhances the idea that a good citizen is one who is involved in electoral forms of political participation (H1) and/or the idea that a good citizen should be engaged in more extensive forms of political participation (H2). Testing the effect of electoral participation on citizenship norms is empirically challenging. First of all, simply comparing voters with nonvoters is likely to bias the results as those who vote in elections are a
✩ I thank Olle Folke, Nils Hertting, Markus Holdo, Sebastian Axbard, Roza Khoban, Fredrik Sävje, the Polsek seminar at Uppsala University, and three anonymous reviewers for their helpful comments. E-mail address:
[email protected]. 1 In this paper, civic engagement is defined as ‘‘working to make a difference in the civic life of our communities and developing the combination of knowledge, skills, values and motivation to make that difference’’ (Ehrlich, 2000, vi).
https://doi.org/10.1016/j.electstud.2019.102090 Received 11 December 2018; Received in revised form 12 September 2019; Accepted 28 September 2019 0261-3794/© 2019 Elsevier Ltd. All rights reserved.
Please cite this article as: Zohreh Khoban, Electoral Studies, https://doi.org/10.1016/j.electstud.2019.102090
Electoral Studies xxx (xxxx) xxx
Z. Khoban
be altered among eligible non-voters.2 Thus, voter eligibility is the theoretically accurate way to study the relationship between voter-related experiences and citizenship norms.3
highly selected group (Blais, 2000; Brady et al., 1995; Wolfinger and Rosenstone, 1980). This particular group is likely to differ in political norms in the first place (potentially this is why some of them decided to vote). Moreover, this strategy only captures the effect of actual voting, not the effect of voter eligibility, which is of theoretical interest in this paper. Another strategy is to compare all individuals who are eligible to vote with all individuals who are not. However, also this comparison is problematic as citizenship norms differ between generations (Dalton, 2008). In order to tackle these issues, I follow an increasingly common practice in the study of political socialization (e.g. Dinas, 2012; Meredith, 2009; Mullainathan and Washington, 2009), and estimate the effect of voter eligibility by using the voting age restriction as a threshold in a regression-discontinuity design. Thus, I compare those who turned 18 (and thus became eligible to vote) a short period before an election with individuals who turned 18 just after the same election. This design ensures that the assignment of the eligibility to vote is as good as random for individuals close to the threshold, and that these individuals are comparable in terms of personal characteristics and social experiences. As a consequence, the design enables me to capture the causal effect of voter eligibility on citizens’ perceptions of good political behavior. I find that voter eligibility does not enhance the idea that a good citizen engages in electoral forms of political participation. However, it boosts the idea that a good citizen engages in more extensive forms of political participation. The support for the extensive forms of participation increases by 5%–7% at the voting age threshold. This finding contradicts the assumption that elections only ask citizens to cast a ballot. It shows that elections help in deepening citizens’ civic engagement by fostering desirable norms of citizenship.
The expectation that voter-related experiences could affect citizenship norms stems from a long tradition of scholars who have claimed that the power exercised through both the right to and the use of the right to political participation generates a sense of political ownership and responsibility (Kaufman, 1960; Mill, 1862; Pateman, 1970; de Tocqueville [1835-40] 1969). The agenda of these scholars has varied substantially. So has the type of participation they have focused on. However, the recurring idea is that involvement in decision-making makes citizens feel like co-owners of society, and urges them to ask what they can do to improve it. Since this process of claiming a common good as one’s own is considered to imply the development of citizens’ character, it has been described as an ‘‘educative effect’’ of political participation (Mansbridge, 1995). In the case of suffrage, this effect is acknowledged by Mill: [I]t is from political discussion and collective political action that one whose daily occupations concentrate his interests in a small circle round himself, learns to feel for and with his fellow citizens, and becomes consciously a member of a great community. But political discussions fly over the heads of those who have no votes, and are not endeavoring to acquire them. . . . nothing depends on the decision they may arrive at, and there is no necessity and very little inducement to them to come to any (Mill, 1862, p. 172).4 By treating citizenship norms as the outcome in this study I stretch the process of the educative effect of political participation from considering citizens’ attitudes about themselves as social and political beings, to also involving their image of how good citizens should behave. I believe this theoretical extension is reasonable because the feeling of co-owing society presupposes the recognition of co-existence (there is a ‘‘co’’ in co-ownership), which is likely to highlight the need for collective engagement. In other words, citizens who exercise power through political participation, and develop political ownership and responsibility, are likely to stress collective actions for taking such responsibility. Thus, the rise of political ownership and responsibility is likely to bring about an emphasis on the idea that good citizens are politically engaged.
