Political congruence between adolescence and their parents: evidence from a quasi-experimental local elections in the city of Ghent (Belgium)

Abstract

Lowering the voting age is often criticized due to the alleged political immaturity of young voters, which is considered as negative for the quality of democracy. We adopt a unique approach to this issue by measuring political congruence between adolescents and their parents to ascertain whether ideological views and salient positional issues, such as attitudes on immigration, differ between different age groups in adolescence. Using a household survey in the city of Ghent (Belgium), first we compare political ideologies and attitudes toward immigration for those of age between 15 and 19, and their parents. We find that adolescent children tend to resemble the political ideologies of their parents, and they also tend to have similar attitudes about immigration. Second, we find that ideological resemblance is even higher among adolescents that regularly talk about politics with their parents. Finally, we do not find evidence that political congruence with parents differs by the age of adolescent children or their right to vote. Therefore, we conclude that political ideology and positional views are formed in an early stage of adolescence, i.e., before the age of 15. Our findings have relevance for the debate about lowering the voting age, as it points to already clearly formed ideological views and political attitudes at early stages of adolescent life.

Introduction

Declining political participation manifests itself by a decline in partisanship and generally lowering turnout rates across countries and across various types of elections (Blais and Rubenson 2013; Gray and Caul 2000). This global phenomenon has worried scholars, with the fear that this trend will endanger democratic legitimacy more broadly defined (Fieldhouse et al. 2007). With, on average, fewer voters participating in elections, both policy makers and social science scholars have searched for possible solutions to increase electoral participation, especially among younger age groups, and one of the recurring suggestions in this debate is to lower the legal voting age (Zeglovits and Aichholzer 2014). However, the empirical evidence with regard to the consequences of lowering the voting age is rather mixed (Chan and Clayton 2006; Wagner et al. 2012; Bergh 2013; Eichhorn 2014). There are two major arguments within this debate. First, scholars have argued that the fact that adolescents will be allowed to vote for the first time at a lower age, will make it easier for them to develop a habit of turning out to vote. This assumption relies on the idea that citizens who have participated in their first elections when they were allowed to vote are more likely to continue to vote in future elections (Highton and Wolfinger 2001; Franklin 2004). The difference is that at the age of 16, adolescents are still likely to live with their parents and to attend a local high school. In turn, these two factors, the school environment and parents, are more likely to influence young first time voters to cast votes in elections, something that according to Franklin (2004) will lead to a lifelong voting habit. The influence of parents and the school environment for those at 18 or above, on the other hand, is more limited due to a higher probability of changing one’s living environment (i.e., starting university education, new employment, more likely to leave household of their parents).

Second, an early introduction to electoral participation is expected to increase voters’ interest in politics and to improve their political knowledge, which in turn will increase the quality of the vote (Stiers et al. 2020). High turnout rates and the quality of vote are crucial for democracies to function, as higher voter participation increases the legitimacy of elected governments. Also, the quality of vote is essential to ensure the connection between voters’ needs and the political decision that will be made.

Our study contributes to this debate about the consequences of electoral participation at a lower age, by ascertaining whether adolescents indeed have distinct and clearly formed views on political matters, i.e., whether they have reached political maturity.¹ Our study adds to the literature on early voting, which has paid less attention to the political maturity of adolescents. Using our unique dataset, we are able to investigate when ideological and positional views are being formed, and whether the emergence of ideological and positional views indeed takes a major step forward between 15 and 19 years old. We are able to further look into the political congruence between adolescents and their parents, asserting possible distinct generational political views or a high level of political agreement within families. This approach directly refers to the family and other socialization mechanisms, as tools to increase participation in elections among young voters (Franklin et al. 2004), as we are able to investigate the importance of family factors in forming one’s ideological views and political attitudes.

