Department of Methods in the Social Sciences, University of Vienna, Rooseveltplatz 2/4, 1090 Vienna, Austria
Markus Wagner: ta.ca.eivinu@rengaw.sukram; David Johann: ta.ca.eivinu@nnahoj.divad; Sylvia Kritzinger: ta.ca.eivinu@regniztirk.aivlys
∗Corresponding author. Tel.: +43 1 4277 49902; fax: +43 1 4277 9499. ta.ca.eivinu@regniztirk.aivlys
1Tel.: +43 1 4277 49912; fax: +43 1 4277 9499.
2Tel.: +43 1 4277 49910; fax: +43 1 4277 9499.
Received 2011 Feb 24; Revised 2012 Jan 19; Accepted 2012 Jan 20.
Copyright © 2012 Elsevier Ltd.
Open Access under CC BY-NC-ND 3.0 license
Abstract
Critics of giving citizens under 18 the right to vote argue that such teenagers lack the ability and motivation to participate effectively in elections. If this argument is true, lowering the voting age would have negative consequences for the quality of democracy. We test the argument using survey data from Austria, the only European country with a voting age of 16 in nation-wide elections. While the turnout levels of young people under 18 are relatively low, their failure to vote cannot be explained by a lower ability or motivation to participate. In addition, the quality of these citizens' choices is similar to that of older voters, so they do cast votes in ways that enable their interests to be represented equally well. These results are encouraging for supporters of a lower voting age.
Keywords: Input legitimacy, Political participation, Teenage vote, Turnout, Voting age
Highlights
► Citizens under 18 may lack the motivation and ability to participate in elections. ► We examine the political motivation and ability and their impact on turnout and vote choice quality for citizens under 18. ► We use a survey from Austria, the only country with a voting age of 16. ► Their reasons for not voting are not based on a lack of political motivation and political ability. ► Their quality of vote choice is no lower than among older voter cohorts.
1. Introduction
The level of turnout at elections is often seen as an indicator of the health of a democracy [], yet there is a general trend towards declining rates of electoral participation in Western Europe [e.g. ; ; ]. This has led to fears that democratic legitimacy may decline as elections increasingly fail to act as the ‘institutional connection’ [] between citizens and the state.
In light of these developments, it has been suggested that the minimum voting age should be lowered to 16 [e.g. ; ; ]. Supporters of such a reform argue that lowering the voting age would have a positive impact on electoral participation. This is because young people under 18 are likely to still be in school and live with their families, two factors that have been shown to encourage turnout through a variety of socialisation mechanisms [; ; ]. In the long term, this higher level of participation at a young age may then facilitate the early development of a habit of voting [e.g. ; ]. Of course, lowering the voting age is not only justified as a way to stop the decline in turnout. For example, it is also seen as a way to ensure that the interests of young citizens are represented in the political system [].
However, the proposed reform is not without its critics. The main argument made against lowering the voting age is that young people under 18 lack the ability and motivation to participate effectively in the electoral process []. It is suggested that this will lead to low turnout rates, comparable to – if not even lower than – those observed among citizens aged 18–25 []. A further consequence would be that citizens under 18 might not make use of their vote as effectively as older voters. While they might vote for the sake of voting, they would not challenge the government to respond to their interests. Thus, their vote choice would be driven more strongly by expressive instead of instrumental considerations [], and their policy views would not be well-represented by political actors.
In this paper, we test whether these critics are right. Are young people under 18 less able and motivated to participate effectively in politics? And do these factors influence whether and how they use their right to vote? If the answer to these questions is yes, then lowering the voting age could indeed have negative consequences for the health of democracy. If the answer is no, then critics are arguably left with fewer arguments why we should oppose lowering the voting age. Instead, we might consider potential positive consequences of the reform, such as tying young people to the democratic process, encouraging the development of a habit of voting and ensuring the representation of their interests.
We examine the choices made by young people under 18 using data from Austria, where in 2007 the voting age at national elections was lowered to 16. Specifically, we use a survey carried out in the run-up to the European Parliament [EP] elections 2009 which over-sampled young people under 26. Austria's reform allows us to examine for the first time whether the critics of lowering the minimum voting age are right. Before, the only possible empirical strategies were either to extrapolate about the behaviour of citizens under 18 from that of voters just over 18 or to study the potential electoral behaviour of young people under 18 in a context where they did not have the vote.
