Populist Attitudes among Teenagers: How Negative Relationships with Socialization Agents Are Linked to Populist Attitudes

While the origins and consequences of populist attitudes in adults are being studied extensively, it is still unknown when populist attitudes might emerge in a person ’ s life. Drawing on the existing literature on political socialization, we focus on populist attitudes during adolescence and explore the contributing role of negative relationships with parents, peers, and teachers. We provide the ﬁ rst comprehensive analysis of populist attitudes among a representative sample of children aged 12 to 18 (mean: 14.66 years) using a uniquedatasetgatheredthroughinterviewsconductedinschools inAustria,Germany,andSwitzerland(n=3,123).Our ﬁ ndingsreveal a strong association between perceived unjust teacher behavior and the level of populist attitudes in adolescents, while the connection of peers and parents with populist attitudes appears to be limited. Further analyses using panel data from the UK support these ﬁ ndings.

T he impact of populism on democracy has been debated heavily in recent years (Rovira Kaltwasser 2012).While some scholars consider populism as a redemptive force (Canovan 1999) that may be able to correct some representational deficits of democracy (Kriesi 2020;Manow 2020), others pointed towards the incompatibility between the populist conception of a general will and elements of liberal democracies like pluralism, freedom of speech, and political compromise (Müller 2016;Urbinati 2019).On the individual level, populist citizens often support general democratic principles, but they criticize democratic practices and the working methods in everyday politics (Pappas 2019;Zaslove et al. 2021).Thus, such "dissatisfied democrats" (Rovira Kaltwasser and Van Hauwaert 2020) could, for instance, more easily be approached by certain political entrepreneurs to support illiberal policy solutions (Wuttke, Schimpf, and Schoen 2023).
The most widespread ideational approach defines populism as a thin-centered ideology that is grounded in a Manichean perception of good versus evil, in which evil elites conspire against the pure and good people, and where the foundation of political decision-making should be based on the general will (volonté générale) of the people (Mudde 2004)."The elite" is often used as a vaguely specified empty signifier and can refer to political actors, economic leaders, journalists, or bureaucrats and the like (Jagers and Walgrave 2007). 1  On the individual level, populist attitudes constitute a non-compensatory concept that consists of multiple components, such as anti-elitism, people-centrism, and a Manichean worldview (Castanho Silva et al. 2018).Citizens are therefore considered as populist only if they endorse all three dimensions at the same time (Wuttke, Schimpf, and Schoen 2020).
Some scholars argue that populist attitudes can be conceived as dispositions that remain dormant until they become activated by external triggers like corruption or misrepresentation (Hawkins, Rovira Kaltwasser, and Andreadis 2020)-hence, they might be relatively stable over time (Ardag et al. 2020).Other scholars showed however, that there may be variations in populist attitudes as response to external stimuli (Rhodes-Purdy, Navarre, and Utych 2021; Rico, Guinjoan, and Anduiza 2020).Most recent evidence highlights though, that both may be true for different parts of the population (Schimpf, Wuttke, and Schoen 2023).
It is however unclear when and how populist attitudes are exactly formed.Based on the existing literature on populism, it is often argued that the development of populist attitudes results from an interaction of the individual with the political process.Thus, intentional policy failures of elites, for instance through a mismatch between responsiveness and responsibility of political actors, may cause citizens to form populist attitudes (Hawkins, Rovira Kaltwasser, andAndreadis 2018, 2020).Following this, populist attitudes should begin to form at the time adolescents start to interact with the political system, that is, around voting age.
In a different view, the primacy model of political socialization suggests that political attitudes are malleable until early adulthood and become increasingly stable afterwards (see Deth, Abendschön, and Vollmar 2011;Dinas 2013;Stoker and Jennings 2008).It is therefore not surprising that we find already large gaps in political involvement at very young age (Abendschön and Tausendpfund 2017;Cesarini, Johannsson, and Oskarsson 2014;Deth, Abendschön, and Vollmar 2011;Jennings, Stoker, and Bowers 2009;Prior 2019).Thus, we should expect that populist attitudes can potentially develop already before children make first contact with the act of voting, that is, during their teenage years.