2. Theory and previous studies Empirical studies on citizens’ participation in elections often seek to explain the causal determinants of voting (e.g. Blais, 2006; Geys, 2006). Alongside these studies, a few scholars approach electoral participation as the independent variable. Some of them share the above-mentioned interest in explaining voter turnout, and analyze whether voting is habit forming (Denny and Doyle, 2009; Fujiwara et al., 2016; Gerber et al., 2003; Green and Shachar, 2000; Meredith, 2009; Plutzer, 2002). Others show that voters develop external political efficacy (Finkel, 1985, 1987) and adapt positive attitudes towards the candidate for whom they vote (Beasley and Joslyn, 2001; Mullainathan and Washington, 2009; Regan and Kilduff, 1988). There is also research on whether changing the costs and incentives for electoral participation affect individuals’ attitudes and behaviors. For example, studies show that compulsory voting does not lead to more interested, knowledgeable and engaged citizens (Holbein and Rangel, 2016; Loewen et al., 2008). Furthermore, it has been found that home registration visits and receiving information on voter registration increases individuals’ political interest and information (Braconnier et al., 2017; Shineman, 2018). However, scholars have thus far not studied the effect of elections on the public’s image of a good citizen—that is, its impact on citizenship norms. In this paper I study whether voter eligibility increases the support for the idea that citizens should be politically active. I choose to study the effect of voter eligibility because I am interested in a broad spectrum of voter-related experiences, ranging from knowing about one’s eligibility to vote, to various forms of election-related activities such as reading information from electoral authorities, taking an election quiz, and voting. Hence, I do not limit the study to the effect of voting. This is because the nature of the dependent variable makes it difficult and uninteresting to isolate the act of voting from experiences that enable or help individuals cast a vote. In fact, norms are more likely to be affected by pre-voting information acquisition, and could therefore also
Since I suspect that the educative effect of election-related forms of participation could boost citizenship norms that promote the importance of being a politically engaged citizen, I test whether voter eligibility enhances the idea that a good citizen is one who is involved in electoral forms of political participation (H1) and/or the idea that a good citizen should be engaged in non-electoral (i.e., more extensive) forms of political participation (H2). Previous research has identified these two ideas about political participation as prevailing citizenship norms that emphasize the importance of citizens’ active political involvement (Theiss-Morse, 1993).
2 Individuals might be unaware of their voter eligibility, which would make it a deficient measure of voter-related experiences. However, 91% of the individuals in the dataset responded that voting is a political right. This indicates high levels of awareness. 3 Since voters could be argued to be more likely to engage in elections and to undergo changes in norms, it would be interesting to compare the effect of voter eligibility to the effect of voting. However, due to lack of data on whether people voted I am not able to make such a comparison. 4 Note that Mill seeks to extend the suffrage. The scenario he describes is most likely for those who have no vote and no prospect of obtaining it.
2
Electoral Studies xxx (xxxx) xxx
Z. Khoban
Fig. 1. The effect of voter eligibility on support for electoral and extensive forms of political participation.
Fig. 2. The effect of voter eligibility on support for voting, party membership, political discussions, and participation in activities to benefit the community.
