Introduction
Generational replacement is potentially a powerful driver of political and social change. If the attitudes and behaviour of more recent generations differ systematically from those of previous generations, and if these differences are not the result of life-cycle effects (for instance that people become more conservative as they grow older), public opinion will automatically shift when the oldest generations are gradually being replaced by younger ones. Some recent observations from Europe and North America suggest that there are indeed substantial differences between generations in political attitudes and behaviours. On average, the younger generations across the EU are more supportive of European unification than older generations (e.g., Down and Wilson, Reference Down and Wilson2013) and in the UK they are more likely to have voted ‘Remain’ in the Brexit referendum (e.g., Frese et al., Reference Frese, Härkönen and Hix2025; Hobolt, Reference Hobolt2016). Across Europe, they are more willing to support measures to protect the environment and to reduce global warning than older generations, and they are overall more likely to vote for Green parties (e.g., Goerres Reference Goerres2008; Lichtin et al., Reference Lichtin2023). In the US, younger generations are more liberal than older generations (e.g., Jocker et al., Reference Jocker2024; Norris and Inglehart, Reference Norris and Inglehart2019). At the same time, the younger generations in the US and Europe are less likely than older generations to turn out at elections (e.g., Franklin, Reference Franklin2004; Grasso, Reference Grasso2014). Because of these differences between generations, it is not surprising that many social scientists are interested in generational differences to explain political and social change. In this review article, we focus on generational differences, reviewing in particular the booming field of research on the role of generations in electoral processes.
The academic interest in generations is far from new. In modern sociology, Mannheim (Reference Mannheim and Kecskemeti1952 [1928]) was the first to develop a theoretical conceptualisation of generations. Ever since the publication of The American Voter electoral researchers have realized that political socialization plays a key role in understanding voting behaviour (Campbell et al., Reference Campbell1960). The idea is that young people would develop a psychological attachment to a specific party as part of their self-identity. This sense of ‘party identification’ (party ID) would then be a strong determinant of their actual party choice in specific elections. To the extent that party ID is stable during people’s lives and to the extent that young citizens adopt the party ID of their parents, we would expect individual electoral behaviour as well as aggregated election results to be very stable. So, the earliest studies on generations predicted stability rather than change.
However, this perspective shifted with the rise of the protest generation in the 1960s, a generation that, on average, had acquired different values than their parents. Inglehart’s (Reference Inglehart1977) famous theory of the ‘Silent Revolution’ is based on the idea that the post-war generation that was socialized in a period of peace and prosperity, would be more likely to prioritize postmaterialist values than the generations before. As these younger generations would gradually replace the older generations, postmaterialist values would slowly become more important – hence the title The Silent Revolution. Inglehart’s theory was also based on the idea that political attitudes and behavioural habits are acquired early in life during the most ‘formative years’ and tend to remain rather stable afterwards. This is because people tend to get ‘set in their ways’ as they grow older (Franklin, Reference Franklin2004). But while Inglehart’s theory predicted stability of preferences at the individual level, it predicted macro-level changes as a result of generational replacement.
Besides the different (historical) contexts in which generations are socialised, there is a second reason why generations might differ systematically from each other: composition. Generations often differ on characteristics that are related to political attitudes and behaviours, most notably religion, education, occupation and ethnicity. In many democracies, newer generations are on average better educated, more secular, work in different areas (e.g., the increase of the service sector in comparison to industry and agriculture), and are more ethnically diverse than previous generations. These variables are related to political engagement, turnout and party choice. For instance, if Evangelical Christians are likely to support the Republican Party in the US, and younger generations consist of fewer Evangelicals than older generations, generational replacement will hurt the Republican Party (all else equal). In survey research, the role of compositional differences can be estimated easily by comparing models with and without controlling for these variables.
A more challenging difficulty in the study of generations is that one needs to disentangle three components of change: Age, Period and Cohort (APC). ‘Age effects’, also referred to as ‘life-cycle effects’, are changes in political attitudes and behaviours that result from growing older. Some of the observable differences between age groups, like the fact that younger citizens are more likely than older ones to vote for Green parties (e.g., Lichtin et al., Reference Lichtin2023) and less likely to vote for the British Conservatives (Tilley and Evans, Reference Tilley and Evans2014), are partially the result of life-cycle effects. To the extent that differences between age-groups result from life-cycle effects, the progressive youngsters will become more conservative as they grow older. In this case, generational replacement will not lead to any meaningful political changes. In fact, the demographic changes resulting from the ageing of the populations of many advanced democracies would then lead to more conservatism at the aggregate level.
Period effects are changes in political attitudes and behaviours that affect all age groups and all generations to the same extent. If a governing party is involved in a huge scandal and loses support in society across all age groups and generations, we call this a ‘period effect’. Meanwhile, the term ‘cohort effect’ refers to birth cohorts, a term used interchangeably with generations. A generation, or birth cohort, is a group of people born in the same period, who are expected to have similar formative experiences because of the historical context in which they grew up (e.g., Inglehart, Reference Inglehart1977; Mannheim (Reference Mannheim and Kecskemeti1952 [1928]). Cohort effects are systematic differences between different birth cohorts, which exist even when controlling for period and age. To acquire a good understanding of the nature of political change (or stability), one needs to disentangle Age, Period and Cohort-effects, which is methodologically challenging because the three are perfectly multicollinear.
Even though the theoretical underpinnings of political socialization and of generational differences are far from new, the field has attracted much renewed interest over the past decade. One reason is probably that there has been a lot of political change recently. To understand these changes, it seems important to assess to what extent these changes are driven by generational replacement. Another reason might be that data availability, in combination with methodological progress and fast computers, have facilitated research in this field. As pointed out, with Age, Period and Cohort-effects being perfectly multicollinear, one needs to impose quite some restrictions on the models. For instance, to have enough power to test for time trends and generational and life-cycle effects, one needs large data sets that span multiple decades. Some of the larger cross-national trend studies, like the Comparative Studies of Electoral Systems (CSES), the European Social Survey (ESS) and the European Elections Studies (EES) have now been running for more than 25 years, and various household panels include more regular measures on political attitudes and behaviours. Thus, the possibilities to estimate APC-models have increased rapidly.
Before discussing recent research on generations and electoral processes, we will first discuss some conceptual issues, like at which stages in life are people politically socialized? What are generations and how do we define and classify them?
