Are the Young Undemocratic? Evidence from a Conjoint Experiment

ABSTRACT Are younger generations less supportive of democracy than older generations? This article adds to ongoing scholarly debates on young people’s support for democracy. We contribute to the field by using evidence from a conjoint experiment embedded in a survey with a representative sample of the Finnish population (n = 1030) to examine whether support for core democratic principles is weaker among the younger generations (Generation Z and Millennials) than the older generations (Generation X and Baby Boomers). Our results do not support the expectation that the younger generations in Finland would be systematically less committed to democratic norms than the older generations. However, we find some generational differences in responses to different democratic norm violations. Our findings make a second contribution to the field of youth and democracy by extending our analyses to the youngest generations of age, Generation Z, whose democratic support has not yet been broadly examined. Our study contributes both to the ongoing debate on democratic deconsolidation in established democracies as well as to the literature on young people’s attitudes towards democratic institutions.


Introduction
Are younger generations increasingly losing their support for democracy in developed democracies?How young people view democracy is important for democracy's future, as democratic beliefs and attitudes are developed early in life (Denemark et al., 2016;Quintelier & van Deth, 2014).For a while now, young people have displayed a declining interest in the institutions of representative democracy, embodied by elections and political parties (e.g.Bennett, 2008;Chou, 2017;Furlong & Cartmel, 2007;Grasso et al., CONTACT Janette Huttunen janette.huttunen@abo.fiSupplemental data for this article can be accessed at https://doi.org/10.1080/00344893.2023.2242386.

2018
).However, several theorists have argued these trends reflect the fact that the young prefer more direct, elite-challenging and engaging forms of political participation, rather than a lack of support for democracy as a regime type (Inglehart, 1997;Dalton, 2008Dalton, , 2016;;Pickard, 2019).Other scholars, however, have continued to display a significant worry over young people's support for, and their attachment to, democracy (see e.g.Denemark et al., 2016;Foa & Mounk, 2016, 2017).
In a widely publicised article, Roberto Stefan Foa and Yascha Mounk argued that support for democracy is declining in established Western democracies, and that the 'steep' erosion of democratic support was particularly notable among the younger citizens and especially among the millennial generation (Foa & Mounk, 2016, for media coverage, see, e.g.Breene, 2017;Howe, 2017).Most recent studies have not generally found support for these arguments, and have concluded that overall, the youngest cohorts in established European democracies continue to support democracy as a political regime type, but in some European countries, the youngest cohorts were more receptive to undemocratic ways of governing (Wuttke et al., 2020a, p. 2, see also 2020b, see also the responses by Alexander & Welzel, 2017;Norris, 2017 to the more general arguments in Foa & Mounk, 2016).The evidence of young people's support for democracy in established Western democracies is thus not conclusive.Yet, it represents a very important question for democracy's future globally.If younger generations are systematically more open to non-democratic ways of ruling, trends in generational voter replacement could lead to a serious decline in the share of the population that supports democratic ways of governing in the future (Wuttke et al., 2020a).
In this article, we explore whether the younger generations are less supportive of democracy than the older generations using experimental evidence from Finland, an established Nordic democracy.Our study contributes to the emerging literature on democratic support in established democracies that has begun to measure citizens' commitment to democratic norms with experimental methods rather than relying on answers to direct survey questions that may be subject to 'social desirability bias' (Carey et al., 2020;Graham & Svolik, 2020;Svolik, 2020;Saikkonen & Christensen, 2022, on social desirability bias, see Wallander, 2009;Hainmueller et al., 2014).Yet, to our knowledge, no studies have yet used experimental evidence to examine whether we could find sub-group differences in the support for democracy between various generations.We examine these questions using data from a conjoint experiment embedded in a nationally representative survey carried out in Finland in May and June 2020 (Christensen & Saikkonen, 2020).Conjoint experiments estimate respondents' evaluations of, for example, profiles of hypothetical political leaders that vary along a number of attributes, such as their respect for democratic norms (Bansak et al., 2021).Conjoint experiments are particularly suited for our analysis for two reasons.First, they are shown to mitigate problems associated with social desirability bias (Horiuchi et al., 2021).Indeed, several recent studies have found that distinct sub-groups of citizens in democracies do not express a commitment to central democratic principles when this is measured indirectly with survey experiments, even though a majority of the survey respondents state support for central democratic principles when asked directly (Svolik, 2020;Saikkonen & Christensen, 2022).Second, conjoint experiments are particularly well suited to measure the complex choices that respondents make between several dimensions, such as antidemocratic behaviour and other characteristics of politicians (Bansak et al., 2021).