3. Research design and data
the threshold, the process that determines treatment assignment can be considered as good as random, which ensures that these units will be similar with respect to potential confounders (Dunning, 2012). Following Coppock and Green (2016), Dinas (2012), Holbein and Rangel (2016), Meredith (2009), Mullainathan and Washington (2009) and Nyhan et al. (2017), I use the voting-age restriction as a threshold in the RDD analysis. This means that in order to identify the relationship between voter eligibility and citizenship norms, I compare individuals who turned 18 (and thus became eligible to vote) a short period before a national election with those who turned 18 just after the same election. The advantage of comparing individuals for which voter eligibility is a matter of which day of the month (and month of the year) they are born is that the distribution of the eligibility to vote can be considered as if random close to the cutoff. This in turn means that known and unknown confounding variables will be balanced across those with and without voter eligibility. Thus, factors
To study the effect of voter eligibility, I compare individuals who, due to voting-age restrictions, had the opportunity to vote for the first time in the most recent national election and those who did not. Feelings of political ownership and responsibility are thought to be what drives the positive relationship between voter eligibility and support for the idea that a good citizen is one who is engaged in politics. Therefore the opportunity to vote for the first time implies a most likely effect, since first-time voting is more likely than subsequent voting to enhance the feeling of political power, which in turn could generate political ownership and responsibility. I use a regression discontinuity design (RDD) to test the causal effect of voter eligibility. RDD approaches can be applied when there is a threshold determining the independent variable, above or below which individuals or units cannot influence their position. For units very near 3
Electoral Studies xxx (xxxx) xxx
Z. Khoban
such as age, education and income will not drive the relationship between voter eligibility and citizenship norms. In contrast, a method that would compare all first time voters to all those who are not eligible to vote would have to include controls for variables such as age and level of education. Even if these most basic controls are included, it is impossible to figure out and control for all relevant factors that could explain differences in norms between these two groups. To carry out the RDD analysis I estimate two separate regression functions (one above and one below the voting-age threshold) and calculate the estimated effect as the difference between the two estimated intercepts. More specifically, I estimate a local linear regression of support for different citizenship norms on the number of days between the individual’s 18th birthday and election day, and compute weights by applying a triangular kernel function on the distance of each observation’s number of days from the cut-off. I follow common practice and select an optimal bandwidth that minimizes mean-squared error (MSE). Since MSE minimization leads to bandwidth choices that are too large for conventional confidence intervals to be valid, I use the robust bias corrected confidence intervals together with the bias corrected point estimates developed by Calonico et al. (2014, 2017). For implementation, I use the rdrobust software.5 The data in this paper is obtained from the civic education study (CivED) carried out by the International Association for the Evaluation of Educational Achievement (IEA) in 1999–2000. The IEA collected the data by surveying over 50,000 upper-secondary students about their civic concepts and attitudes. Since the data contains information on the respondents’ exact birthday, respondents can be categorized according to whether they were eligible to vote in the last national election in their country. The time period between the last national election and the date of survey varies between one month and about two and a half years between different countries. The median time between elections and date of survey for countries included in the main analysis is 15 months. See Online Appendix A for information on the date of survey and the date of election in each country. The surveyed individuals in the CivED dataset are 14–24 years old. To compare similar individuals I restrict the sample to those who where 17–19 years old at the time of the election.6 This limits the sample to about 14 000 observations. Of the 16 countries in the dataset, I include 10 in my main analysis: Chile, Czech Republic, Denmark, Estonia, Latvia, Poland, Portugal, Slovenia, Sweden, and Switzerland. Three countries are excluded on the basis that they were not free (according to the Freedom House Index) at the time of the election. This is because it is unlikely that citizens in non-free countries would feel that the opportunity to vote gives them political power and ownership.7 Three countries are excluded because their sample contains less than 10 individuals with voter eligibility. See Online Appendix A for details. The outcome variables in this study are the ones used in section B of the CivED survey. This section contains 15 statements about how a good citizen behaves. The students were asked to rate the importance of each statement on a scale from 1 ‘‘not important’’ to 4 ‘‘very important’’. Two statements in this section are suitable for measuring respondents’ support for electoral forms of political participation: (a) ‘‘an adult who is a good citizen votes in every election’’ and (b) ‘‘an adult who is a good citizen joins a political party’’. A further two statements capture support for more extensive forms of electoral participation: (c) ‘‘an adult who is a good citizen engages in political discussions’’ and (d) ‘‘an adult who is a good citizen participates in