Generations and political socialization
Political socialization refers to processes through which individuals acquire their political beliefs, attitudes, values, and behavioural patterns that are expected to remain relevant later in life. As people grow older, they tend to ‘get set in their ways’ and become less likely to change (e.g., Dinas et al., Reference Dinas2024; Easton and Dennis, Reference Easton and Dennis1969; Franklin, Reference Franklin2004). Young people are therefore more susceptible to the formation of (new) attitudes and behavioural patterns, and that is why this period in life is often referred to as the ‘most impressionable years’ or ‘formative years’. Obviously, opinions and behavioural patterns are not fixed when people turn 30 or 40. Over the course of their lives, people are influenced by their family, friends and the media, and they adapt to changing circumstances (e.g., Alwin and Krosnick, Reference Alwin and Krosnick1991; Tyler and Schuller, Reference Tyler and Schuller1991). Yet, the formative experiences provide a home base (or ‘long term equilibrium’) to which habitual voters return after temporary changes in party preferences (e.g., Van der Eijk and Franklin, Reference Van der Eijk and Franklin2009). Generational differences will be particularly large if (1) generations have very different formative experiences, and (2) these formative experiences have a much stronger impact on attitudes and behavioural patterns than the experiences that happen later in life. When we speak of ‘political socialization’, we refer to the process by which attitudes are formed early in life (during the ‘formative years’).
There is no consensus, however, as to which period in life is the ‘most impressionable’ or ‘most formative’. Developmental psychologists have found evidence that basic orientations, in particular the formation of social identities and categorizations, are already formed during pre-adolescent years (see e.g., Sapiro, Reference Sapiro2004 for an overview). These insights have inspired research on the political attitudes of children and adolescents (e.g., Campbell, Reference Campbell2008; Hooghe, Reference Hooghe2004; Hooghe and Wilkenfeld Reference Hooghe and Wilkenfeld2008; Torney-Purta et al., Reference Torney-Purta2004; Van Deth et al., Reference Van Deth2011). However, for the formation of political attitudes and electoral preferences, most political scientists consider adolescence and early adulthood, roughly from 15 until 25 years of age, as the period in life that is most ‘formative’. The strongest learning effects were found to take place around the age of 18 (Bartels and Jackman, Reference Bartels and Jackman2014; Rekker et al., Reference Rekker2019; Schuman and Rodgers, Reference Schuman and Rodgers2004). While there might be reasons to expect that these learning effect take place mostly around the age of 18 per se, they could also be related to the age when people become eligible to vote in most countries. As a case in point, research on voting at 16 in Austria showed that political interest increases right at the time when voters reach the age at which they become eligible to vote, while no other differences were observed amongst the various young age-groups (Glantschnigg et al., Reference Glantschnigg2013; Zeglovits and Zandonella, Reference Zeglovits and Zandonella2013).
As mentioned previously, the term ‘generation’ is normally used to define a group of citizens who were born around the same time and who (therefore) experienced the same socializing context during their formative years (e.g., Mannheim Reference Mannheim and Kecskemeti1952 [1928]; Neundorf and Smets, Reference Neundorf and Smets2015; Smets and Neundorf, Reference Smets and Neundorf2014; Stoker, Reference Stoker2014). Historical events with a major impact such as the fall of the Berlin Wall, and the subsequent fall of communist regimes, might exert period effects as well as cohort effects. The fall of communism affected all citizens who lived in countries with such regimes, so everyone was affected immediately. To the extent that this affected citizens’ attitudes across the board (irrespective of age and generation), these should be seen as ‘period effects’. However, the fall of the Berlin Wall can also be used to distinguish generations: a generation that was politically socialized under communism and a generation that had their most formative experiences in a democratic state. This way of classifying generations is conceptually most logical, as it builds upon the findings in the socialization literature.
Yet, doing so is less straightforward than it may seem. First, one would need to create country specific classifications of generations, that are linked to the specific historical developments in each country, and which are relevant for the process being studied. In their study of generations and realignment in the Netherlands, Van der Brug and Rekker (Reference Van der Brug and Rekker2021) classified generations based on changes in the Dutch party system. As a consequence, their classification cannot be ‘automatically’ applied to other contexts, nor in a comparative setting. Second, when drawing boundaries between generations, it seems inevitable that people are classified as belonging to the same generation even though they have experienced rather different situations during their formative years. If the most formative years for political socialization are in the years that someone is between 15 and 25 years old, what does this mean for someone who was born in 1970 in East Germany? The fall of the Berlin Wall was in 1989 when she turned 19. This leads to her being partially socialized in a communist regime but then being socialized in a united German democratic state. This exemplifies how difficult it is to draw an exact boundary that distinguishes between people socialized during or after communism.
Because of these difficulties, different scholars have used different classifications, using sometimes the same labels for generations with different years of birth. In the US, many scholars use a categorization of generations that was popularized by the Pew Research Center (Dimock Reference Dimock2018): Greatest generation (born between 1910 and 1927), Silent generation (1928–1945), Boomer (1946–1964), Generation X (1965–1980), and Millennials (1981–1996). In the Western European context, many scholars now use the classification that was first proposed by Grasso (Reference Grasso2014): Pre-WWII Generation (born between 1909 and 1925), Post WWII Generation (1926–1945), the 60s-70s generation (1946–1957), the 80s Generation (1958–1968), the 90s Generation (1969–1980), and Millennials (1981–1996). More recent studies began to include the generation born after 1996, which is referred to as Gen Z (e.g., Dimock, Reference Dimock2019). Studies in Central and Eastern Europe often distinguish between those being socialized in a communist regime and those socialized more recently (e.g., Neundorf, Reference Neundorf2010). Yet, these studies usually do not focus on electoral behaviour.
Obviously, such classifications are rather crude. Someone born in 1996 may have more socializing experiences in common with someone born in 1997 than with someone classified as belonging to the same generation, who was born in 1982. Moreover, historical circumstances differ a lot between countries. While many Boomers in Northwestern Europe enjoyed a period of freedom and high employment in the late 1960s, Boomers in Spain, Portugal and Greece lived under an extreme right dictatorship and many American men of their generation were sent to fight a war in Vietnam. While they were socialized in the same years, their socializing experiences were clearly very different. It is not surprising therefore that studies of generational differences tend to conduct many robustness checks with alternative classifications, or with estimations of smooth curves to assess how sensitive the results are to arbitrary classifications.
When estimating life cycle-effects, scholars run into similar problems. To estimate their models, it is essential to categorize the respondents in different age groups that reflect certain stages in life. Based on insights from developmental psychology, researchers have distinguished the following phases in life: adolescence (ages 16–25), early adulthood (26–35), middle adulthood (36–65), and late adulthood or retirement age (65+) (e.g., Bartels and Jackman, Reference Bartels and Jackman2014; Dassonneville, Reference Dassonneville2017; Rekker et al., Reference Rekker2015; Steinberg, Reference Steinberg2010). These categorizations probably work well if one is mainly interested in the estimation of generational differences and ‘only’ needs these age categories as control variables. Yet, if the main purpose is to track patterns of political socialization across time, these categories may be a bit too crude, especially in early life phases.