Our survey indicates that the intention to turn out was indeed relatively low among citizens under 18 in the 2009 EP election. Using the self-assessed likelihood of voting on a scale of 0–10, under-18s have a low average intention of turning out, with a mean score of 5.91. This is lower than among respondents aged between 18 and 21 [6.24] and between those aged between 22 and 25 [6.98], while respondents over 30 have a mean score of 7.38.
Is this pattern due to the fact that Austrians under 18 are particularly unable or unwilling to participate in politics? Our findings show that this is not the case. First, measures of political interest, knowledge and non-electoral participation indicate that young people under 18 are not particularly unable or unwilling to participate in political life. Second, these factors do not help to explain their lower turnout rates, so we cannot say that young citizens fail to vote for reasons that are particularly troubling for democratic legitimacy. Finally, there is no evidence that the quality of vote choices among citizens under 18 is any worse than that of older voters.
We begin this paper by discussing in greater depth existing arguments regarding the political behaviour of citizens under 18 and the potential effects of lowering the voting age in terms of democratic legitimacy, focussing on turnout and the quality of vote choice. After describing the survey, we provide a brief descriptive account of young people's motivation and ability to engage in politics. We then turn to a multivariate analysis that explores the reasons behind turnout decisions of citizens under 18. Finally, we examine the quality of vote choice among these voters.
2. Citizens under 18 and the democratic process
In the scholarly debate democratic legitimacy includes two dimensions: input and output legitimacy []. This paper focuses on the input dimension of democratic legitimacy. Input legitimacy refers to the idea that “[p]olitical choices are legitimate if they reflect the ‘will of the people’ – that is, if they can be derived from the authentic preferences of the members of a community” [: 6]. Input legitimacy requires citizens who are motivated and competent and who engage in reasoned arguments in collective decision-making processes. As a result, input legitimacy may be negatively affected by lowering the voting age if this only serves to extend suffrage to citizens who are not motivated or able to participate in decision-making in this way. Simply put, the central question is whether citizens under 18 have the ability and motivation to participate effectively in elections.
Why might we expect this not to be the case? argue that young people under 18 are simply not politically ‘mature’ enough to take part in the electoral process, and they define this ‘maturity’ precisely as the ability and motivation to participate. They measure the ‘political maturity’ of young people under 18 using political interest, party identification, political knowledge and attitudinal consistency. According to , those under 18 fail to score high enough on any of these indicators. They suggest that these differences cannot be explained by the fact that in the UK those under 18 do not yet have the vote and therefore have no incentive to become involved in politics. Instead, citing , they argue that the teenage brain may simply not be ready to vote at 16. However, point out that so far no neurological evidence has been put forward to prove this point, while show that teenage citizens possess the same cognitive sophistication as young adults. It is perhaps more likely that these age differences may exist due to a universal life-cycle effect, with younger voters simply not yet having developed the political interest, knowledge and sense of duty that comes with age [].
Thus, from this critical perspective young citizens under 18 lack the ability and motivation to engage effectively in politics. Since our aim is to test the arguments made by critics of lowering the voting age, our hypotheses are as follows:
H1a:
Young citizens under 18 are less able to participate in politics effectively than older voters.
H1b:
Young citizens under 18 are less motivated to participate in politics effectively than older voters.
2.1. Citizens under 18 and turnout
Enlarging suffrage to include young people under 18 may have consequences for the level of turnout. On the one hand, some scholars argue that turnout numbers may improve, especially in the longer term, as young people under 18 are more easily and more lastingly mobilised to vote due to socialisation effects [e.g. ]. On the other hand, critics put forward the argument that it could also be that young people under 18 simply mirror the low levels of turnout found among those aged between 18 and 21 [e.g. ].
However in this paper, we are not concerned with the levels of turnout themselves. For one, to examine the development of a habit of voting requires a longer-term perspective than cannot be achieved just two years after the voting age was lowered. Moreover, looking exclusively at the level of turnout should not be the only way to address whether declining electoral participation is worrying. As pointed out, it is particularly concerning when decisions not to vote are a reflection of disenchantment, indifference or a lack of capabilities [].