To our knowledge, our study is therefore the first to examine young generations' commitment to key democratic principles with experimental evidence.Moreover, to our knowledge, our article is also the first study to explore the level of democratic support among the youngest generation of age, Generation Z.Much has been written about the Millennials, who are often still even in the 2020s referred to as the 'young' generation, despite that the oldest Millennials approach 40 years of age.Exploring the support for democracy also in the young Generation Z, those born in 1997 or after (Dimock, 2019), is our second contribution to the literature.
In this article, we explore whether the younger generations (Generation Z and Millennials) are less supportive of democracy than older generations (Generation X and Baby Boomers) in Finland. 1 We examine support for (representative) democracy as a political regime type and focus on the potential lack of commitment to fundamental democratic norms.To preview our findings, our results do not in general lend support to the suggestion that the younger generations would be systematically less supportive of democracy in Finland.However, we find some interesting generational differences in responses to different democratic norm violations.These are particularly pronounced among the Baby Boomer generation.These results can suggest that the reactions to violations of specific democratic principles may differ between various generations.
The article is structured as follows: First, we outline the findings of the previous literature on young people's support for democracy, followed by our hypotheses and their theoretical implications.After discussing the research design, we present the results from our analyses in two sections.We end the article with a concluding discussion.

Young People, Democracy and Democratic Deconsolidation
The current debate on young people and democracy is often related to political participation and whether the young are politically active or passive.A large set of literature has examined the age and generational differences in electoral turnout (see e.g.Bhatti et al., 2012;Blais et al., 2004;Nemčok & Wass, 2021;Wass, 2007).Due to the decreasing trends of youth engagement in democratic institutions, mostly elections (e.g.Bennett, 2008;Furlong & Cartmel, 2007;Grasso et al., 2018), the young have been seen as apathetic.This has led to debates on the normative question of whether the potential passivity of the young is a threat to democracy (Amnå & Ekman, 2014), even if some suggest that citizen passivity is not necessarily a negative feature but can even be an asset for democracy (see Amnå and 'standby citizens' (Amnå & Ekman, 2014)).The question of young people's disengagement with democracy and democratic norms has been, however, prominent in many fields.For example, in citizenship education literature the focus has often been on the democratic 'character' building (see Mills, 2022), or how young people can 'become' (democratic) citizens though (civic) education, as empirical evidence show that formal education can have significant democratic returns (Hoskins et al., 2008).However, one common point of departure in youth participation literature in political science is the notion that instead of passivity, which in turn would entail a negative relationship with democracy, the young merely prefer political engagement in different ways than older generations do.Perhaps most notably, Inglehart's (1997) post-materialism theory argues that due to societal modernisation since WW2 and growing emphasis on post-materialistic values such as quality of life and individual self-expression, the patterns for political engagement for younger generations have changed into more elite-challenging and active participation.Societal value change, alongside rising educational levels and easier access to political information (Dalton, 2007;Dalton et al., 2001) have resulted in a shift in youth engagement.Instead of displaying support for the democratic system by voting in elections, the young want to engage in more engaged, autonomic, sporadic, and active forms of non-institutionalised political participation (Inglehart, 1997;Dalton, 2008;Hustinx & Roose, 2016, p. 95;Norris, 2004).Thus, the decline in young people's election participation does not necessarily reflect a declining commitment to democracy among the young.However, while the discussion on the nature of young people's political participation is rich and nuanced, the main focus of our article is on young generations' support for basic democratic principles.