Turning to period effects, scholars engaging in APC-modelling do not normally classify different periods on a theoretical basis. The common practice is to add year-dummies (or survey-wave dummies) to the model. In this way, the models control for oscillations resulting from sudden events, as well as changes in survey modes.
A different approach to conceptualize generations is taken by Kostelka and Blais (Reference Kostelka and Blais2021). The authors choose countries’ GDP to measure whether cohorts have experienced a certain level of affluence at the time of their political socialization. Generations are thus measured as cohorts born before or after economic transformations. This operationalization is thus largely comparable across contexts, even though economic transformations can differ from country to country.
Having introduced some of the core concepts in research on political socialization, we move to the main focus of this literature review: providing an overview of (new) research in the field of generations and electoral processes. Two main research traditions can be distinguished. The first tradition focuses on processes of political socialization during the formative years, focusing especially on the role of socializing agents like schools, parents, peers, and increasingly (social) media. When it comes to electoral behaviour, this research focuses mainly on the role of parents and peers. The population being researched are adolescents and young adults. The second research tradition studies long term changes in a specific outcome variable, resulting from generational replacement. The focus is on the population eligible to vote, and the purpose is to compare over time between generations and age groups. This research tradition consists mainly of APC-models, which pool survey studies over a long period of time and aim to disentangle Period, Age and (birth) Cohort effects. Research in this tradition tends to focus mainly on two outcome variables: turnout and party choice, even though early electoral research, starting from The American Voter (Campbell et al., Reference Campbell1960) treated them as two aspects of the same social and political processes.Footnote 1 Recently, several studies also focus on the determinants (or correlates) of party choice, to assess whether generational replacement contributes to patterns of realignment or dealignment. In our literature review, we will distinguish between these research traditions. We begin our discussion with research focusing on processes of political socialization.
Studies of processes of political socialization
Research on political socialization tends to focus on the development of values, attitudes and predispositions during the formative years, which are expected to be rather stable later in life. Research emphasizes the roles of the main agents of political socialization: parents, peers, teachers and (social) media. In this overview, we focus mainly on the socializing roles of parents and peers.Footnote 2
Starting with the seminal work by Jennings and Niemi (Reference Jennings and Niemi1968), early studies on political socialization focused mainly on parents as socializing agents. Thereby, using parent-child pairs for data collection, political socialization was considered as a trickle-down effect with parents transmitting their values to their children. Later, scholars also included a different perspective on political socializations with the possibility of children influencing their parents, the so-called trickle-up effect (McDevitt and Chaffee, Reference McDevitt and Chaffee2002), this being all the more important once children grow older and get access to information, for instance via social media, and on issues (e.g., climate change) that parents may not obtain or are less interested in. This research thus emphasizes that political socialization does not work in a unidirectional way from parents to kids, but also the other way around (see also Durmuşoğlu et al., Reference Durmuşoğlu2023).
Already Jennings and Niemi’s results showed a transmission effect of party identification from parents to their children, which was also confirmed by later studies showing that parents transmit their partisan preferences to their children (Achen, Reference Achen2002; Boonen, Reference Boonen2015; Coffé and Voorpostel Reference Coffé and Voorpostel2010; Hooghe and Boonen, Reference Hooghe and Boonen2015; Jennings et al., Reference Jennings2009; Kroh and Selb Reference Kroh and Selb2009; Zuckerman et al., Reference Zuckerman2007). In the European context, research on the transmission of political orientations has focused on party preferences and ideological positions. Using adolescent-parent pairs in the Netherlands, Durmuşoğlu et al. (Reference Durmuşoğlu2023) find that parent’s party preferences impact directly their children’s preferences, with political discussions again being an important moderator. They also find evidence that children influence the political preferences of their parents. Furthermore, they show the overall great importance of political ideology in political socialization processes in multi-party systems in Europe, with parent’s left-right positions indirectly impacting adolescents’ party preferences.
For other attitudes these transmission patterns could not be found, hinting already at the importance of other socializing agents such as peers. Following Jennings and Niemi’s work (Reference Jennings and Niemi1968), Tedin (Reference Tedin1980) – and also Dalton (Reference Dalton1980) – shows that parents can generally have a strong influence on political attitudes, but an important prerequisite for political influence is that political issues are communicated, and that young people are aware of the attitudes of their parents as well as of their peers (see also below). This ‘perceptual accuracy’ plays a decisive role: if young people know the attitudes of their parents or friends, the correlations between their own attitudes and those of these people are higher. Jennings et al. (Reference Jennings2009) confirm this finding: parents’ politicization and cue communication is key for the transmission of political orientation and values.
Overall, the literature shows that while parents are certainly influential, peers can also play an important role in the political socialisation of young people next to their families (Quintelier and Meeusen Reference Quintelier and Meeusen2016). This is not too surprising as teenagers and young adults become increasingly independent from their families and focus more on social interactions with their peers: Young people thus develop their social identity and political attitudes through interactions with their peers. Quintelier and Meeusen (Reference Quintelier and Meeusen2016) show that young people tend to have friends with similar political interests and behaviour. This similarity is reflected in political participation, political discussions and attitudes such as trust or prejudice towards minorities. However, important moderators in these transmissions are the homogeneity of peer networks and frequency of discussions: the more homogenous the network and the more frequent the discussion, the larger the effect on patterns of political participation and attitude similarity. Discussions and social interactions thus seem to be crucial factors in the political socialization processes through peers.
How do the parents and peers interact in the political socialization process? Tedin (Reference Tedin1980) showed that the general attitudes of young people tend to be relatively close to the political attitudes of their parents. However, when it comes to specific opinions, they are often closer to their peers. This means that although young people are influenced in an ideological direction that is very similar to that of their parents, their specific views are often closer to those of their friends. However, partisanship seems to be more strongly influenced by parental socialization than through peers. Focusing on the development of political attitudes through adolescence and early adulthood, research however shows that there could be a shift over time. Younger adolescents are influenced mostly by their parents, while at a later stage, their peers tend to become more important (e.g., Dey, Reference Dey1996), in addition to (social) media, school and voluntary organizations (Quintelier, Reference Quintelier2015).