We argue that adolescence is a pivotal phase for receptivity to populist ideas.Adolescents go through tremendous cognitive and socio-cognitive growth throughout this period, allowing them to think about political problems in more abstract and complex ways.
During this time, the role of environmental variablesnotably the family, peers, and school-is crucial and these proximal environments are critical in forming youth political orientation because they are where young people first meet politics.But while some origins of populist attitudes like grievances or emotions are probably working in the same way for adults and adolescents, others are likely to differ.In particular, a low sense of political efficacy and a perceived lack of representationthat is, the perceived lack of will or the inability of political actors to respond to individual or social grievances, a perception of having little say in politics, and a sense of injustice-are often a central motivator for the development of populist attitudes (Castanho Silva and Wratil 2023;Geurkink et al. 2020;Hawkins, Rovira Kaltwasser, and Andreadis 2020;Spruyt, Keppens, and Van Droogenbroeck 2016).However, adolescents often lack experience with the political system that is necessary to form the perception of a representation gap.Thus, we argue that socialization agents function as entities or authorities in accordance with which adolescents can generate populist attitudes.If adolescents perceive themselves as being treated unjustly by their teachers or if they have few possibilities to bring in their opinion in their circle of friends or at home, they can also be likely to develop populist attitudes.
In this study, we investigate whether and how socialization agents contribute to the formation of populist attitudes among adolescents.In particular, we ask whether negative relationships with parents, peers, and teachers are associated with having stronger populist attitudes.The study contributes to our understanding of the formation of populist attitudes as it is the first to explore its pervasiveness and correlates among a representative sample of adolescents (n=3,123, mean age =14.66 years) in multiple countries using a unique dataset gathered through interviews conducted in schools throughout Austria, Germany, and Switzerland.The results indicate that perceived unjust teacher behavior is strongly and positively associated with populist attitudes among children, whereas the role of peers is more limited and non-existent for parents.Furthermore, we find that the association between teacher perception and populist attitudes might increase with age.Additional evidence from UK panel data supports these findings.

Socialization Processes and the Formation of Populist Attitudes
In a developmental perspective, the teenage years are characterized by biological, emotional, and social changes that can challenge existing structures and identities (Hurrelmann and Quenzel 2019; National Academies of Sciences, Engineering, and Medicine 2019).Without having had any previous experience with the political system themselves, children and adolescents often already have attitudes toward social and political issues.According to Weisberg (1980), three different models of political socialization can be distinguished, depending on the age phase.First, the primacy model, according to which political values are imparted as early as infancy.Here, fundamental political orientations and values are learned in the family, which can structure later political action and thinking.Second, the intermediate-period model, which takes place during adolescence.As cognitive development progresses, political concepts are now comprehended and, for example, an understanding of political engagement is developed.Third, the recency model, which holds that even after adolescence, cumulative experiences with politics are made and political thinking and behavior can change.Following this intermediate model of political socialization, children are able to develop affective connections to socio-political objects early in life, for example in the form of party identification (Campbell et al. 1960), whereas cognitive aspects mature during later stages of socialization.
In the absence of the opportunity to draw on their own experience, political socialization is the primary process for developing political beliefs and behaviors via social interactions and collaboration with others, such as parents, peers, and adult role models (Flanagan 2013).Thus, political socialization can be understood as the individual process of learning patterns corresponding to one's own societal position as mediated through various societal agents (Hyman 1959).
The teenage years and young adulthood are further characterized as "impressionable years" meaning that adolescents' (political) attitudes are particularly malleable through personal experiences or political events due to cognitive changes and development and the search for identity and community (Deth, Abendschön, and Vollmar 2011;Dinas 2013;Ghitza, Gelmand, and Auerbach 2023;Stoker and Jennings 2008).While adolescents often find themselves in a trade-off between the individual development of a persona and, for example, the need for attachment to their peers, a pluralist society does not provide instructions how to cope with such conflict.Thus, we expect that the groundwork for the development of populist attitudes-which are characterized, among others, by a clearly structured black-and-white worldview (Manicheanism)-can already be laid during adolescence and thus be influenced by political socialization processes.Since political attitudes become more stable over the lifecourse (Bacovsky and Fitzgerald 2023;Denny and Doyle 2009;Firebaugh and Chen 1995;Plutzer 2002;Russo and Stattin 2017), it is particularly important to uncover the mechanisms that lead to the development of populist attitudes in early life.