A significant worry over young people's support for basic democratic norms was raised in two articles published in 2016 and 2017, where Foa and Mounk argue that there is a widening and alarming generational gap in support for democracy.They argue that Millennials (defined as those born in the 1980s) see democracy as less essential than older generations; Millennials are disillusioned with liberal democratic institutions and are more attracted to authoritarian alternatives than older generations are, which manifests itself in voting for anti-establishment parties and candidates (Foa & Mounk, 2017).The decline in support for democracy is, according to Foa and Mounk, owed more to the generational than age effect: it is specific for the young millennial generation, and not associated with youth independent of time period.If true, the decline of democratic support among the young is thus not caused by the trend that people would hold more extreme political preferences at a young age that would mellow as they grow older (as the life-cycle approach to political preference formation would suggest).Rather, they argue that the social and societal changes that Millennials have experienced, such as individualisation and digitalisation and the rise of economic insecurity (Foa et al., 2020;Wuttke et al., 2020a), have led to this age cohort to increasingly turn away from democracy as a regime type.Similarly, Denemark et al. (2016) examined generational differences in the support for democracy in established democracies, and found evidence of decreasing support for democracy among the young, especially in Anglophone countries.
Foa and Mounk's claims on the broader trends in the loss of democratic support in developed democracies faced immediate criticism in the literature (see Alexander & Welzel, 2017;Norris, 2017;Voeten, 2016).The responses ranged from criticism towards the reliability of the empirical findings (Norris, 2017) to criticism of whether democracy is comparable for different generations (see esp.Alexander & Welzel, 2017).However, some studies that have focused more on examining generational differences have found support for the existence of generational differences in democratic attitudes.Voeten (2016) found that Millennials were somewhat more favourably inclined towards non-democratic ways of ruling; nevertheless, this seemed to be more of an age rather than a cohort effect.A recent study that revisited this debate (Wuttke et al., 2020b) found that, overall, the youngest cohorts in established European democracies continued to support democracy as a political regime type.However, they did find that in some European countries the youngest cohorts were more receptive to undemocratic ways of governing (2020a, p. 2).The evidence on young people's support for democracy is thus not conclusive.
Yet, all of these studies have analysed the differences in youth attitudes toward democracy through traditional survey questions.As outlined above, our article is part of an emerging literature that has begun to measure the support for democracy in established democracies, focusing especially on the differences between different sub-groups of the population, with experimental evidence.We contribute to this literature by examining democratic support among the young with experimental evidence from Finland.Finland is an established Nordic democracy, with a consensus-driven political culture and generally high levels of democratic satisfaction (Bäck et al., 2016;Kestilä-Kekkonen, 2015;Rapeli & Koskimaa, 2020) and institutional trust (Kestilä-Kekkonen & Söderlund, 2016).However, previous research has found distinct sub-groups with a weaker commitment to democratic norms even in Finlandsuggesting that democratic deconsolidation may be an issue even in 'consensus democracies with relatively low levels of political polarization' (Saikkonen & Christensen, 2022).Therefore, the commitment to democratic norms does not appear to be uniform among the Finnish population and this makes it warranted to investigate whether there are generational differences in Finland when it comes to commitment to democratic principles.
In comparisons with other Nordic countries, which often provide the most natural point of comparison due to similar political cultures, Finland appears to differ in two ways: Finland scores lower in political interest and in internal efficacy than the other Nordic countries (Rapeli & Koskimaa, 2020).In the latter measures, Finnish citizens exhibit lower levels even in comparison with Southern European countries (Rapeli & Koskimaa, 2020).However, in general, the Nordic countries resemble each other, also in the case of young people and democracy.Similar trends in youth participation with declining participation rates in elections are visible also in Finland, despite that the decline is not as deep as in some other countries (see Nemčok & Wass, 2021).Finnish young people tend to have rather traditional views on politics, they are sceptical of their abilities to influence political decision-making within representative democracy and among the least motivated to participate politically in European comparisons despite displaying high levels of political knowledge (Myllyniemi, 2014).The Nordic young people are in general relatively passive with regard to political participation, yet they are knowledgeable and democratically engaged (Lieberkind & Bruun, 2021).Due to the similarities, the Finnish case is likely to be comparable with the rest of the Nordic countries regarding the generational support for democracy.