Transmission effects do not have to work in a linear way. Partisanship, ideological predisposition, socio-economic status or gender are important moderators that research has investigated when analysing socialization effects. Focusing primarily on political ideology, Van Ditmars (Reference Van Ditmars2023) using household panel data for Switzerland and Germany shows similar patterns for the transmission of left-right ideology: the ideological positioning is highly similar between parents and children. However, the results are more heterogenous once she takes gender into account. Young women or daughters are less likely to take over the ideological positions of their parents but place themselves more likely to the left regardless of their parents’ ideology. Political socialization processes on ideological transmissions are thus strongly mediated by gender, also hinting at a stronger impact of societal transformations on young women’s ideological positions. Van Ditmars (Reference Van Ditmars2023) describes this as a gender-generation gap with daughters being ideologically socialized more independently than sons, which might have long-term effects on gender realignment.
Ares and Van Ditmars (Reference Ares and van Ditmars2023) focus on changes in the socio-economic status of parents and their offspring. Using household panel data for the UK, Switzerland and Germany, they again show that parental class affiliation matters for their children’s party preferences, and here in particular the association between parents’ working class belonging and children’s likelihood to vote for a left party, even though the latter being part of the middle class as adults. The authors argue that this transmission frame may cease to work for the next generations with their parents then being middle class resulting in lower support for left-wing parties in the future. However, this long-term prediction does not take into account changes in the policies of left parties, which could make them more attractive to middle class voters.
Generational differences, life cycles and political change
How can political socialization be related to political change? And to which extent is this age or cohort related? Most research in political socialization and political change focuses on changes in turnout and party choice, to which we turn next.
Turnout and its determinants
The decline in electoral turnout is often connected to political socialization research. Konzelmann et al. (Reference Konzelmann2012) explore whether life cycle and generational differences are related to declining turnout rates. Thereby, the life-cycle effect states that people participate more frequently in elections as they grow older because they familiarise themselves more with the political system and go to the polls out of habit. Meanwhile the cohort effect takes into consideration that socialisation in different historical periods shapes the voting behaviour of a generation. Turnout of future pensioners might be different compared to today’s pensioners because they have internalised different political attitudes and patterns of behaviour. For instance, when younger generations grow older, they will probably be less likely to turn out than earlier generations when they were at the same age. This is generally attributed to changed norms in relation to political engagement during the most impressionable years. In other words, today’s norms will affect their turnout behaviour in the future.
Based on registry data,Footnote 3 Konzelmann et al. (Reference Konzelmann2012) find for Germany that the decline in voter turnout in recent decades is mainly due to negative period effects – these are political events that influence the voting behaviour of all age groups – and only partly due to cohort effects. However, the difference in voter turnout between generations is likely to narrow in the coming decades, because the generations that are now being replaced are the most different from those entering later. The newer generations that are now entering the electorate are more similar to the generations just before them.
Other studies provide evidence of larger cohort effects. Based on life-course data from the Max-Planck-Institute, Becker (Reference Becker2002) demonstrates that cohort effects explain most of the declining turnout rates during the period 1953-1987. Based on pooled cross-sectional data, similar findings are presented for national elections in Canada (Blais et al., Reference Blais2004), Finland (Wass Reference Wass2007), and Sweden (Persson et al., Reference Persson2013). Their APC-models show that declining turnout is largely driven by cohort effects, with younger generations turning out less and less. Next to generation-specific differences in political attitudes and experiences, Persson et al. (Reference Persson2013) emphasise both the importance of institutional and individual factors: A decline in party membership and the increase in party fragmentation have contributed to younger generations voting less. Above all, the fragmentation of the party system makes it more difficult for younger voters to make clear decisions, which leads to lower voter turnout. This also fits Dassonneville’s (Reference Dassonneville2013) conclusion that it is mainly the political supply that leads younger generations to behave differently.
Drawing on a large data set that aggregates all post-1945 democratic elections data until 2017 (in total 1,421 legislative and presidential elections) with the percentage of registered voters casting a vote as dependent variable, also Kostelka and Blais (Reference Kostelka and Blais2021) show that the decline in turnout is mainly due to cohort effects. This finding is supported by the in-depth analysis of four large-scale longitudinal datasets where turnout is measured at the individual level, namely the Comparative Study of Electoral Systems, the British Election Study, the American National Election Study and the Canadian Election Study. The empirical evidence they draw on is that the global decline in voter turnout is due to generational replacement, with economic development affecting citizen’s sense of civic duty.
Most APC-studies of turnout focus on national elections, but several studies also focus on European Parliament (EP) elections.Footnote 4 Turnout was always lower at EP elections compared to national elections due to its second-national-election nature, but in relative terms, decreased even more than in national elections over time: from 66% in the first election in 1979 down to 46% in 2009. Bhatti and Hansen (Reference Bhatti and Hansen2012) provide evidence of substantial cohort effects, which partially drive this downward trend. Those born after 1970 show the lowest turnout rates due to lower interest in politics, elections in general and EP elections in particular – thus, EP elections being of particular importance to observe cohort effects in turnout. In a more recent study, Li (Reference Li2025) focuses on the increase in turnout in the 2019 EP elections which rose to its highest level in two decades. This development was in contrast to the expectations of researchers who had expected that voter turnout would continue to fall. The increase in turnout among younger voters was particularly striking. Using an APC-model, Li finds that while voter turnout in national elections continued to decline in many European countries, participation in EP elections is increasing among younger generations. This indicates that the younger generations increasingly see EP elections as important as national elections and do not base their decision to vote solely on national issues, potentially leading to a decreasing gap between turnout in national and EP elections amongst this age group. Li concludes that this strengthens the legitimacy of the EP elections and that young voters play a key role in the development of the democratic process in Europe. Since the effects are small and turnout at a young age is relatively unpredictable, it would be useful to study this in the context of the 2024 EP elections to assess the robustness of the reversed cohort effects amongst national and EP elections, which Li observed in data from 2019.
In these various APC-models of turnout, life-cycle effects (age) tend to be substantially smaller than generational differences. However, several studies do find evidence of a relatively small curvilinear pattern, with citizens around the age of 50 featuring the highest participation rates, and lower turnout rates amongst younger and older citizens. This life-cycle pattern was observed in national elections in Denmark, Finland and the US (Bhatti et al., Reference Bhatti2012),Footnote 5 and in EP elections (Bhatti and Hansen, Reference Bhatti and Hansen2012). Also, when studying first-time voters, evidence is found for a small age effect. Voters who have just reached the age at which they are allowed to vote – this is 18 years in most countries and 16 years in some – are slightly more likely to turnout to vote as older first-time voters (Bhatti et al., Reference Bhatti2012; Aichholzer and Kritzinger, Reference Aichholzer, Kritzinger, Eichhorn and Bergh2020).