Second, we are able to compare findings among different age groups, within a quasi-experimental setting, where we examine whether ideological views and political attitudes differ between age groups among late adolescents. Our quasi-experimental approach allows to test the effect of electoral participation, since the sample of adolescents in this study is divided between those who have the right to participate in elections (above 18), those who are allowed to vote in the Ghent mock elections (16 and 17), and the baseline group (age 15) that was not allowed to vote in any election. The unique quasi-experimental data comes from the city of Ghent in Belgium, where a mock election was held with a lowered voting age for the local elections in 2018. We used this setting to administer a household survey to all adolescents in the city between the age of 15 (non-voting age), 16–17 (first time mock trial voting age), and 18–19 (regular voting age), and their parents.

We test the effects of the right to vote on political maturity among adolescents. In other words, we test whether the act of voting leads to more pronounced or different positional views on political issues among this group. Our study relies on congruence as a measure of similarity of political views between adolescents and their parents, in order to assess the political maturity of the adolescents. Since parents are assumed to be long-term political actors with some degree of political maturity, similar political outcomes (i.e., congruence) between adolescent children and their parents would indicate similar levels of political maturity. If, however, the distance in positional views on political matters is high among adolescents and their parents, then this might bring into question the maturity of political reasoning among the younger sample (although this could lead in both directions).²

This paper proceeds as follows: We first describe the literature on the topic of the lowering of the voting age and political maturity of young adolescence, before elaborating our theoretical framework and possible age group differences. Subsequently, we proceed with our data collection strategy in a quasi-experimental approach and the methods we use in our study. The result section elaborates our findings and points to the final conclusion as well as recommendations for further research in the topic.

Literature review

Family socialization as a factor of political maturity

Traditionally, it has been assumed that children are socialized mainly within the family. Although there is also an effect of the education system and the peer group, in general, it is safe to assume that parents still have a strong effect on the political attitudes and behaviors of their children (Iyengar et al. 2018). In some of the older literature, it was assumed that parents have a strong effect on the attitudes of their children, and that they actually try to influence these attitudes. In the current era, however, this perspective has largely been abandoned, and most authors would now see this intra-family socialization as a reciprocal process (McDevitt and Chaffee 2002). While parents will have an effect on the attitudes of their children, it is just as likely that children will influence the attitudes of their parents. Most studies indeed show there is a remarkable degree of congruence between the political attitudes of children and parents, and in the literature two important causal mechanisms can be distinguished.

Some authors emphasize the occurrence of a sorting mechanism that leads to the reproduction of social stratification (Burns et al. 1997). In general, parents and their children will share the same socio-economic background, as resources are shared within the household. Furthermore, children of highly educated parents on average will receive better education opportunities, so they can aspire to reach the same kind of economic status as their parents once they enter adulthood. This close resemblance with regard to the access to material resources and social status, will also imply that adolescents and parents develop the same set of ideological preferences.

A second potential causal mechanism is more closely related to the content of the ideological beliefs. Typically, parents and children talk about political issues, and this is seen as a prime mechanism for transmitting political attitudes and values from one generation to another (Hooghe and Boonen 2015). Indeed, in families where there is less discussion about political topics, children often do not even have reliable knowledge about the political and ideological positions of their parents, and therefore it can hardly be expected that they have the same ideological preferences as their parents. Nevertheless, in this reasoning too, it is clear that one should be cautious with regard to the direction of causality. Levinsen and Yndigegn (2015) have shown that when children have the perception (correctly or not), that there is a large ideological distance between themselves and their parents, they will tend to avoid having political discussions with their parents. Most likely they want to avoid a sharp disagreement within the family, and they therefore refrain from entering into this kind of discussion. The fact that we routinely observe a correlation between intra-family coherence on the one hand, and the intensity of political discussion on the other, therefore, should not be taken as an indication that we have established a real causal mechanism in this regard. What we do know, however, is that a high degree of politicization within the family will be associated with a closer resemblance between parents and children (Hatemi and Ojeda 2021). Discussions within the family are one possible (and important) indicator for this politicization of the family context, but other potential indicators are watching the television news together, or having a newspaper arriving in the household.