Lower levels of turnout among citizens under 18 do not automatically indicate that this pattern is due to a lower ability and motivation to participate. Other reasons may underlie this decision. First, young voters may privilege new modes of political participation over traditional forms of electoral participation [], ‘bypassing the electoral routes’ [: 165]. Electoral participation is not the only way that a democratic bond between citizens and the political system can be created [e.g. ; ; ; ]. Young voters may be particularly likely to choose other forms of participation due to longer schooling years, exposure to other forms of informal civic education, higher information levels, new information channels and a decrease in party affiliation [e.g. ]. Second, young voters may simply see voting itself as less of a civic duty [e.g. ; ; ]. They may have a more individual calculus of the utility of voting and rely more heavily on the assessment of the importance of election outcomes []. Thus, analysing only turnout rates per se is not enough to provide a good picture of the status of input legitimacy as we also need to take the underlying motives into account. In other words, we need to know whether citizens under 18 fail to vote because of a lower ability and motivation to participate effectively. If this is the case, then this undermines input legitimacy; if not, then lower turnout is perhaps less worrying.
In sum, we argue that the quality of the electoral participation of citizens under 18 is particularly unsatisfactory if low turnout can be explained by a low willingness and motivation to engage in politics. We will therefore test the following two hypotheses:
H2a:
The lower turnout of young people under 18 can be explained by their lower ability to participate in politics.
H2b:
The lower turnout of young people under 18 can be explained by their lower motivation to participate in politics.
2.2. Citizens under 18 and the quality of vote choice
Just because citizens go to the polls does not mean that they will be well-represented by those they elect. As Lau et al. argue: “[V]otes freely given are meaningless unless they accurately reflect a citizen's true preferences” [: 396]. Citizens should be able to select accurately between political actors and make a choice that is consistent with their own views, attitudes and preferences [e.g. ]. If voters under 18 take choices that do not reflect their interests and attitudes, then this will limit their substantial representation []. The arguments presented earlier that citizens under 18 may lack the requisite ability and motivation to participate [] would also lead them to be less inclined to think carefully about their decision and therefore choose parties that do not reflect their preferences. They may fail to take choices that represent their interests well. Thus, there would also be negative consequences for democracy if the choices made by voters under 18 are less well-linked to their actual preferences than those of older voters. On the other hand, if the decisions of voters under 18 reflect their preferences as well as they do in older age groups, then the critics' arguments have no empirical basis. We would have no reason to believe that the interests and preferences of voters under 18 would be less well-represented.
Our final hypothesis therefore tests this last argument by critics of lowering the voting age and reads as follows:
H3:
The quality of vote choice among voters under 18 is lower than among older voters.
3. Data and methods
Until now, empirical research on the effects of lowering the voting age has had to take one of two unsatisfactory approaches. The first method has been to assume that under-18s are little different from those just over 18, justifying the use of evidence from the voting behaviour of young citizens aged 18 and older [e.g. ]. The second approach uses data on citizens under 18 before they have the right to vote [e.g. ]. Studying electoral participation for those who do not have the right to vote has a considerable flaw: without the right to cast a ballot, there is no rational incentive for citizens to increase their interest and knowledge in politics. Simply having voting rights may encourage people to gather information and become politically active in other ways [; ]. To test correctly whether the electoral participation of under-18s matches the quality of that of their older peers, we therefore need a case where such young citizens have the right to vote.
Austria is the only country in Europe that has a voting age of 16 for national elections. The reform was passed by the Austrian parliament in 2007, and since then, young people under 18 have cast ballots at a series of elections, including for the national parliament in 2008, the European Parliament in 2009 and the presidential elections in 2010. Austria thus provides the first opportunity to examine the political participation of under-18s in a nation-wide election, at least in a stable advanced industrial democracy. The specific data used in this paper are from a pre-election survey [n = 805] conducted at the end of May and the beginning of June 2009, so in the weeks directly before the European Parliament election []. Voters between 16 and 25 were over-sampled for this survey [n = 263], making this dataset particularly suitable to our research questions. We take advantage of the over-sampled segment of Austrian voters to compare 16- and 17-year olds to voters between 18 and 21, 22 and 25, 26 and 30 and to voters over 31.
We assess the ability and motivation to participate effectively in politics using three measures. The ability to engage in politics is evaluated using political knowledge, which we measure by assessing whether respondents correctly place the Social Democrats [SPÖ] to the left of the two far-right parties [FPÖ and BZÖ] and the People's Party [ÖVP]. We measure the motivation to participate effectively in politics using political interest and the willingness to consider various forms of non-electoral participation. The respondents' interest in politics is measured as the average of answers to eight questions tapping attention to politics in general and the EP campaign in particular. The variable was rescaled to range from 0 to 1, and the alpha reliability coefficient of this scale is 0.81. We measured non-electoral political participation by asking respondents to rate on a four-point scale their hypothetical willingness to engage in a series of political activities: contacting a politician, collecting signatures, working for a non-governmental organization, taking part in a legal demonstration and working on a campaign. We also create an overall index for non-electoral political participation using the average answer to the five questions. The scale ranges from 0 to 1 with an alpha reliability coefficient of 0.75.