activities that benefit people in the community’’. These variables are analyzed both pairwise and separately. See Online Appendix B for details on the coding of all outcome variables. 4. Results As is standard in RDD analyses, I start with a graphical analysis. Fig. 1 shows RD plots in which the number of bins is chosen to mimic the underlying variability of the raw data (Calonico et al., 2014). The horizontal axes represent the distance in days between each observation’s 18th birthday and the election day. Thus, the vertical lines in the middle of the plots illustrate the voting-age threshold. The vertical axes represent support for the two electoral forms of political participation combined (left plot) and the two more extensive forms of political participation combined (right plot). The width of the graphs correspond to the optimal bandwidths suggested by Calonico et al. (2014, 2017). See Online Appendix C for RD plots that are not restricted by the optimal bandwidth estimator, and that include confidence intervals for local means within each bin to illustrate the variability of the underlying data. The fitted line in the left plot in Fig. 1 shows a drop in support for electoral forms of political participation at the voting-age threshold. However, the drop would have been less noticeable if the solid lines were not plotted, which indicates that the effect might be ambiguous. The plot to the right shows a jump in support for more extensive forms of political participation. According to the fitted line in this figure, the support for extensive forms of participation increases from about 2.67 to 2.8 (i.e., about 5%) on a scale from 1 to 4. In contrast to the lefthand plot, the average effect at the cutoff in this plot is detectable by observing the bins around the threshold. Fig. 2 displays the results for the four values separately and confirms the directions of the more general relationships suggested in Fig. 1. For electoral forms of participation, Fig. 2 shows that the approval rating for the statements that a good citizen votes in every election and joins a political party drops at the voting-age threshold. As in Fig. 1, none of the drops is clearly noticeable. For the two more extensive forms of political participation, Fig. 2 shows that the approval rating for the statement that a good citizen engages in political discussions increases from about 2.35 to 2.5 (i.e., about 6%) at the threshold, while the rating for the statement that a good citizen takes part in activities to benefit people in the community increases from about 3 to 3.2 (i.e., about 7%). In accordance with Fig. 1, the shifts in support for extensive forms of participation are more apparent than those for electoral forms of participation. The corresponding regression results for these figures are shown in Table 1, which confirms that there is no effect of being eligible to vote on support for electoral forms of political participation. The point estimates are small and statistically insignificant. This is the case for both the combined measure (column 1) and for voting (column 2) and party membership (column 3) separately. However, the positive effects of voter eligibility on support for extensive forms of political participation (columns 4, 5 and 6) are statistically significant at the 5% or 10% level. As Table 1 shows, these findings are largely robust to estimations using country fixed effects or using a second degree polynomial. The exception is the support for political discussion, which is slightly lower and no longer significant at the 10% level when using country fixed effects.8 In summary, the findings in this section provide support for H2, but not for H1. When comparing individuals who became eligible to vote shortly before the election and individuals who became eligible to
5
Software available at https://sites.google.com/site/rdpackages/rdrobust. More precisely, I restrict the sample to individuals who were eligible to vote up to one year before and one year after the election day. As can be seen in Online Appendix A, the voting age for national elections is 18 in all countries. However, the voting age in Estonia and Switzerland is 16 for local elections. 7 I compare free and non-free countries in an effort to trace the mechanism in Online Appendix F. 6
8 Online Appendix D shows the effect of voter eligibility on support for extensive forms of political participation separately for each country. The point estimates are in general similar to the aggregated regression, but less precisely estimated.
4
Electoral Studies xxx (xxxx) xxx
Z. Khoban
Fig. 3. Bandwidth sensitivity tests.