Hypotheses: Young People and the Support for Democracy
The current worries about young people's declining commitment to democracy in established democracies have largely centred on generational analysis (Denemark et al., 2016: Foa & Mounk, 2016;2017).We, therefore, focus on testing the expectation that younger generations would be less committed to democratic norms than older generations.Our article uses data from a pre-registered conjoint experiment conducted in Finland in 2020. 2 In line with recent studies (Carey et al., 2020;Graham & Svolik, 2020) this experiment tested citizens' commitment to democracy by exploring whether citizens are more or less likely to choose political leaders who transgress key democratic norms.
Democracy is a multidimensional concept (Dahl, 1971;1991), but functioning democracies generally share four fundamental dimensions: free and fair elections, respect for civil rights, respect for the legitimacy of the political opposition and respect for the system of separation of powers and the rule of law (Mettler & Lieberman, 2020).The experiment that we analyse here measured respondents' reactions to the trampling of the latter two key democratic norms in Finland, the respect for the legitimacy of the political opposition and the system of separation of powers and the rule of law (Christensen & Saikkonen, 2020).Worryingly, the 2019 Finnish parliamentary election campaign saw an increase in physical attacks on electoral candidates during campaign events (Wass et al., 2020).Moreover, there have been concerns over increasing targeted (online) harassment of judicial and other government officials (Illman, 2020).In our analysis, we contrast the effects among the different generational sub-groups, i.e. between the 'younger citizens', that is, the members of Generation Z and Millennial generations, to those among 'older citizens', that is, citizens belonging to Generation X and the Baby Boomer generation.
The first democratic principle that the experiment focused on relates to the fundamental democratic norm of tolerating the legitimacy of the opposition (Levitsky & Ziblatt, 2018;Mettler & Lieberman, 2020).Representative democracy is based on the competition between politicians for the support of the voters (Schumpeter, 1942).However, the principles of democratic politics entail that competition over these issues should be fought solely by peaceful means.Indeed, in his seminal book Juan Linz (1978) identified several warning signs for antidemocratic politicians.Building on this work, Steven Levitsky and Daniel Ziblatt (2018, pp. 21-24) argued that the denial of the legitimacy of political opponents as well as tolerance or encouragement for political violence can be considered central indicators of authoritarian politicians.The experiment, therefore, tested how willing citizens are to condemn politicians who exhibit these 'warning signs' by either inciting political violence against their opponents or failing to condemn this sort of behaviour.In practice, the experiment measures the effect that committing these democratic transgressions has on the favourability of the hypothetical politician trampling democratic norms.If younger citizens are less committed to democracy than older citizens are (Foa & Mounk, 2016), these effects should be weaker among the younger citizens (Generation Z and the Millennials) than the older citizens (Generation X and Baby Boomers).Accordingly, we test the following hypothesis 3 : H1: The effects of not condemning or inciting physical attacks against opposition candidates are weaker among younger citizens.
The second fundamental democratic principle that we examine concerns the rule of law and respecting the system of separation of powers in democracies.Democracy entails equality between all the citizens and their rulers (Dahl, 1991) and thus the rule of law can be considered a central democratic prerequisite (O'Donnell, 2004).Trying to undermine judicial independence is often one of the early warning signs of democratic backsliding (Carey et al., 2020;Chiopris et al., 2021;Levitsky & Ziblatt, 2018).The second norm that the experiment tested was therefore politician's judicial deference to court decisions, and whether politicians argue that judicial decisions should be respected even though this may be politically disadvantageous for their party.Again, we are interested in whether these effects would be weaker among young citizens when compared with older citizens, as this would signify a weaker commitment to democratic norms among the younger citizens.Accordingly, we test the following hypothesis: H2: The effects of disrespecting judicial officials are weaker among younger citizens.