Previous research has identified three main factors that contribute towards the development of populist attitudes.First, (subjective) status loss, a poor socio-economic position, and relative deprivation have long been linked to stronger support for populist parties (for example, Gidron and Hall 2017;Pettigrew 2017).Although these mechanisms have been mainly reported for adults (Spruyt, Keppens, and Van Droogenbroeck 2016), we believe that they also hold for adolescents, as there is strong evidence for the link between parental socio-economic status and adolescents' political attitudes (Akee et al. 2020;Holbein 2017).Second, emotions like anxiety and anger are seen as not only shaping political attitudes in general (Marcus 2000(Marcus , 2022;;Marcus et al. 2019) but also specifically shaping the support for populist parties (Rico, Guinjoan, and Anduiza 2017;Salmela and von Scheve 2017).Finally, a low sense of political efficacy and a perceived lack of representation are often a central motivator for the development of populist attitudes (Castanho Silva and Wratil 2023;Geurkink et al. 2020;Hawkins, Rovira Kaltwasser, and Andreadis 2020;Spruyt, Keppens, and Van Droogenbroeck 2016).In most cases this pertains to a lack of will or the inability of political actors to respond to individual or social grievances, a perception of having little say in politics, and a sense of injustice (Betz 2019).Populist tendencies can thus be put into action through poor democratic governance and purposeful policy failure by political elites.Previous research identified the major forces behind such a trend in the form of widespread systemic corruption in developing countries or growing conflicts between political elites' accountability and responsiveness in developed countries, as they grow more ideologically distant from their constituents (Hawkins et al. 2018, Hawkins, Rovira Kaltwasser, andAndreadis 2020).This frequently results in the emergence of political players who start to draw attention to the flaws in democracy as it exists now, but whose power is ultimately curbed by democratic actors through political discussion and decision-making.Populist rhetoric ultimately draws on these worries and offers clear friend-and-foe images, making it easier for vulnerable individuals to perceive themselves as part of the supposedly "good" group of the pure people and thus regain some feelings of control over their situation (for example, Fritsche et al. 2013;Greenaway et al. 2015;Hogg and Gøtzsche 2021;Widmann 2021;Wirz 2018).
However, while we believe that the influence of deprivation and emotions is quite similar among adolescents and adults, we assume that a perceived lack of political efficacy and representation develops differently.Although the connection between responsibility and responsiveness can be made for adults who regularly engage with the political process, this does not apply for adolescents who are at the verge of developing a political consciousness.In particular, adolescents often lack the political knowledge and experience necessary to form the perception of a representation gap that develops after years of experience with the political system (see further Easton and Dennis 1969;Huber, Jankowski, and Wegscheider 2023).Thus, we believe that socialization agents like parents, peers, and teachers play an important role (Koskimaa and Rapeli 2015;Quintelier 2015).If adolescents perceive themselves as being treated unjustly by their teachers or if they have few possibilities to bring in their opinion in their circle of friends or at home, they can be likely to develop populist attitudes, too.Previous research has shown, for instance, that the assessment of the fairness of teacher behavior, particularly in terms of relational and procedural justice, has an influence on the legitimacy of teachers' authority as well as the assessment of institutional authorities outside of school (Gouveia-Pereira et al. 2003).Others found that even preschoolaged kids are already attentive to procedural justice in addition to distributive justice (Grocke, Rossano, and Tomasello 2015).Hence, we focus on the influence of socialization agents on populist attitudes.