Table 1 The effect of voter eligibility on support for electoral and extensive forms of political participation. Electoral
Baseline specification Voter eligibility
Robust bias corrected 95% CI Order Loc. Poly. Bandwidth left/right Eff. Number of obs L Eff. Number of obs R Observations Second order polynomial Voter eligibility
Robust bias corrected 95% CI Order Loc. Poly. Bandwidth left/right Eff. Number of obs L Eff. Number of obs R Observations Country fixed effects Voter eligibility
Robust bias corrected 95% CI Order Loc. Poly. Bandwidth left/right Eff. Number of obs L Eff. Number of obs R Observations
Extensive
Combined (1)
Vote (2)
Party (3)
Combined (4)
Discussion (5)
Active (6)
−0.0467 (0.075)
−0.0494 (0.090)
−0.0657 (0.094)
0.132** (0.060)
0.155* (0.092)
0.203** (0.090)
[−0.193,0.100] 1.000 96.5 1237 813 13,233
[−0.225,0.126] 1.000 107.3 1601 944 13,726
[−0.250,0.118] 1.000 91.6 1118 791 13,319
[0.015,0.249] 1.000 109.4 1586 912 13,144
[−0.025,0.334] 1.000 102.8 1394 877 13,329
[0.026,0.379] 1.000 66.8 757 619 13,596
−0.0687 (0.083)
−0.0461 (0.099)
−0.0916 (0.104)
0.202** (0.085)
0.222* (0.118)
0.223** (0.101)
[−0.232,0.095] 2.000 146.7 2618 1137 13,233
[−0.240,0.147] 2.000 169.3 3431 1339 13,726
[−0.296,0.112] 2.000 137.6 2396 1105 13,319
[0.035,0.370] 2.000 101.6 1350 858 13,144
[−0.010,0.454] 2.000 116.1 1798 972 13,329
[0.025,0.422] 2.000 106.9 1526 926 13,596
−0.0304 (0.073)
−0.0203 (0.088)
−0.0567 (0.092)
0.137** (0.062)
0.137 (0.092)
0.254*** (0.086)
[−0.174,0.114] 1.000 86.4 1000 748 13,233
[−0.194,0.153] 1.000 94.9 1248 839 13,726
[−0.238,0.124] 1.000 92.6 1139 801 13,319
[0.015,0.258] 1.000 102.3 1371 864 13,144
[−0.043,0.317] 1.000 96.3 1266 825 13,329
[0.085,0.422] 1.000 63.2 718 593 13,596
Notes: Bias corrected point estimates and robust standard errors (in parentheses) following Calonico et al. (2014, 2017). Running variable is distance of each observation’s number of days to being eligible to vote. Estimate is average treatment effect at cut-off estimated with local linear regressions with triangular kernel and MSE-optimal bandwidth chosen according to CCT implementation. ***Indicate significance at the 1 percent level. **Indicate significance at the 5 percent level. *Indicate significance at the 10 percent level.
5. Robustness checks
vote shortly after, voter eligibility does not seem to increase support for the idea that a good citizen engages in electoral forms of political participation. However, it does increase support for the belief that a good citizen engages in more extensive forms of political participation.
To test whether the results in the previous section are sensitive to the choice of bandwidth, I examine how the estimates from the local linear specifications change as I alter the estimation window over a wide range. The results of this analysis are presented in Fig. 3, which plots the estimates and their 95% confidence intervals. As the figure shows, for bandwidths larger than about 50 days, the estimates for support for extensive forms of political participation are significant at the 5% level over a wide range of bandwidths. The estimates for electoral forms of political participation are not significant for any bandwidth. These results support the findings in the regression analyses
The positive effect of voter eligibility on support for extensive forms of participation is eye-opening as it shows that elections play a part in fostering civic engagement—at least on a normative level. Also noteworthy is the absence of a positive relationship between voter eligibility and support for electoral forms of political participation. While I cannot investigate the reasons for the absence of an effect here, I will suggest some explanations in the concluding section of the paper. 5
Electoral Studies xxx (xxxx) xxx
Z. Khoban
above by showing that the estimates in the regressions are not sensitive to the choice of bandwidth. To validate that the difference between those with and without voter eligibility is as good as random, I conduct two robustness checks. Here, I provide a summary of these tests. Online Appendix E contains the corresponding tables and figures. First, I run the McCrary test of balance in the density of observations around the threshold in order to detect whether any manipulation is taking place around the threshold. I find no significant difference in the number of persons on either side of the eligibility threshold, which rules out the possibility of such manipulation. Second, I examine whether predetermined variables that should not be affected by the treatment are balanced across the threshold. For this part of the analysis, I perform graphical analyses as well as placebo regressions on four predetermined covariates: sex, number of family members at home, daily newspaper at home, and socialization with friends. This analysis suggests that the covariates are balanced around the threshold. Taken together, the robustness checks help verify that assignment to voter eligibility is as good as random close to the eligibility threshold. To account for the fact that my running variable is discrete and contains several observations per value, I conduct a number of additional robustness checks (see Online Appendix E.3). First, I follow the recommendations of Lee and Card (2008), and analyze the