These two causal mechanisms therefore lead to different expectations. The first approach assumes that there will be a higher level of congruence between parents and their children when the family has access to various resources (income and education level), while the second approach assumes that the degree of politicization within the household is crucial in this regard. Our first line of analysis investigates these two possible mechanisms.

First, we look at the congruence between parents and their adolescent children, which is defined as the distance in political views. We hypothesis that

H1

Adolescent children resemble the political ideology and political attitudes of their parents.

Second, we investigate the two possible mechanism through which this relation is established.

H2a

Parent’s socio-economic status (education and income) increases ideology and political attitudes congruence with their adolescent children.

H2b

Talking about politics within the family increases ideology and political attitudes congruence between parents and their adolescent children.

The right to vote as a factor of political maturity

Political maturity as a general concept to assess the quality of the vote is the main reason why it is important to understand more about the intra-family congruence with regard to political attitudes. A recurring argument against lowering the voting age is based on the assumption that adolescents lack the required political maturity to develop their own political attitudes in an independent manner (Chan & Clayton 2006). The basic argument is that adolescents would be easily influenced, whether it is by their parents, their peers, or by popular “influencers” on the social media. If that would be the case, we should indeed be able to observe in our dataset that young adolescents (i.e., the age of 15) behave differently than later adolescents (i.e., the age of 19). If young voters develop their own, independent position, differences in these age groups would be an argument to have a rather late age to vote.

Although there is a consensus among scholars that increasing turnout rates and strengthening the quality of vote have a positive impact on democratic legitimacy, there is a lack of agreement about whether lowering the voting age to 16 will contribute to these aspirations. Within the literature, there are clearly opposing views on this issue. For example, Chan and Clayton (2006) use data from the United Kingdom to argue that adolescent citizens lack the political maturity to be effectively involved in elections. Eichhorn (2014), on the other hand, finds that electoral and non-electoral participation is higher among 16-year-old and 17-year-old Scottish voters, compared to their peers in the rest of the United Kingdom, who are not allowed to participate in elections. Similarly to Eichhorn (2014), Stiers et al. (2021) show that the right to vote increases attention to politics among young people in Belgium. Also, Wagner et al. (2012) argue that the quality of the vote among 16 years olds in Austria is similar to the quality of the vote of 18 years old, although they also note that turnout rates are slightly lower among the youngest age group. In contrast, Bergh (2013) using data on elections in Norwegian municipalities finds there is a significant gap in political maturity between 16 and 17 years old compared to older groups. It is clear from this short overview, therefore, that empirical results about the effects of lowering the voting age are mixed. Our second line of analysis therefore investigates when this political maturity is created, and if the right to vote is related to the political maturity of adolescent children. Thus we test:

H2a

The congruence between parents and adolescent children with regard to political ideology and political attitudes differs between those who are at legal age (18 and above) with full rights to vote, compared to minors (bellow 18).

H2b

The congruence between parents and adolescent children with regard to political ideology and political attitudes differs based on age of the adolescent.

Data and methods

Data collection

We collected data on the occasion of the city of Ghent mock elections, where for the first time, adolescents at the age of 16 and 17 were invited to vote by the city authorities. Following this unique opportunity, we organized a large-scale post-election survey, where all citizens between 15 and 19 years received a survey questionnaire, accompanied by an additional survey questionnaire directed to one of their parents. We received the dates of birth and addresses of all citizens of the city of Ghent between the ages 15 and 19, which included a total number of 11,016 adolescents (the whole population). Every adolescent received a survey and two personalized by first name letter reminders, followed by a final round of survey questionnaires aimed for those who did not responded to the previous requests. The first round of surveys, one envelope per household, was sent immediately following the local elections on October 14th, 2018.³ A reminder letter to those who have not responded was sent on November 12th, 2018, followed by a second reminder, which also included questionnaires, sent the first week of December. Each questionnaire had a unique number connected to an expected respondent from our database of all young people between age 15 and 19 residing in the city of Ghent. We received 3735 questionnaires in total, out of which 2360 were completed by adolescents (15–19 years old) and 1375 questionnaires were completed by parents. Our final response rate for adolescents is 21.6 percent. For the analysis of this paper, we only use those adolescents for which we have a match with a parent, so we can construct a unique dyad of a child and a parent.