An important factor driving turnout is habituation. Dinas (Reference Dinas2012) shows that early voting experiences have a strong and long-lasting influence on voting behaviour later in life. People who have already voted in an election at a young age are also more likely to turn out in later elections. This effect is particularly pronounced if the first election in which one participates is a presidential election, as this attracts greater attention and mobilization effects compared to less important elections (such as congressional elections) (e.g., Dinas et al., Reference Dinas2024). If the first election is a second-order election (such as an election for the EP), this depresses turnout later in life (Franklin and Hobolt, Reference Franklin and Hobolt2011).Footnote 6 At the same time, Dinas (Reference Dinas2012) shows that the influence of early electoral experiences diminishes over time but gets stronger again at certain points in life. Thus, not bringing younger generations to the polls can have long-lasting effects as citizens are not growing into a voting habit with turnout declining as generations with more habitual non-voters are slowly replacing generations with more habitual voters.
Overall, these results show that voter turnout is not just a question of the individual life course, but it is also strongly influenced by the historical, economic and social conditions of the respective generation, and that the general decline in voter turnout depends more on deeper, generational changes in values and attitudes towards political participation. Young people feel less obliged to vote (lower civic duty) (Kostelka and Blais, Reference Kostelka and Blais2021) and are increasingly sceptical about the political effectiveness of their votes (Blais and Rubenson, Reference Blais and Rubenson2013), show higher political alienation (Valgardsson, Reference Valgardsson2019) and lower political knowledge, the latter being an important factor for turnout (Stockemer and Rocher, Reference Stockemer and Rocher2017). Furthermore, a growing number of elections and referendums (e.g., regional and EP elections, referendums) leads to ‘election fatigue’ and reduces the motivation to participate regularly in elections (Kostelka and Blais, Reference Kostelka and Blais2021).
Party choice
Political socialization has been a central topic of research into party choice for more than sixty years. Age differences in party choice may be the result of life-cycle effects and/or generational differences.
As mentioned previously, in one of the most influential studies in the field, Campbell et al. (Reference Campbell1960) proposed that people would develop a sense of party identification (party ID) during their formative years, which would remain stable later in life. In the US context, party ID was and still is the strongest determinant of party choice, certainly among those who (strongly) identify with one of the parties. These people will almost always vote for the candidate of the party they identify with. Yet, Campbell et al. (Reference Campbell1960) also observed some clear age-effects in the sense that partisan attachments increased with age. In later work, Converse (Reference Converse1969) proposed that it was not age as such, but the number of years of being partisan that strengthens feelings of partisanship. There would be a self-reinforcing mechanism at work, whereby feelings of partisanship are being reinforced through being partisan. Building upon insights from social psychology and Fiorina’s (Reference Fiorina1981) model of retrospective voting, Bartels (Reference Bartels2002) proposes that people evaluate political events through the lens of their partisan identities, so that these identities are reconfirmed and strengthened.
A related theory is proposed by Dinas (Reference Dinas2014), who argues that partisan attitudes are the result of repeatedly voting for the same party. After voting for a party, people tend to have a more positive view of that party, which helps to justify their decision. In this way, partisanship is strengthened by voting for that party. Research in these traditions do not predict clear age effects in party choice. Some voters develop feelings of partisanship at a relatively young age, and these feelings become stronger over time resulting in stability at the individual level in party choice. Others do not develop such strong feelings of partisanship, and their party choices will be more volatile. Yet, this model does not make predictions about shifts in party support at the aggregate level as a result of aging.
However, there are theoretical reasons to expect that people will tend to become more conservative as they grow older. Different life phases are associated with different material conditions. In different stages of their lives, people have a career, get married, buy a house, get a mortgage, etc. This includes that later in one’s life, one tends to have a better income, more savings, and, to the extent that these material circumstances are related to political attitudes, political attitudes may thus change systematically more to the economic right as one grows older (e.g., Binstock and Quadagno, Reference Binstock and Quadagno2001; Tilley and Evans, Reference Tilley and Evans2014). As a result, people might become more conservative in their political attitudes, and consequently, older people would then be more likely to support conservative parties.Footnote 7
Even though there are theoretical reasons to expect life-cycle effects in party choice, there are not many studies that have tested the hypothesis that people become more likely to shift to right-wing and/or conservative parties as they age. Goerres (Reference Goerres2008) tested the hypothesis in Germany (1977–2002) and Britain (1964–2001) and found no evidence that voters switch to ‘economically conservative’ parties as they age. However, Tilley and Evans (Reference Tilley and Evans2014) tested the same hypothesis in a combined APC and panel design and find that British voters are indeed more likely to support the Conservatives when they age. Also, Peterson et al. (Reference Peterson2020) analyse a rather unique panel survey of American high school seniors and their parents, who were interviewed in 1965 and re-interviewed in 1997. Their study shows that there is generally much stability on all of the indicators. However, despite the overall stability, there were also some statistically significant shifts, in particular when it comes to party ID and ideological self-placement. As people grow older, they were more likely to switch in a conservative rather than in a liberal direction on these two indicators.
Geys et al. (Reference Geys2022) present similar findings, based on a Norwegian rolling panel running from 1977 to 2017. While there is also much stability in partisan preferences, they do find that older citizens are significantly more likely to switch to a party that is more to the right than to a party that is more to the left, while young voters are more likely to shift in the other direction. A similar result is reported by Rekker (Reference Rekker2024), who conducts a more traditional APC-analysis on a pooled data set of 258 election studies from 21 countries. He also finds small life-cycle effects, indicating that people over the age of 65 are somewhat more likely to vote for a conservative party. Based on European Social Survey data of 12 European countries, Lichtin et al. (Reference Lichtin2023) also find an age effect regarding support for Green parties. Controlling for generational differences, their APC-analyses show that the oldest age group is significantly less likely to support the Greens than younger age groups.
If we summarize the findings regarding life cycles and party choice, the effects seem to be quite small, and especially those who have repeatedly voted for a party tend to become more partisan over time (e.g., Dinas, Reference Dinas2014). The overall story is thus one of stability and stabilization, rather than change and life cycles. However, small life-cycle effects are nevertheless observed, and these tend to support the notion that, on average, people become (slightly) more conservative as they grow older.
Two implications follow from the fact that political views and feelings of partisanship tend to stabilize at the individual level over time. First, socializing experiences during the formative years are important.Footnote 8 Second, with life-cycle effects being small and age groups differing from each other substantially, these are indicative of generational differences. This is indeed confirmed by APC-models of party choice, which tend to show substantially larger generational differences than life-cycle effects.