effect of voter eligibility on participatory norms using clustered standard errors at level of the running variable. The point estimates in this analysis are slightly larger and more precisely estimated in these specifications.9 Second, I conduct a local randomization analysis (Cattaneo et al., 2015), and compare the difference in means between individuals who became eligible to vote on the election day and those who became eligible to vote one day after the election. I also compare individuals with a distance of five and ten days respectively from the voter eligibility threshold. The point estimates are similar to the one in the baseline specification. However, these tests include few observations, and the estimates are not statistically significant. Finally, I also collapse the data at the daily level and conduct my main analysis on this sample. These results also remain similar to the baseline estimates, although not significant at conventional levels. In addition to the above tests, I create so-called placebo thresholds at false values of the forcing variable. I check if a positive effect on support for extensive forms of political participation occurs at the date of election for previous and subsequent years (when elections did not take place). I find no effects at the placebo thresholds (see Online Appendix E.4). Taken together, the robustness checks in this section provide strong support for the identifying assumption.
different social backgrounds and different civic education curricula, I have observed that the effect is larger among those with lower educated parents and those who have not learned to care about politics in school, i.e. among individuals with more opportunity for political ownership and responsibility to grow.10 Further, I have found that the positive effect of voter eligibility is absent in non-free countries, where eligibility is less likely to generate a feeling of political ownership. These results, which can be found in Online Appendix F, weakly indicate that feelings of political ownership and responsibility could be relevant for understanding the relationship between voter-related experiences and normative support for civic engagement. However, future research should investigate the validity of this proposed mechanism. The finding of a positive effect of voter eligibility on support for extensive forms of political participation raises the question of why voter eligibility does not also spur the idea that a good citizen engages in electoral forms of political participation. While I cannot provide clear answers, there are many plausible explanations. If I adhere to the educative theory of political participation, a possible explanation could be that voting and party activity are perceived as a commitment to social order rather than an engagement in community life (cf. Dalton, 2008). Attitudes towards these forms of political participation could therefore remain unaffected by increased feelings of societal coownership. Alternatively, a positive effect of voter eligibility on support for electoral participation does occur but fades away because of an anticlimactic experience of voting, or disappointment with the result or aftermath of the election. The results of this paper also prompt questions about the persistence and behavioral implications of the findings, which should be explored in future research. The consistency between citizenship norms and individual actions has not been thoroughly studied (Dalton, 2008). Previous research has suggested that most people are involved in politics in ways consistent with their citizenship perspective (Theiss-Morse, 1993). However, it has also been argued that other factors could have a more significant effect on political actions. For example, research on voting persistence has found that individuals with a strong habit to vote maintain this habit without necessarily holding appropriate motivations, goals, and emotional states (Aldrich et al., 2011). Even if the relationship between individuals’ motivations and political behavior would be weak, the findings in this paper challenge a key methodological assumption in previous research on voting as habit-forming: that the only channel through which past eligibility affects subsequent turnout is through past participation (Gerber et al., 2003). By finding an effect of voter eligibility on citizenship norms, this paper provides support to the assumption that any instrumental variable approach to estimating persistence in voting is likely to capture the joint effect of past voting and other experiences that voter eligibility might entail (Meredith, 2009). The findings in this paper are also interesting from a participatory democratic point of view. Regardless of the behavioral implications, the findings should increase perceptions of the importance of elections for those who cherish deep civic engagement. Elections might not be sufficient in a society that aims at profound citizen involvement, but they can contribute to that goal by fostering desirable norms of citizenship among those who are eligible to vote. Thus, the recent search for new institutions that could facilitate in-depth political participation should be accompanied by efforts to safeguard and emphasize voter eligibility and the activities it brings about.