By lowering the age group to 15, we were able to collect data from three different adolescent groups, those without any right to vote (age 15), those who were allowed to participate in mock elections (age 16 and 17), and those that have a regular right to vote (above 18), and the comparison between these two groups will be central to our analysis.

Variable operationalization

Our study has two main dependent variables: (1) respondents’ left–right ideological position and (2) respondents’ score on anti-immigrant attitudes, which are, in Belgium, as in numerous other European countries, highly salient political issues. Both these variables are considered to be important factors of political maturity. For example, those young voters who are considered to be politically mature would be expected to have coherent expression of their ideological positioning, with their own positional views (Bergh 2013, p. 93). Along these lines we would expect to observe variation both with attitudes (e.g., on immigration) or ideology by age, if political maturity happens at age 18 or when one is granted the chance to participate in elections. For measuring political ideology, participants were asked to place themselves on a 11-point scale (ranching from 0 to 10) following a standard ideology self-assessment question:

“The terms ‘left’ and ‘right’ are often used in politics. Where would you place yourself on this scale, where 0 is left 5 is center and 10 is right.”

Since our subject of interest is the congruence between children and parents, we modeled political ideology in two distinct ways. First, we used the self-placement of adolescents on this scale as our dependent variable, and used the self-placement of the parent of that adolescent as an independent variable. Our second approach is to calculate the congruence of adolescents with their parent’s political ideology, by using the distance between the position of the adolesent and the position of the parent. This variable is based on the absolute difference between these two positions from the two members of the same dyad.

Our second line of investigation concerns attitudes toward immigration, as a salient issue in Belgian politics (Hooghe and De Vroome 2015). For this analysis, we used the same strategy as with political ideology, by first including adolescents’ attitudes toward immigration as our dependent variable, and the attitude of their parent as our independent variable. Second, we calculated the distance between the adolescent’s attitude about immigration, and that of the parent.

Our attitudes toward immigration variable is a composite score of three questions that measure immigration attitudes. Respondents were asked to place themselves on a 5-point Likert scale, following three statements about immigration. They are as follows:

• 1. “Immigrants are generally good for Belgian economy”

• 2. “The culture in our country is threatened by immigrants”

• 3. “Immigrants cause more crime.”

The scales were than matched to follow the same overall meaning, where 1 expresses a negative attitude toward immigration and 5 expresses a positive attitude toward immigration. The three questions load well on one factor with Cronbach’s alpha 0.77 for young adolescents and 0.80 for parents, forming a strongly coherent latent concept.

Previous research has shown that in families where there is a lot of political discussion going on, there is a stronger congruence of the attitudes of parents and children (Hooghe and Boonen 2015). We therefore also include a variable measuring how often adolescents talk about political topics with their parents. To construct this variable, respondents were asked to report on a 4-point scale (with 1 being never and 4 being often) how often do they talk with their parents about political and social issues.

We control for several factors that might influence the congruence between adolescents and their parents. With our controlling variables, we focus on the environment in which adolescents were raised, by looking at their parents’ education level, as well as the number of books in the parent house—another approximation of education and knowledge within the household (Torney-Purta 2002). Second, we use the financial situation of the parent, as an indicator for the financial situation of the family household. Research has shown that families with higher income provide more opportunities for their children to receive a higher level of education, which in return is associated with higher levels of political participation and political knowledge (Kallio et al. 2016).