It is always difficult to measure turnout using survey questions due to the problems of over-reporting, sample selectivity, social desirability bias and the stimulus effects of pre-elections interviews [e.g. ; ; ]. There is evidence that the pre-election turnout intention questions are the best available predictor of whether a person is likely to vote []. Respondents might be more honest regarding their actual intention to turn out when presented with a scale in which people can indicate uncertainty and reluctance without declaring directly that they might abstain. Therefore, we use turnout intention as our dependent variable. We measure propensity to turn out with a question asking respondents to state their certainty of voting in the upcoming EP election on a scale of 0–10. In our sample, 54.1% of respondents gave a vote intention score of 8 or higher and 41% a score of 9 or higher. This compares favourably to the 46% who actually voted on 7 June 2009.
Examining the intention to turn out in an EP election gives us also the advantage of studying an election with lower overall turnout; this could reduce the social desirability bias as people might be less reluctant to declare that they will not vote when abstention is a more common phenomenon. We take into account the specific EU nature of the election by including EU-specific versions of core variables in our regression models and by including a control variable concerning views on European integration.
4. Results
We present our results in three steps. First, we present descriptive findings on the ability and motivation to participate in politics among young people under 18. Next, we examine the causes underlying turnout decisions before finally examining the quality of vote choice.
4.1. The ability and motivation to participate effectively
Critics of lowering the voting age argue that citizens under 18 have a lower motivation and ability to engage in politics than older citizens. We test this by considering three measures widely used in the literature to capture these constructs [e.g. ]: interest, knowledge and non-electoral political participation.
Fig. 1 presents the age group means for political interest and knowledge. We see, first, that interest in politics is by no means particularly low among under-18s; indeed, it is the second-highest average of the four age groups under 30. However, in spite of their apparent interest in politics, political knowledge is somewhat lower among under-18s compared to the other three groups of young voters. However, it is worth noting that this difference is significant in a two-tailed t-test only for the comparison with 22- to 25-year-olds. Moreover, a cautious interpretation of these results is required since we only have one knowledge question. Nevertheless, there is some indication that political knowledge might be lower among under-18s. This may be due to the fact that young citizens do not yet have the experience necessary to place parties correctly on a left–right scale. There is thus some support for H1a, i.e. that citizens under 18 are less able to participate in politics.
Open in a separate window
Fig. 1
Interest and knowledge, age group differences. Note: mean values by age group shown; bars indicate 95% confidence intervals around the mean; dashed line indicates overall mean; see for question details.
Citizens can also engage in politics outside of elections. In Fig. 2, we present the average scores for each age group across all five activities we asked about, with the top-left panel presenting the results for the overall index. It is clear that the youngest citizens' willingness to participate in non-electoral politics is relatively high and no different from the overall mean. Young citizens are particularly likely to say that they would take part in a demonstration. Overall, citizens under 18 are just as motivated to take part in political life as older age groups. This preliminary result indicates that there is little evidence in favour of H1b.
Open in a separate window
Fig. 2
Non-electoral political participation, age group differences. Note: mean values by age group shown; bars indicate 95% confidence intervals around the mean; dashed line indicates overall mean; see for question details.
Finally, we further illustrate the attitudes of young people under 18 using measures of democratic disaffection and alienation, specifically the three indicators institutional trust, satisfaction with democracy and the perceived impact of politics on one's life [Fig. 3]. These are related to measures of the motivation to engage effectively in politics. Surprisingly, our data show that trust in institutions among citizens under 18 is significantly higher than the overall mean among all citizens, so there is no indication at all of disaffection using this classical measure. In addition, satisfaction with national and European democracy among citizens under 18 is actually higher [and often significantly so] than among older citizens. Turning to the bottom-right panel, younger citizens do not differ from older citizens in the proportion who say that either the national parliament or the EP has a ‘strong’ impact on them personally. Given this evidence and their satisfactory level of political interest, there is strong descriptive evidence that the youngest citizens in Austria are not particular ‘turned off’ by electoral politics, democracy and political institutions in general – results which are quite contrary to those reported in the literature [e.g. ]. There is therefore so far little evidence in favour of H1b, so that there is little indication that citizens under 18 are less motivated to participate effectively in politics.