Previous studies have found that even in Finland some sub-groups of the population are willing to ignore politicians' undemocratic actions when there is a congruence in ideological and policy goals between the politicians and the respondents (Saikkonen & Christensen, 2022).As this is true also in Finland, there is a reason to explore whether there are also sub-group differences in democratic support among different generations.

How We Operationalize Generations
We divide the respondents into four generations: Baby Boomers, Generation X, Millennials, and Generation Z. Baby boomers, defined as those born between 1946 and 1964 is a registered generation in the USA, thus definitions of Baby Boomers are typically constant across studies. 4The oldest respondents in our data are born in 1950, therefore we define here those born between 1950 and 1964 as Baby Boomers.The children of the Baby Boomers (born in the late 1960s and 1970s) started to be labelled 'Generation X' or 'Gen X' during the 1990s (Pickard, 2019, p. 42).In our study, Generation X is defined as those born between 1965 and 1980.
The Millennials, also known as Generation Y, are born after Gen X, and the Millennials is the generation that has received the most attention in the democratic deconsolidation literature (see Alexander & Welzel, 2017;Foa & Mounk, 2016;Norris, 2017;Voeten, 2016).In these studies Millennials are typically defined as 'those born in the 1980s', however, here we use a more specific definition of the generational lines.We follow the definition used by the Pew Research Center (Dimock, 2019), which defines Millennials as those born between 1981 and 1996.
We contribute to the research field by extending our analysis of young people's support for democracy to the youngest generation that has reached adulthood as of 2020, that is, Generation Z, also colloquially referred to as 'Zoomers'. 5Pew Research Centre defines Generation Z as those born in 1997 and after (Dimock, 2019).The end birth year of the generation has not been assigned yet.The data contains only those members of Generation Z who had the right to vote at the time the data was collected.Therefore, the respondents born between 1997 and 2002 are categorised as members of Generation Z in our analyses.
Our operationalisation of the different generations is summarised in Table 1.

Data and the Conjoint Experiment
In order to explore our research question 'Are younger generations (Generation Z/the Millennials) less supportive of democracy than older generations (Generation X and  1965-1980Millennials 1981-1996Generation Z 1997-2002 Boomers)?',we make use of a dataset collected by Christensen and Saikkonen (2020). 6 The dataset consists of a conjoint experiment embedded in a nationally representative survey 7 of the population in Finland with regards to age, gender and region of living, with a sample size of 1030 via an online panel recruited by Qualtrics.
The conjoint experiment allows us to examine the respondents' preferences on democracy in an indirect way (see e.g.Graham & Svolik, 2020).In a conjoint analysis, respondents are presented with sets of alternatives that have randomly varied attributes and the respondents need to evaluate the sets (Hainmueller et al., 2014).The attributes are different theoretically relevant values, captured by different characteristics that are assumed to affect the evaluations made by the respondents.In this particular conjoint experiment, the respondents were presented with two randomised profiles of hypothetical candidates for prime minister of Finland and asked which profile they would prefer.The profiles (also referred to as politicians or leaders in this article) were randomised on seven attributes.The first two attributes pertained to the two essential democratic norms: respecting the physical integrity of the political opponents and respecting the decisions of judicial officials.The other attributes that were included in the conjoint experiment were ideology (leftist-centrist-rightist); position on immigration policy; and two background characteristics that generally make the profiles more realistic to respondents, that is, gender and the level of education; as well as whose wishes the decisions made by the imaginary politician will reflectordinary citizens, political elites or social groups.All the attributes and their levels are presented in Table 2. Hainmueller et al. (2015) found that despite worries of lacking external validity due to, e.g.response bias, stated preference experiments, in particular paired conjoint design, can measure real-world behaviour remarkably well.In this 'forced choice' design, respondents have to choose which one of the two profiles in each paring they would support, which allows similar trade-offs as in real-life elections (Hainmueller et al., 2014).We use the conjoint data to explore whether there are generational sub-group differences in the effects of politicians committing democratic transgressions on politicians' favourability.Thus, we test whether the commitment to key democratic norms is weaker among the younger generations than among the older generations.