6. Concluding remarks In this paper I have examined whether voter eligibility increases support for the idea that a good citizen engages in electoral forms of political participation (H1) and/or the notion that a good citizen engages in extensive forms of political participation (H2). Comparing individuals who were eligible to vote up to one year before and one year after the election day, I have found support for H2, but not H1. According to the theory that motivates this study, the power exercised through the right to and experience of political participation produces feelings of political co-ownership and responsibility. These feelings are in turn likely to strengthen the norm that a good citizen is politically engaged. I am not able to test whether an increased sense of co-ownership and responsibility is the causal mechanism in the positive relationship between voter eligibility and support for extensive forms of political participation. However, in comparisons of individuals with
Appendix A. Supplementary data Supplementary material related to this article can be found online at https://doi.org/10.1016/j.electstud.2019.102090.
9 Note that the change in point estimates is due to the fact that I am using bias correction suggested by Calonico et al. (2014, 2017).
10
6
However, these differences are not statistically significant.
Electoral Studies xxx (xxxx) xxx
Z. Khoban
References
Geys, B., 2006. Explaining voter turnout: A review of aggregate-level research. Elect. Stud. 25 (4), 637–663. Goodin, R.E., 2008. Innovating Democracy: Democratic Theory and Practice After the Deliberative Turn. Oxford University Press, Oxford. Green, D.P., Shachar, R., 2000. Habit formation and political behaviour: Evidence of consuetude in voter turnout. Br. J. Political Sci. 30 (4), 561–573. Holbein, J., Rangel, M., 2016. Does Voting Have Upstream and Downstream Consequences? Evidence from Compulsory Voting in Brazil, Working Paper. Kaufman, A.S., 1960. Human nature and participatory democracy. In: Freidrich, C.J. (Ed.), Responsibility, NOMOS III. Liberal Arts Press, New York, pp. 266–289. Lee, D.S., Card, D., 2008. Regression discontinuity inference with specification error. J. Econometrics 142 (2), 655–674. Loewen, P.J., Milner, H., Hicks, B.M., 2008. Does compulsory voting lead to more informed and engaged citizens? An experimental test. Canad J. Political Sci. 41 (3), 655–672. Mansbridge, J., 1995. Does participation make better citizens? Good Soc. 5 (2), 1–7. Mansbridge, J., Bohman, J., Chambers, S., Christiano, T., Fung, A., Parkinson, J., Thompson, D.F., Warren, M.E., 2012. A systemic approach to deliberative democracy. In: Deliberative Systems: Deliberative Democracy at the Large Scale. Cambridge University Press Cambridge, pp. 1–26. Meredith, M., 2009. Persistence in political participation. Q. J. Political Sci. 4 (3), 187–209. Mill, J.S., 1862. Considerations on Representative Government. Harper and Brothers, New York. Mullainathan, S., Washington, E., 2009. Sticking with your vote: Cognitive dissonance and political attitudes. Am. Econ. J. Appl. Econ. 1 (1), 86–111. Newton, K., 2012. Curing the democratic malaise with democratic innovations. In: Geissel, B., Newton, K. (Eds.), Evaluating Democratic Innovations: Curing the Democratic Malaise? Routledge, New York, pp. 3–20. Nyhan, B., Skovron, C., Titiunik, R., 2017. Differential registration bias in voter file data: A sensitivity analysis approach. Am. J. Political Sci. 61 (3), 744–760. Pateman, C., 1970. Participation and Democratic Theory. Cambridge University Press, New York and Cambridge. Plutzer, E., 2002. Becoming a habitual voter: Inertia, resources, and growth in Young adulthood. Am. Political Sci. Rev. 96 (1), 41–56. Regan, D.T., Kilduff, M., 1988. Optimism about elections: Dissonance reduction at the ballot box. Political Psychol. 9 (1), 101–107. Shineman, V.A., 2018. If you mobilize them, they will become informed: Experimental evidence that information acquisition is endogenous to costs and incentives to participate. Br. J. Political Sci. 48 (1), 189–211. Smith, G., 2009. Democratic Innovations: Designing Institutions for Citizen Participation. Cambridge University Press, New York and Cambridge. Theiss-Morse, E., 1993. Conceptualizations of good citizenship and political participation. Political Behav. 15 (4), 355–380. Wolfinger, R.E., Rosenstone, S.J., 1980. Who Votes? Yale University Press, New Haven and London.