Open in a separate window
Fig. 3
Alienation, indifference and impact of parliaments, age group differences. Note: mean values by age group shown; bars indicate 95% confidence intervals around the mean; dashed line indicates overall mean; see for question details.
In sum, citizens under 18 do not differ in terms of their democratic disaffection and their motivation to participate in politics from older age groups. However, they do have relatively little knowledge, indicating that they might be less able to participate. In the next section, we will examine whether age group differences in the motivation and ability to participate effectively in politics can help explain lower turnout levels among citizens under 18.
4.2. Turnout motivations of citizens under 18
We now turn to examining the causes of lower turnout by following the approach of and . The starting point is a very basic regression model with dummy variables for the four age groups under 31 as well as a series of fundamental socio-demographic controls. In the following steps, a series of independent variables are added to the model. This allows us to test whether different age groups decide not to vote because they lack the ability or motivation to participate effectively in politics; if this is the case for the youngest cohort, the age coefficient for the under-18s should decrease [or even cease to be significant] when the relevant variables are added to the model. Importantly, the cases included in each model are the same, so the most basic model only includes respondents for which all variables in the fullest model are available. Due to a heteroscedastic error distribution robust standard errors are used.
We present six OLS regression models predicting voter turnout [Table 1]. Model 1 includes four age dummy variables for those aged 16 and 17, 18–21, 22–25 and 26–30. The reference group are citizens aged 31 and older. The model also includes four socio-demographic controls: education, gender, rural residence and migration background. Previous research indicates that education is an important predictor of the propensity to turn out in almost all democracies [e.g. ; ; ; ]. We also include migration background as this may be related to the level of resources and to the level of connectedness one feels to the political system; we therefore expect having a migration background to have a negative effect on turnout []. Living in a rural area may increase turnout as social capital and thus the pressure to vote may be higher there [e.g. ; ].
Table 1
OLS regression results.
VariablesModel 1Model 2Model 3Model 4Model 5Model 6Basic modelBasic model + controlsKnowledgeParticipationInterestFull modelAge [ref: 31+] 16–17−1.005 [0.517]−1.244** [0.458]−1.210** [0.467]−1.272** [0.438]−1.09* [0.424]−1.100** [0.419] 18–21−1.562*** [0.361]−1.138** [0.375]−1.163** [0.366]−1.201*** [0.353]−0.429 [0.347]−0.600 [0.332] 22–25−0.690* [0.329]−0.708* [0.306]−0.771* [0.306]−0.936** [0.302]−0.295 [0.294]−0.576* [0.293] 26–30−1.040** [0.354]−1.118*** [0.330]−1.109*** [0.328]−1.040*** [0.306]−0.547 [0.293]−0.565* [0.284]Political knowledge0.703* [0.277]0.672** [0.253]Non-electoral participation3.396*** [0.499]2.423*** [0.511]Political interest6.815*** [0.689]5.841*** [0.709]Education1.332*** [0.250]1.013*** [0.251]1.012*** [0.249]0.659* [0.258]0.749** [0.242]0.533* [0.245]Female0.075 [0.228]0.063 [0.216]0.089 [0.216]0.007 [0.208]0.154 [0.202]0.126 [0.197]Rural residence0.156 [0.229]0.195 [0.222]0.179 [0.221]0.233 [0.213]0.254 [0.208]0.257 [0.202]Migration background−0.392 [0.306]−0.513 [0.289]−0.521 [0.293]−0.431 [0.285]−0.432 [0.270]−0.393 [0.275]EU attitude−0.0002 [0.474]0.063 [0.474]−0.496 [0.461]−0.135 [0.429]−0.409 [0.427]Trust in institutions3.043*** [0.773]2.964*** [0.776]2.718*** [0.755]1.559* [0.748]1.464* [0.738]Dem. satisfaction: National0.135 [0.281]0.126 [0.282]0.255 [0.276]0.103 [0.264]0.184 [0.263]Dem. satisfaction: EU0.311 [0.259]0.302 [0.260]0.224 [0.258]0.479 [0.247]0.385 [0.246]Impact of national parliament1.246*** [0.265]1.163*** [0.263]1.011*** [0.254]0.380 [0.253]0.256 [0.251]Impact of European Parliament0.998** [0.309]1.024** [0.314]0.851** [0.296]0.419 [0.289]0.422 [0.286]Constant7.217*** [0.245]3.960*** [0.529]3.522*** [0.553]3.479*** [0.504]3.044*** [0.498]2.413*** [0.496]R20.0550.160.1690.2190.2730.308N699699699699699699
Open in a separate window
*p