Our second line of interest is to investigate whether political attitudes are being formed at age of 18 or earlier. This is relevant for the more general debate on when adolescents actually can be considered as full, mature citizens. To do so, we test whether the current voting age of 18 is a predictor of ideology formation. In the first model, we use age as a scale. This allows us to investigate the differences that are associated with one unit (i.e., one year) increase in age. Second, we group those respondents bellow 18 as our baseline group, and the respondents above 18 as our comparison group. Finally, we create three categories, with respondents at age of 15 as our baseline group, our second group consists of those of age of 16 and 17, and third we have those belonging to the regular voting age group 18 or above. This way we are able to calculate differences in voting behavior at different age stages, using several approaches, both by year and by group, to ascertain whether mock elections or the right to vote are associated with different political preferences.

Finally, we control for the question whether the parent adheres to a national (Belgian), or regional (i.e., Flemish) identity.⁴ Appendix 2 (Electronic supplementary material) provides summary statistics of all variables used in our models.

Methods

Since our dependent variables are continuous scales, we use multivariate linear regressions (OLS) in all our models 1–6, for both Tables 1 and 2.⁵ Our methodological approach is based on two sets of dependent variables. The first set of dependent variables measures ideology and positional attitudes of adolescents, explained by the ideology and positional attitudes of their parent (models 1–3 in Table 1, 2). The second approach focuses on the distance between adolescents and their parents concerning both ideology and positional attitudes (models 4–6 in Tables 1, 2). We use three distinct age groups to model our analysis, which allows us to investigate the first set of dependent variables. The same strategy with three specifications of age groups is used to investigate distances as our dependent variable.

Table 1 Ideological congruence between adolescents and their parent

Table 2 Congruence between adolescents and their parent regarding attitudes for immigration

Results

We divide the results in two parts. First, we focus on the analysis of ideological congruence between adolescents and their parents (Table 1), after which we present the analysis of their attitudes toward immigration (Table 2).

Ideological congruence (left–right)

Following Table 1, with an analysis on the left–right ideology congruence between adolescents and their parents, we consistently find evidence of a strong positive association between these two variables (see models 1 to 3, Table 1). Parents’ ideology is highly correlated with that of their children, with a unit increase in parents’ self-placement being associated with almost 0.57 units increase in the self-placement of the adolescent. In other words, if the parent is identifying herself toward the right of the ideology scale, the likelihood for her adolescent daughter/son to identify on the right side of the scale increases too. All models are significant at the 0.01 level. We conclude that ideological congruence between adolescents and their parents clearly is present in our sample confirming hypothesis 1.

Having established this strong correlation between the two ideology positions within the family dyad, we go a step further, by investigating the possible mechanism that could explain this congruence (model 4 to 6, Table 1). Here, we look at family socialization (H2a) and political communication within the family (H2b) as mechanisms to explain ideological congruence. Models 4 to 6 in Table 1 focus on education level, number of books in the house and family income as part of the socialization mechanism and the political communication mechanism, where adolescent respondents report how often they talk politics with their parents. Our expectation is that socio-economic background of the parents (education and family income) will have effect on ideology congruence, although we do not necessary hypothesis about the direction of this effect.⁶ Second, we expect that those who talk more often about politics with their parents are more likely to have similar ideological views, as these talks can help to identify more correctly the position of other family members.

Following these two hypothesis, we find no evidence that socialization factors have an effect on the ideological congruence. However, we find evidence that indeed those who talked more often about politics with their parents have more closely related ideological views with that of their parents. Across all our models (4–6, Table 1) this communication mechanism is significant at the 1% level (p < 0.01). We thus have a strong indication that one of the transmitting paths for ideology congruence lies in talking about politics within the family. In other words, the more political discussions within the family, the stronger the similarity in ideological views. For robustness, we also use an interaction term in models 1–3, Table 1 between the communication mechanism (talks politics with parents) and left–right ideological self-placement of parent. In line with the above, we find positive and significant interaction effect between the two variables, re-confirming our congruence analysis that those who talk politics with their parent tend to have similar political views (see Appendix 3 (Electronic supplementary material), Table 1).