Our study is the first to assess the pervasiveness of populist attitudes among adolescents along with specific child-related correlates.To our knowledge, previous studies either surveyed young adults who are at the end or way beyond the impressionable years or they study other facets such as right-wing radicalism or extremism that were post-hoc deemed as aspects of populism.Thus, youth populist thinking, so far, is considered inconsistent and varies by context and time, especially in the context of voting behavior (Noack and Eckstein 2023).For instance, right-wing populist parties are gaining more and more votes from young people, as can be seen in the 2022 elections in both France and Italy.Relatedly, Westheimer (2019) demonstrated that in the United States that about one-quarter of young adults consider democracy a (very) poor form of government and prefer a "strong leader."In Germany, the Shell study showed that almost one-quarter of 12-to 25-year-olds are open to populist ideas and 9% are even open to national populists views based on a measure of populist attitudes that wrongly mixes anti-EU sentiments with xenophobia and Islamophobia (Shell Deutschland Holding GmbH 2019).
In summary, while populism is not exclusive to youth, research suggests that adolescence is a critical period for the susceptibility to populist ideas.During this time, adolescents experience significant cognitive and socio-cognitive development, allowing them to think about political issues in more abstract and nuanced ways.While the role of adolescents and the stability of political views in adulthood should not be disregarded, the influence of environmental factors-particularly the family, peers and school-cannot be overlooked.These proximal contexts play a crucial role in shaping youth's political positioning and are where young individuals first encounter politics (Hatemi and Ojeda 2021;Jennings, Stoker, and Bowers 2009).

The Role of Parents in the Political Socialization Process
Existing research emphasizes parents as one of the main political socialization agents.Children spend most of their time with their parents, who have a high impact on their daily lives.In this context, studies focus, for instance, on the influence of parents on the formation of party identification (Kroh and Selb 2009), political ideology (Van Ditmars 2023; Weiss 2023), right-wing extremist attitudes (Oepke 2005), and political participation (Verba, Schlozman, and Burns 2005).Parents can influence the development of their children's political orientations both through the way they treat their children and through their socioeconomic status (Jungkunz and Marx 2024;Neundorf and Smets 2017;Prior 2019).Existing studies show that parents with higher socioeconomic status are more likely to have children with higher education (Neundorf and Smets 2017).The children's level of education, in turn, influences their political interest and knowledge.In addition, parental socioeconomic status contributes to the emergence of class-specific political orientations (Jennings, Stoker, and Bowers 2009).
Intergenerational transmission by parents can be more overt or more subtle.Social learning theory (Bandura 1977) argues that parents act as role models and thus foster parent-child coherence in political attitudes and values (Gniewosz, Noack, and Buhl 2009).Existing research also shows that successful transmission in this way relies on the child's correct perception of parental values and attitudes (Knafo and Schwartz 2004).Parents exert a substantial influence on their children's political awareness and engagement.Research suggests that highly politicized parents have the potential to foster positive civic orientations, encouraging their children's active participation in politics (Beck and Jennings 1982).Successful transmission of political values occurs more frequently when families maintain a politicized environment, as consistent political signals are provided by parents (Jennings et al. 2009).Furthermore, parents as role models lead to the imitation and adoption of political behaviors and attitudes (Dryer 1998).
However, such a one-step approach ignores how the family environment and the characteristics of the parent or child influence the transmission process.Following current research, transmission occurs in two steps instead.Children must first perceive their parents' political attitudes and then decide whether to adopt or reject the perceived orientation for their own position (Hatemi and Ojeda 2021).According to this view, transmission is a function of both the parent and the child, since neither the correct perception nor the adoption of parental orientations alone reflects actual transmission.At the same time, this view also implies that transmission can bypass the child's perception and thus occur unintentionally, implicitly, or indirectly, and is overall a reciprocal process as opposed to a one-way transmission (Hatemi and Ojeda 2021).