In the largest study of this type, Rekker (Reference Rekker2024) shows that the steady decline of Christian Democratic parties over the past decades is largely due to generational replacement. Their support is large among the oldest generations which are being gradually replaced by generations among whom support for Christian Democratic parties is low. Conservative and Social Democratic parties are facing similar challenges, while Green, liberal and more radical left parties might gain with the recent generations being most supportive of Green parties (e.g., Lichtin et al., Reference Lichtin2023). Findings for far-right parties are more diffuse, and they tend to differ per country (Rekker, Reference Rekker2024). However, in his study of ten countries, Mitteregger (Reference Mitteregger2025) shows that if we zoom in on the political competition both within the right block and the left block, younger generations who identify with the left are more likely than older generations to support the Greens, while older generations on the left are more likely to support Social Democratic parties. Younger voters who see themselves as right-wing tend to be more likely to support the far right, compared to older voters, who are more likely to vote for more traditional right-wing parties, in particular Conservative and Christian Democratic parties. A complicating factor in these types of studies of the far-right is that some of these parties have formed quite recently, so that it is problematic to include them in APC-models. Also, to disentangle the impact of age and generation, APC-models must assume that age effects are stable over time and across generations. This may not always be a valid assumption, when the youngest voters of today are sensitive to the messages of far-right influencers on social media. Van der Brug et al. (Reference Van der Brug2026) find a weak but significant negative age effect on support for the far-right in the EP elections of 2024, which did not yet exist in 2019. Changes might be occurring among the youngest age group, which APC-studies cannot yet pick up. Another complicating factor is that young voters are generally more likely to vote for newly founded parties than older voters, because they have not yet formed strong partisan identities (Rekker, Reference Rekker2022).
In addition to these large scale, cross-country comparative APC-studies of party choice, several case studies offering long time series of individual countries have also been published. Yet, when studying party choice over a longer period of time, the dependent variable poses challenges. If new parties enter, or if existing parties merge or disappear, the dependent variable changes. Even though valid research designs exist to study changes in electoral behaviour when the dependent variable itself changes over time (e.g., Van der Brug, Reference Van der Brug2010; Van der Brug and Rekker, Reference Van der Brug and Rekker2021; Van der Eijk et al., Reference Van der Eijk2006),Footnote 9 APC-studies of party choice are clearly more straightforward if the country has a stable party system that is dominated by a limited number of large parties. This may be the reason why these single country studies of party choice tend to focus largely on the US, the UK and Germany.
Studies in the US context have observed substantial generational differences in the popularity of the two parties. The Democratic Party tends to be more popular among Millennials and Gen Z than the Republican Party (e.g., Deckman, Reference Deckman2024; Fisher, Reference Fisher2018; Norris and Inglehart, Reference Norris and Inglehart2019; Twenge, Reference Twenge2023). This is partially the result of differences in the composition of these generations in terms of religion, education and ethnic diversity. The youngest generations tend to be more secular, ethnically diverse and better educated, which makes them more liberal on all kinds of issues, while at the same time the Republican Party has been increasingly dominated by its most right-wing and conservative faction, since the rise of the Tea Party and Trump. If the partisan preferences of the different younger generations remain stable, it would thus become increasingly difficult for the Republican Party to win the popular vote in presidential elections.
Studies in the UK have also shown quite substantial generational differences in party choice, yet not as large as in the case of the US. Tilley and Evans (Reference Tilley and Evans2014) show that generations socialized in periods of Conservative rule, are more likely to vote Conservative later in life than generations socialized when Labour was in office. Focusing on political attitudes, rather than party choice, Grasso et al. (Reference Grasso2019) find similar generational differences between ‘Thatcher’s children and Blair’s babies’, as they jokingly call them. German APC-studies of generational differences in party choice, show declining support for the two mainstream parties SPD and CDU/CSU among younger generations paired with a higher popularity of the Greens (Goerres, Reference Goerres2008; Steiner, Reference Steiner2023).
APC-studies of electoral behaviour in other countries focus almost always on dependent variables that are more straightforward, such as turnout (see the previous section), or electoral volatility (e.g., Dassonneville, Reference Dassonneville2013). An exception is a recent study of Lisi et al. (Reference Lisi2021) in four Southern European countries in the years since the financial crisis in 2008. Focusing on support for the mainstream parties, they find no evidence of generational differences, nor life-cycle effects. They detect mainly period effects, namely the effects of the financial crisis per se. Yet, this may be the result of the specific conditions (the economic crisis) and the short time series, which hinders the possibility to disentangle age, period and cohort effects.
To summarize, APC-models of party choice find evidence of small life-cycle effects – as do studies focusing on other dependent variables. Across the board, young people are more likely to support progressive or liberal parties, and people tend to vote slightly more conservative as they grow older. In aging societies, this life-cycle effect will contribute somewhat to more conservatism, but the effects are small. In most countries, generational replacement contributes much more to the electoral fortunes of parties than the effect of aging. As the popularity of parties within generations still shows some footprints of their popularity when these generations were in their formative years, these changes are very gradual but also persistent. The traditional mainstream parties, especially the Christian Democrats and (to a lesser extent) the Social Democrats, are affected. Findings for the Far Right are, however, less consistent across countries. Some of these new parties are relatively popular among younger generations, while other parties tend to attract relatively more older voters. The pattern is still somewhat unclear, and a complicating factor is that, all other things equal, first-time voters are more likely to vote for new parties.
Determinants of party choice and political socialization
Several scholars have argued that there has been a realignment across Western Europe in the past decades, which implies that (socio-cultural) issues like immigration and European integration have become increasingly important in structuring partisan preferences (e.g., Bornschier Reference Bornschier2010; Ford and Jennings Reference Ford and Jennings2020; Kriesi et al., Reference Kriesi2008; Stubager Reference Stubager2013). Some scholars have suggested that there are theoretical reasons to expect that processes of realignment have a generational component (e.g., Franklin, Reference Franklin Mark1992; Gougou and Mayer, Reference Gougou and Mayer2013; Van der Brug, Reference Van der Brug2010; Wagner and Kritzinger, Reference Wagner and Kritzinger2012; Walczak et al., Reference Walczak2012). The theoretical idea is that young people near the age at which they are allowed to vote, learn about the party system, which is structured around the conflict dimensions that are salient at that specific point in time. If the most important conflicts are about issues such as social housing and unemployment benefits, and if the party system is structured around these socio-economic issues, voters will familiarize themselves with this structure of the party system. Their positions on socio-economic policies will then become important determinants of their party preferences. Since we know that experiences in these formative years often leave a lasting impression on behaviours, attitudes and perceptions, these determinants of party preferences can be expected to remain important later in life. This theoretical argument is much related to the US-based literature on issue evolution, which describes how new issues emerge on the political agenda and how they become embedded in the dominant conflict dimension that separates the parties (Carmines and Stimson, Reference Carmines and Stimson1981). In this model, new voters are the ‘most likely agents’ of issue evolution, because they will gradually replace older generations who were not socialized with these newer issues (Carmines and Stimson, Reference Carmines and Stimson1981; see also Wagner and Kritzinger, Reference Wagner and Kritzinger2012).