What is worth noting, however, is that we do not expect talks about politics between adolescents and their parents to lead to a specific ideological preference. In other words, we expect that both sides, those on the left and those on the right side of the political spectrum, will use communication within the family as a mechanism to achieve closer congruence within the family. The non-significant effects in models 1 to 3 indeed suggest that this process occurs in a similar manner across the entire political spectrum. Second, although we expect talking about politics will be associated with a stronger resemblance between adolescents and parents, we still need to acknowledge that those models (4 to 6, Table 1) explain less than one percent of the variation in the dependent variable. Nevertheless, we expect that talking about politics matters, as do other factors in connecting the ideology of parents with that of their children. Interestingly enough, we do not find evidence that the financial situation of the family or the education level of the parents have a positive impact on ideological congruence. Following the ideology congruence we find support of hypothesis 2b, but not for hypothesis 2a.

Our second line of inquiry deals with the question when exactly these ideological views are being created. The reasoning behind this leads to the debate of lowering voting age and if the right to vote has a direct effect on the quality of the vote, where ideology and positional attitudes congruence between adolescents and their parents are our tests of the political maturity of adolescents. However, using a variety of comparing age groups and different specifications as our baseline line group, we consistently fail to find any difference among age groups concerning ideological congruence with that of their parents. We do not find significant differences with congruence between adolescents and their parents, having similar impact across different age groups, although we use several different specifications of age. All results lead to one single conclusion: no evidence is found in our sample that those who are of voting age (18 and above) or those who belong to the mock election group (16 and 17) differ ideologically from those who do not have any right to vote and are at the age of 15 (see models 1–3, Table 1). Similarly, we do not find evidence of different ideology congruence by age (if the respondent is of legal age or not), neither with ideological self-placement, or with ideological distance as our modeling approach. Second, the right to vote seems to have no effect on the ideological views or ideological congruence. Thus, we find no evidence in support for both hypothesis H2a and H2b.⁷

Congruence: attitudes toward immigration

Our second line of research investigates the political attitudes of adolescents with regard to immigration. Since immigration constitutes one of the most salient issues in Belgian politics, with political attitudes strongly polarized about this issue (Billiet and De Witte 1995), we test whether adolescents and their parents share similar views on this topic. Table 2 models 1 to 3 present the results using parents’ attitudes about immigration as independent variable, while models 4 to 6 measure distances.

With this line of analysis we find confirmation that congruence is important part of peoples’ political attitudes, with the attitudes of adolescents correlating quite strongly with those of their parents. More specifically, with one unit increase toward positive attitudes on immigration among parents, we find an increase of positive attitudes toward immigration among their children increased by 0.33 units. This means that one-third unit increase of adolescents’ attitudes toward immigration is associated with a unit increase in parents’ attitudes toward immigration, where parents having positive attitudes toward immigration are more likely to influence positive attitudes toward immigration among their children. We thus are confident that both political attitudes and ideologies of adolescents reflect those of their parents, confirming hypothesis 1.

The second line of our research investigates the mechanisms through which this attitudinal congruence is transmitted. Differently from ideology, we find that the communication mechanism, where adolescents talk about politics with their parent is less robust as compared to ideology.⁸ Also, we find that the financial situation within the family influences the congruence with regard to immigration attitudes to some extent. First, those that often talk about politics with their parents are more likely to have similar attitudes on immigration, but the effect size is fairly minor (only model 4 and model 6 are significant at the 5 percent level, Table 2). Second, similar pattern emerges when considering family finances. Financial situation within the family translates into closer views on the immigration (see models 1 to 3, Table 2), where higher positioning in terms of family finance is associated with more positive views on the immigration. However, this mechanism is not very robust when considering models 4 to 6, Table 2. Our analysis suggest (1) congruence between parents and their adolescent children's political attitude exists. (2) Both, socializing (income) and communication mechanisms to some extent influence this relation.