In our analysis of the conjoint data, the dependent variable is whether a given profile was chosen or not in a comparison.We estimate the average marginal component effects (AMCE) using linear regression (Hainmueller et al., 2014).AMCE is the effect of an individual treatment component.We obtain the AMCEs for all attributes simultaneously by running a single regression of the choice outcome on the set of dummy variables for the attribute values (Hainmueller et al., 2014).In our analysis, the AMCE is interpreted as the average change in the probability that a profile will win support when the profile includes the listed attribute instead of the baseline attribute value (see Table 2 for baseline values (R)) (Hainmueller et al., 2014).Thus, when we measure the effect of a candidate inciting or accepting a democratic transgression, we measure the effect that such a democratic transgression has on the probability that the candidate is chosen.The causal effect of an attribute may interact with respondents' background characteristics (Hainmueller et al., 2014), such as age.In this article, we focus on estimating conditional AMCEs (Bansak et al., 2021) for the generational sub-groups, that is, analysing whether these effects differ between the generations.We estimate conditional average marginal component effects (conditional AMCEs) by conducting an interaction regression analysis for all attributes simultaneously (Hainmueller et al., 2014) between the attributes in the conjoint experiment and the four different generations to detect possible interactions between age and the effects of democratic transgressions on candidate favourability.
Alongside the conditional AMCEs, we focus particularly on examining the marginal mean estimates, and the differences between the generational sub-groups.Marginal means tell us the level of favourability toward profiles with a particular feature level when ignoring all other features (Leeper et al., 2020).It allows us to verify the popularity of a given attribute level without using a particular reference category, as the conditional AMCE results can be sensitive to the reference category selected (Leeper et al., 2020).Marginal mean estimates reflect the percentage of respondents that have chosen said alternative (of profiles) every time that alternative has been shown.
Our sample matches the general population in Finland well and therefore it is not necessary to weight the data. 8We report the results in coefficient plots, as is recommended by Hainmueller et al. (2014).

Empirics: Generations and Support for Democracy
We explore the differences in generations' democratic attitudes based on the conjoint experiment data.Based on the two hypotheses we are testing, we expect that the effect of the two measured democratic transgressions, condemning or inciting violence against opposition and disrespecting judicial decisions, on leader favourability is weaker among the younger generations (Generation Z and Millennials) than the older generations (Generation X and Baby Boomers).The results are presented in Figure 1.The regression results can be found in the Supplementary Appendix.
Hypotheses 1 and 2 suggest that we should find that the effects of democratic transgressions on leader favourability should be weaker among the younger generations (Generation Z and the Millennials) than the older generations (Generation X and the Baby Boomers).Here we therefore first examine whether the results show substantive and systematical differences between these generation groups.The conditional AMCE estimates, presented in the left hand panel of Figure 1, suggest that the effect of committing democratic transgressions on leader favourability is negative for all of the generational sub-groups, and there are some differences in the effects between the different groups (which we discuss further below).However, neither the conditional AMCEs nor the marginal means estimates, presented in the right hand panel, suggest substantive and systematic differences when the two younger generational groups (Generation Z and the Millennials) and the two older generational groups (Generation X and Baby Boomers) are compared.Indeed, the marginal means estimates among the Millennials (part of the younger generation group) and Generation X (part of the older generation group) are mostly of similar magnitude.Therefore, we cannot conclude that the younger generations (Generation Z and the Millennials) in Finland would be systematically more lenient towards democratic norm violations by politicians when compared with the older generations (Generation X and Baby Boomers).
However, there are some interesting differences in effects between the generations.When examining the marginal means estimates in the right hand panel of Figure 1, we can see that the negative effects of disrespecting judicial independence on candidate favourability are most pronounced among the Baby Boomer generation.Moreover, these effects are less pronounced among Generation Z.However, when we examine generational differences in reactions to candidates inciting violence against opponents we find that the effects are least pronounced among the Baby Boomer generation.Thus, for Baby Boomers, there is a pronounced negative effect on a candidate's favourability if the leader does not respect judicial decisions, but the negative effects of inciting violence against political opponents are somewhat less pronounced among this generation.These results suggest that different generational groups may respond differently to the violations of various democratic principles.These results are in line with previous studies, which suggest that the reactions to different democratic violations differ between population sub-groups (see, e.g.Carey et al., 2020).