Thus, we would expect that traditional cleavages such as social class and religion remain to be important for the oldest generation who were politically socialized in the years of frozen cleavages, but much less so for the newer generations. Studies of generational differences in the predictors of party choice found support for this idea. Franklin (Reference Franklin Mark1992) showed that the decline of cleavage voting had a clear generational component, even though social cleavages also became smaller predictors of party choice over time within each generation. Van der Brug (Reference Van der Brug2010) confirmed this and showed that left-right was the strongest predictor of party choice for the generation socialized in the 1970s. Walczak et al. (Reference Walczak2012) had similar findings but found no support for their hypothesis that short-term predictors of the vote would have the strongest effects for the youngest generation.
Gougou and Mayer (Reference Gougou and Mayer2013) use insights from political socialization to explain the seemingly contradictory observation that in some countries the youngest generation is more likely to support the far-right than older generations, even though they are more liberal in terms of their attitudes on the core issues of the far-right: anti-immigration and euroscepticism. The explanation is that these issues are less important for older generations, which makes them unlikely to vote for a party campaigning on this issue, while attitudes on immigration are a more important determinant of party choice among the younger voters. Even though the group of younger citizens with anti-immigration attitudes is relatively small compared to older citizens, they are more likely to vote for the far right because they base their party choice more on this new issue.
Similar studies have been published in the US, where scholars tend to use the term partisan sorting, which suggest a causal effect of partisanship onto policy preferences. Stoker and Jennings (Reference Stoker and Jennings2008) show that the political attitudes of younger citizens are more strongly aligned with their party ID than those of older generations, especially on issues that emerged recently. Levendusky (Reference Levendusky2009) demonstrates that younger cohorts are better sorted than older ones, while Twenge et al. (Reference Twenge2016) find that the correlation between partisanship and ideological self-placement is stronger for younger citizens than for older ones. Yet, the main drawback of these studies is that they do not disentangle age, period and cohort differences.
Recently, several APC-studies have been published, which do focus on the question whether generational differences exist in the correlates of party preferences. In US-based APC-studies, scholars have found different patterns. While Phillips (Reference Phillips2022) finds quite limited generational differences in the correlation between party identification and ideological self-placement, Jocker et al. (Reference Jocker2024) focusing on a larger number of issues find that the generation of Baby Boomers are most strongly sorted on most issues, more so than the generations before them, but also more than the most recent generations. In part, this is a consequence of the fact that a larger group of voters in the youngest generation does not identify as belonging to any party.
In their study of the Netherlands, Van der Brug and Rekker (Reference Van der Brug and Rekker2021) find that dealignment focusing on the weakening effect of religion and social class on party choice coexists with realignment across generations. This means that education and positions on new issues are stronger determinants of party preferences for younger generations than for older ones. Another study which suggests that generational replacement contributes to patterns of realignment comes from Haffert and Mitteregger (Reference Haffert and Mitteregger2023), who demonstrate that the urban-rural divide in voting behaviour is larger among the younger generation than among the older.
Similarly, there are several studies focusing on generational differences in gender gaps in voting behaviour and ideology (e.g., Dassonneville, Reference Dassonneville2021; Harsgor, Reference Harsgor2018; Inglehart and Norris, Reference Inglehart and Norris2003; Off et al., Reference Off2025; Shorrocks, Reference Shorrocks2018; Van Ditmars, Reference Van Ditmars2023). Some 50 years ago, women were more likely than men to support Conservative and Christian democratic parties. Nowadays women are more likely to vote for Green and left-wing parties and less likely to support the far-right. It seems plausible that there is a generational component underneath these changes, since recent generations were socialised in a period with different gender roles than previous one. Yet, the jury is still out. Shorrocks (Reference Shorrocks2018) presents evidence of a generational component in the gender gap in left-right self-placement. Meanwhile, Dassonneville (Reference Dassonneville2021) points out that the differences exist mostly when comparing current generations to the oldest ones, but that the newest generations do not differ significantly from those before them. Milosav et al. (Reference Milosav2026) and Off et al. (Reference Off2025) show that the gender gap in far-right support is largest among the youngest age group. Also, Grasso and Shorrocks (Reference Grasso and Shorrocks2026) find that women are more supportive of government intervention than men (which makes them more prone to voting for the left), but they find no significant differences between generations in the magnitude of these gender gaps. An interesting perspective comes from Harsgor (Reference Harsgor2018), who shows that the recent increase in the gender gap in voting in the US is largely a consequence of changes in the composition of the electorate. The gender gap has been rather stable and quite sizable between black men and women, while the relative share of black citizens has increased in the US.
Focusing more on issues and ideology, Jocker et al. (Reference Jocker2025) conduct an APC-analysis of issue alignment in 17 European countries. Issue alignment refers to the strength of the relationship between vote choice and issue preferences. They find that younger generations are more strongly aligned on some of the new issues, as well as on the general left-right dimension. Yet, the patterns they observe show quite a lot of variation across different countries. Dassonneville and McAllister (Reference Dassonneville and McAllister2026) explain such country differences in the relationship between age and party choice by characteristics of the party system. If parties on the left and right compete on the basis of socio-cultural issues, young voters are more likely to support left-wing parties. If they compete on socio-economic issues, age and party choice are unrelated. An open question, however, is whether younger voters are more strongly aligned on left-right in those countries where parties compete more on socio-cultural issues.
The idea that the ideological left-right dimension is not meaningful for young voters thus seems to be incorrect. Importantly, within the youngest generation, left-right positions are more strongly correlated with socio-cultural issues than within older generations (e.g., Mitteregger, Reference Mitteregger2025; Rekker, Reference Rekker2016; Steiner, Reference Steiner2023). However, perceptions of left-right positions of parties as well as the substantive understandings of the meaning of left-right are remarkably similar between generations (Durmuşoğlu and Van der Brug, Reference Durmuşoğlu and van Der Brug2025). So, it seems that these differences between generations and countries are mainly driven by the salience of socio-cultural issues for different age groups and in a party system.