Aldrich, J.H., Montgomery, J.M., Wood, W., 2011. Turnout as a habit. Political Behav. 33 (4), 535–563. Beasley, R.K., Joslyn, M.R., 2001. Cognitive dissonance and post-decision attitude change in six presidential elections. Political Psychol. 22 (3), 521–540. Blais, A., 2000. To Vote or Not to Vote? The Merits and Limits of Rational Choice Theory. University of Pittsburgh Press, Pittsburgh. Blais, A., 2006. What affects voter turnout? Annu. Rev. Political Sci. 9, 111–125. Böker, M., 2017. Justification, critique and deliberative legitimacy: The limits of mini-publics. Contemp. Political Theory 16 (1), 19–40. Braconnier, C., Dormagen, J.-Y., Pons, V., 2017. Voter registration costs and disenfranchisement: Experimental evidence from France. Am. Political Sci. Rev. 111 (3), 584–604. Brady, H.E., Verba, S., Schlozman, K.L., 1995. Beyond SES: A resource model of political participation. Am. Political Sci. Rev. 89 (2), 271–294. Calonico, S., Cattaneo, M.D., Farrell, M.H., Titiunik, R., 2017. Rdrobust: Software for regression-discontinuity designs. Stata J. 17 (2), 372–404. Calonico, S., Cattaneo, M.D., Titiunik, R., 2014. Robust data-driven inference in the regression-discontinuity design. Stata J. 14 (4), 909–946. Cattaneo, M.D., Frandsen, B.R., Titiunik, R., 2015. Randomization inference in the regression discontinuity design: An application to party advantages in the US senate. J. Causal Inference 3 (1), 1–24. Coppock, A., Green, D.P., 2016. Is voting habit forming? New evidence from experiments and regression discontinuities. Am. J. Political Sci. 60 (4), 1044–1062. Dalton, R.J., 2008. Citizenship norms and the expansion of political participation. Political Stud. 56 (1), 76–98. de Tocqueville, A., 1969. Democracy in America. Anchor Books, Garden City, NY, Translated by George Lawerence, ed by Jacob Peter Mayer. Denny, K., Doyle, O., 2009. Does voting history matter? Analysing persistence in turnout. Am. J. Political Sci. 53 (1), 17–35. Dinas, E., 2012. The formation of voting habits. J. Elections Publ. Opin. Parties 22 (4), 431–456. Dunning, T., 2012. Natural Experiments in the Social Sciences: A Design-Based Approach. Cambridge University Press, New York and Cambridge. Ehrlich, T., 2000. Civic Responsibility and Higher Education. Oryx Press, Westport, CT. Finkel, S.E., 1985. Reciprocal effects of participation and political efficacy: A panel analysis. Am. J. Political Sci. 29 (4), 891–913. Finkel, S.E., 1987. The effects of participation on political efficacy and political support: Evidence from a west German panel. J. Politics 49 (2), 441–464. Fujiwara, T., Meng, K., Vogl, T., 2016. Habit formation in voting: Evidence from rainy elections. Am. Econ. J. Appl. Econ. 8 (4), 160–188. Fung, A., Wright, E.O., 2003. Deepening Democracy: Institutional Innovations in Empowered Participatory Governance. London, Verso. Gerber, A.S., Green, D.P., Shachar, R., 2003. Voting may be habit-forming: Evidence from a randomized field experiment. Am. J. Political Sci. 47 (3), 540–550.
7