Our second line of analysis on anti-immigrant attitudes deals with the age difference. Here, similarly as with ideology, we find no evidence that age matters in terms of attitudes toward immigration. We find no evidence that a unit increase in age has an impact on immigration attitudes or for that matter on congruence about these attitudes with the parent. Neither do we find confirmation that participation in elections would matter. Both in models 2 and 5, where we investigate the difference between those with the right to vote and those without the right to vote as our baseline, we find no statistical difference between the two groups. There are also no differences when model 3 and model 6 are considered, i.e., when mock participation in elections or full participation in elections is available for adolescents. Thus, we conclude that the hypothesis that political participation would lead to a higher level of political maturity is not supported by the results of our analysis.⁹ We find no differences in political maturity by age, seen through ideology and immigration attitudes, and we find similar congruence effects of ideology and immigration attitudes with that of parents and adolescents among all age groups.

Conclusion

Our analysis shows that two factors are important regarding ideology and positional attitudes of adolescents. First, we find that adolescents resemble the political ideology and positional attitudes of their parents. This seems to be transmitted through communication channels within the family, when ideology is taken into account. However, the communication channel is less robust when immigration attitudes are concerned. The socialization mechanism, mainly through the family financial situation, on the other hand, might also have influence on the political attitudes (immigration attitudes). Our findings suggest that adolescents follow their parents in both ideological views and political attitudes, and they show that the family remains one of the stronger predictors of future political attitudes of adolescents. To some extent this is more strongly the case in terms of ideology, than with political attitudes, but overall the parents’ views on both dimensions provide strong indications about adolescent children’s political attitudes.

Second, the ideological views and positional attitudes are already formed at age of 15, and we find no different patterns about ideology congruence or immigration congruence between parents and adolescent children across different age groups. This infers that, at least in our sample, there is no evidence of a distinct generation that would suggest different political maturity between ages 15 and 19. Therefore, we believe that political maturity is formed even earlier than what is often assumed. Such findings incline that the age for right to vote could be lowered, without a loss of the quality of the vote. Adolescents already have formed ideological views as they have established opinions regarding policy issues (i.e., attitudes toward immigration), and their votes should be reflected similarly in the elections, no matter if it is a vote from a citizen at age of 16 or at the age of 18. The ability to vote is no different by age and it is not affected by the virtue of the fact that some groups have the right to vote compared to others. Political attitudes thus are surprisingly already formed at age 15, and lowering the voting to 16, we believe would not make a big difference regarding the political maturity of these voters, nor will it have an effect on the option to vote, as some other research has argued (Eichhorn 2014). However, in line with Franklin (2004), it might lead to higher levels of voter turnout, if the habit of voting hypothesis are correct.

However, we should state that our sample comes from the city of Ghent only, which imposes limitations in the way these findings can be generalized across different countries and geographical settings (i.e., rural/urban or cultural differences). Additional studies should test congruent voting in various countries and under different conditions to address the generalizability.

Future research should also focus on two important aspects of voting at 16. First, it should be investigated whether there indeed is evidence of long-term higher voter turnout in countries that have already implemented voting at 16. In this line of research, we should also investigate whether there is a real social demand for lowering the voting age. In the past, the women’s movement and the labor movement waged a hard and long campaign to obtain the right to vote. One does get the impression that the current generation of adolescents is not equally convinced that they should have the right to vote (Douglas 2020). Second, our study suggests the exact timing when political ideology and political attitudes are being formed is before the age of 15. Therefore, further research should investigate at what age do children create their political views? The results of the current analysis seem to indicate that the answer to this question should be at an earlier age which is often not assumed in the current literature.