Conclusions and looking ahead
Over the recent decade, research on political socialization and generational differences has been booming in the social sciences. In this review article, we reflected on such research in the field of electoral processes, which consists of two research traditions. The first research tradition focuses on generational differences in political attitudes and behaviours. Scholars in this research tradition often try to disentangle Age, Period and Cohort effects. The second research tradition focuses on processes of political socialization during adolescence and early adulthood, and on the role of socializing agents such as peers and parents. In this concluding section we will not try to summarize our reflections on this literature, but instead we want to emphasize four general points that we feel are relevant for future research in this field.
First, more knowledge might be obtained if different research traditions would develop less in isolation. One of the common assumptions in APC-analyses on turnout and party choice is that the most impressionable years are between 15 and 25, and there is much evidence supporting this (e.g., Bartels and Jackman, Reference Bartels and Jackman2014; Rekker et al., Reference Rekker2019; Schuman and Rodgers, Reference Schuman and Rodgers2004). Yet, studies of political socialization suggest that some relevant political attitudes are already formed at an earlier age. For example, attitudinal differences between educational levels exist already before puberty (e.g., Kuhn et al., Reference Kuhn2021). It seems worthwhile to study children already at an age where they have little knowledge of political parties and the main conflict divisions between parties. Obviously, such studies should not focus on political parties per se, or on complex issues, but it would be fruitful to study the formation of basic values and ideologies (for an overview, see e.g., Abendschön, Reference Abendschön2010; Sapiro, Reference Sapiro2004; Van Deth et al., Reference Van Deth2011). Interesting work in this field is done by developmental psychologists (e.g., Brown, Reference Brown2011; Gampe and Daum, Reference Gampe and Daum2018; Reifen-Tagar and Cimpian, Reference Reifen-Tagar and Cimpian2022), but it does not yet seem to have fully landed in political science.
Along the same lines, secondly, we note that studies of turnout and party choice have developed as separate research traditions in the field of electoral research. This is not just true in the field of generations, but this is more generally the case. However, foundational studies in the field, like The American Voter (Campbell et al., Reference Campbell1960), treated turnout and party choice as interconnected and partially driven by the same underlying societal processes. The connectedness of party choice and turnout seems obvious. A person who finds all parties on offer unattractive will not be very motivated to turn out. If the parties on offer appeal less to young voters than to older ones, that could be an explanation for low turnout among these voters in a specific election. It is rare to see attempts to integrate turnout and party choice in a single model (but see Franklin, Reference Franklin2022, Reference Franklin2024), and we are not aware of any such studies in the field of generations. We believe, however, that this integration would offer various opportunities for future research in general and political socialization in particular.
A third point is methodological but follows from the substantive discussion of how scholars define and operationalize generations and patterns in life cycles. As argued before, any classification of generations has some element of arbitrariness, because respondents that are classified as belonging to different birth cohorts may share many of the same formative experiences, albeit at a slightly different age. Yet, for political orientations later in life it may not matter much whether you were 16 or 18 when the Berlin Wall fell. Because of the somewhat arbitrary nature of pre-defined classifications, more inductive estimation methods, such as Generalized Additive Modelling (GAM), are attractive alternatives (e.g., Grasso, Reference Grasso2014; Neundorf, Reference Neundorf2010; Valgardsson, Reference Valgardsson2024). These methods enable users to estimate smooth curves of cohort or age, without having to predefine categories. Even though GAM’s have clear advantages, there is reason for caution. Anyone who has applied more traditional approaches to estimate APC-models will know that multicollinearity statistics can easily go through the roof when modelling age as a linear function, and certainly when also including a quadratic term. Scholars who use Bayesian approaches such as GAM tend to be not so worried about multicollinearity. However, when the estimated effects of age and birth cohort run in opposite directions and multicollinearity is extremely high, one may start to worry about the robustness of these estimations. One should be careful not to overinterpret such results, and we would urge researchers to always apply more traditional APC models to test the robustness of patterns found with GAM.
A final topic for future research concerns one of the central assumptions that APC-models make. To disentangle the effects of life cycles and generations, one must assume that life- cycle effects are the same for each generation. Without that assumption, the models cannot be identified. However, one might doubt whether this central assumption is always valid. For many people, getting retired around the age of 65 probably had quite different implication 60 years ago than it does today. Compared to previous generations, the most recent generations live with their parents longer, they study longer, get children later and start working later (e.g., Bialik and Fry, Reference Bialik and Fry2019; Schizzerotto and Lucchini, Reference Schizzerotto and Lucchini2004). These are important life events, which may well have an impact on patterns of electoral behaviour including both turnout and party choice.
Recently, scholars have attempted to assess whether indeed life cycle-effects vary across generations. An interesting contribution comes from Serra (Reference Serra2023) who shows, on the basis of panel data, that there are indeed differences between generations in the life-cycle effects on support for the British Conservative Party. Another indication of differences in life-cycle effects across generations can be found in the US using data from the American National Election Studies (ANES). In the 1950s many first-time voters entered the electorate as an identifier with one of the two parties. Many inherited the party ID of their parents, but the number of identifiers hardly increased in this generation as they grew older. Among later generations, fewer first-time voters entered the electorate identifying with a party, but many formed their party ID later, as a consequence of repeatedly voting for the same party (e.g., Dinas, Reference Dinas2014; Van der Brug and Franklin, Reference Van der Brug and Franklin2018). It seems that more work may be needed into the differences in life cycle patterns between generations and the approach by Kostelka and Blais (Reference Kostelka and Blais2021) may be a possibility to study political socialization across time and context in a comparative way. While this clearly opens new opportunities for research, it is also a challenge for APC-models which assume that life-cycle effects are the same in each generation.
Overall, this review article shows that to understand and explain political and social change, research in political socializations provides important input. Yet, there are new avenues of research to explore with new datasets and methodologies becoming available. More progress can be made by integrating insights from different subfields within political science (such as turnout and party choice), but also by learning from other disciplines (e.g., developmental psychology).
Data availability statement
Not applicable.
Acknowledgements
We are grateful to the reviewers and editors of EPSR for their constructive criticism and support.
Funding statement
Van der Brug’s work on this article was supported by a grant from the Nederlandse Organisatie voor Wetenschappelijk Onderzoek, Grant/Award Number: 406.18.RB.025.
Competing interests
There are no competing interests.
Ethical statement
This is a review article, so there are no ethical issues pertaining to this publication.
Open access