To examine these results further we conducted robustness checks with different generational comparisons and different codings of the age groups.The results from these analyses are presented in Figures S2-S6 which, together with the regression results, can be found in the Supplementary Appendix.The robustness tests show similar results as our main generational analysis.First, we conducted a robustness check where the younger generational groups (Generation Z and the Millennials) and the older generational groups (Generation X and Baby Boomers) were combined into two groups.As seen in Figure S2, there are small differences in the effects between the groups in the judicial independence treatment, but no substantive differences when it comes to the violence treatment.We also conducted robustness checks with different codings of young vs. older: first, 'young' was defined in line with the Finnish legislation (Youth Act Law 1285/2016, 3 §, which defines people under 29 as 'young'), and second, 'young' was defined along the custom within youth studies, focusing on the youngest possible respondents (using age brackets of 18-24; see, e.g.García-Albacete, 2014, p. 80;Huttunen, 2022).These results do not contradict the overall conclusions of our main analysis.Moreover, as the discussion of the democratic deconsolidation literature has largely focused on the Millennials and the Baby Boomer generations (see, esp., Foa & Mounk, 2017), we also examined subgroup differences between the Millennials and the Baby Boomers, as well as compared Baby Boomers with all of the younger generations.In all of the robustness comparisons, we find that there is a negative effect of leader transgressions for all the generations.However, when comparing the Baby Boomer generations to the rest of the population we find that the negative effect of inciting violence for candidate favourability is less pronounced among the Baby Boomers than the younger generations, while the negative effect of disrespecting the decisions of judicial officials is more pronounced among this generation.The results of the robustness tests further suggest that there may be something specific in the Baby Boomer generation that differs from the three younger generations in regards to their attitudes toward democratic norms.
The main goal of our paper has been to examine whether there are systematic differences between the younger generations when compared to the older generations when it comes to the commitment to democratic norms.Our overall results do not suggest that the younger generations would systematically support democratic norms to a lesser degree than the older generations. 9However, the differences between the post-war generation and the other generations in Finland are very interesting and merit further scholarly attention.

Concluding Discussion
There is a widespread worry that young generations in particular are not as supportive of democracy as older generations are.The democratic deconsolidation literature (Foa & Mounk, 2016, 2017) sparked an extensive discussion on 'undemocratic Millennials' (see also Denemark et al., 2016).Newspaper articles both stating that 'Millennials are rapidly losing interest in democracy' (Breene, 2017) as well as asking whether Millennials are giving up on democracy (see e.g.Howe, 2017), have been a part of political journalism ever since.The threat of undemocratic young people has led to policy shifts in e.g.education, with the aim to encourage the active and 'good' citizenship of young citizens (Hoskins et al., 2008;Mills, 2022).In the field of youth studies within political science, there has also been a long-lasting worry about young people's attachment and support for democracy and especially its key institutions.The young are often held responsible for an impending democratic deficit (see Pickard, 2019, pp. 60-61), since young people's support for elections and political parties, the key institutions of representative democracy, has been declining (e.g.Bennett, 2008;Chou, 2017;Furlong & Cartmel, 2007;Grasso et al., 2018;Pickard, 2019).Despite political scientists' explanations that the apparent youth disengagement in institutionalised forms of politics is a shift towards more active and direct participation in non-institutionalised forms of political engagement rather than rejection of democracy (see e.g.Dalton, 2008Dalton, , 2016;;Huttunen & Christensen, 2020;Pickard, 2019), any potential signs that the young generations are less supportive of democracy should be explored.
In line with recent studies that have focused on generational differences in democratic support in established democracies (Denemark et al., 2016;Foa & Mounk, 2016;2017), we explored the question of whether the younger generations (Generation Z and Millennials) are less supportive of democracy than the older generations (Generation X and Baby Boomers).We used data from a conjoint experiment embedded in a nationally representative survey conducted in Finland in May and June 2020.The results suggest that the worry over young people's lack of support for democracy may be overdriven, at least in the Finnish context.Regarding our hypotheses, the results suggest that the younger generations in Finland are not systematically less democratic than the older generations, when support for democracy is measured in an experimental setting mimicking a real-life situation.We found no systematic patterns that would suggest that the younger generations (Generation Z and Millennials) are less committed to central democratic principles than the older generations (Generation X and Baby Boomers).
While our results do not indicate that the younger generations would be especially anti-democratic, we found some interesting differences in the attitudes towards democratic norms between the Baby Boomer generation in Finland and the other three generations.In particular, the negative effect of violating the democratic norm of judicial independence was more pronounced among this generation.The robustness tests also revealed a less pronounced negative effect of inciting violence on candidate favourability when Baby Boomers were compared with the all the other generations.These results suggest that there may be differences in the commitment to democratic norms between the post-war and other generations in Finland.These findings are interesting and merit further scholarly attention.
Our paper contributes to and builds on the emerging literature that has used experimental evidence to examine the extent of democratic support in established democracies (Carey et al., 2020;Graham & Svolik, 2020;Saikkonen & Christensen, 2022).This approach is particularly suitable for examining sub-group differences in political attitudes that may be subject to social desirability bias (see esp.Hainmueller et al., 2014).
However, there are naturally limitations to our study.First, survey experiments measure respondents' commitment to democracy at a given point in time, and this type of evidence does not allow for the identification of life-cycle or other time-dependent effects.Therefore, our findings should be complemented with further studies using panel data.Second, the conjoint experiment treatments include only two dimensions of democracy, and further studies should probe sub-group differences in respondent reactions to different democratic transgressions.Third, our data comes from a sample of Finnish respondents, which may naturally impact the generalizability of our findings.However, previous research has found distinct sub-groups with a weaker commitment to democratic norms even in Finland (Saikkonen & Christensen, 2022).Thus, even 'least-likely-cases' may reveal worrying tendencies, which is why examining the questions of deconsolidation and generations' support for democracy also in these contexts is important.The results are most likely generalisable to the other Nordic countries due to similar political and social cultures (Rapeli & Koskimaa, 2020) and the similarities of Nordic young people (e.g.Lieberkind & Bruun, 2021).Since Finland scores lower than the other Nordic countries in, for example, internal efficacy and political trust, the results are promising for the Nordic neighbours: the support for democratic core norms is likely to be even stronger in the other Nordic countries.Our results are also in line with the previous findings by Norris (2017), andAlexander andWelzel (2017).However, these results should be further tested with samples from other Nordic countries and as well as other established democracies.
The implication of our study is that even if other recent experimental studies have identified distinct sub-groups of the population in established democracies that exhibit a weaker commitment to central democratic principles (Carey et al., 2020;Graham & Svolik, 2020;Saikkonen & Christensen, 2022), our results do not lend support for the notion that support for democratic principles would be systematically lower among the younger generations in Finland.We should thus be cautious of accounts that put the blame for democratic deconsolidation solely on the Millennials or other younger generations.
Notes 7. We make use of both the summary statistics of the responses to the direct survey questions as well as data from the conjoint experiment.The analysis of the responses to the direct survey questions can be found in the Supplementary Appendix.8.A survey/population comparison can be found in the Supplementary Appendix.9. We have also conducted analyses on the direct survey questions on respondents' democratic attitudes to further examine potential differences between the generations' commitment to democracy.We found that the small difference between the attitudes towards judicial independence between the Baby Boomer generation and all the other generations is also visible in the answers to the direct survey questions.See the Supplementary Appendix for more discussion on the analyses of the answers to the direct survey questions.

Disclosure Statement
No potential conflict of interest was reported by the author(s).

Figure 1 .
Figure 1.Differences in the effects of democratic transgressions across generations (conditional AMCEs and marginal mean estimates).

Table 1 .
Classification of generations.

Table 2 .
Conjoint attributes and attribute levels.