Existing research on the conditioning factors of this transmission process shows that the transmission of opinions, attitudes, or values from parents to their children can be more subtle than assumed in social learning theory (Baumrind 1991).Central to this is the style of interaction between parent and child (Oepke 2005).Mediation plays an important role, and it is shown that a successful transmission between parent and child is promoted by authoritative education (balance between discipline and nurturance), while authoritarian education is detrimental to it (Weiss 2023).The authoritarian parenting style is characterized by a lack of warmth and high levels of demands and control, where parents have high expectations but are unresponsive to their children's needs (Kılıçkaya, Uçar, and Nazhgül 2023).Interactions with authority figures in the family, especially parents, thus play an important role.As a result, intergenerational transmission appears to be more than simply adopting parental views. 2 These attitudes are also shaped by the style of faceto-face interactions (Edwards 2004;Torney-Purta, Richardson, and Barber 2004).Strict and authoritarian parenting styles, characterized by strict rules and controls, have been shown to promote, for example, political alienation (Gniewosz, Noack, and Buhl 2009) and right-wing extremist attitudes (Oepke 2005) in adolescence.We thus assume that experiences with strict authorities in terms of negative and non-reciprocal relationships with parents undermine the development of a positive bond between the individual and the broader social world, that is, the institutions of society and the representatives of the political system.This impression of a low position in the power structure may subsequently lead to the development of populist sentiments, mainly in the sense of an anti-elitist attitude.Thus, we form the following hypothesis: H1: A negative, non-reciprocal parent-child relationship is associated with higher levels of populist attitudes.

The Role of Peers in the Political Socialization Process
Peers play a significant role in the political socialization of adolescents and young adults, as they engage in discussions on socio-political issues, share popular culture, and develop a set of values, whether they are common or opposing (Neundorf and Smets 2017).The constant interaction and presence of peers in the lives of young people provide a platform for the formation of opinions and the development of political skills (Verba, Schlozman, and Brady 1995).Through these discussions, adolescents are exposed to diverse perspectives, which contribute to their understanding of democratic principles and economic concepts such as the exchange of goods, services, and information (Quintelier 2015).
It is worth considering the diverse settings in which peer interactions occur.Peer interactions take place within various contexts, including friendship networks, classmates, and other institutional settings such as associations or clubs.Each of these settings may exert distinct influences on political socialization.For instance, peer discussions within friendship networks, characterized by close and personal connections, may foster an environment conducive to open and honest political exploration.On the other hand, interactions within institutional settings, like clubs or associations, may introduce specific group dynamics and norms that shape political attitudes in a different manner (Quintelier 2015).
Meta-analytic evidence further shows that peer (and parental) support is associated with stronger critical reflection and external political efficacy among adolescents (Heberle, Rapa, and Farago 2020).Positive peer experiences amplify this effect, fostering a more comprehensive and nuanced understanding of socio-political landscapes.On the contrary, negative peer relationships could thus result in a lower likelihood to engage in or pay attention to complex discussions and a higher receptiveness for a simplified portrayal of socio-political issues (that is, Manicheanism).
Strong ties with classmates not only foster emotional and psychological well-being but also provide a conducive environment for political exploration.Adolescents, within this secure environment, are empowered to delve into and develop their political beliefs, using these relationships as a blueprint for future interactions both within and beyond the school premises (Noack and Eckstein 2023).Furthermore, the presence of peers inherently introduces a set of social norms.These norms play a pivotal role in determining the behavior and attitudes of each individual within the group (Cochran and Brassard 1979).By understanding and reflecting upon their interactions with peers, individuals can gain valuable insights into broader societal dynamics and discern their position within this larger social structure.However, not all peer interactions are affirming.Negative treatment from peers can be internalized by individuals, leading them to perceive themselves as holding a lower position in the social hierarchy.Such perceptions might cultivate feelings of being marginalized or subordinate to others, potentially leading to resentment or skepticism towards perceived opinion leaders or authority figures.With this in mind, we assume that: H2: A negative relationship with peers is associated with higher levels of populist attitudes.
The Role of Teachers in the Political Socialization Process Young people spend a significant part of the day at school, where they gain experience over many years (Abdelzadeh, Zetterberg, and Ekman 2015).In addition to the task of teaching knowledge and skills, schools also enable interpersonal experiences.From research on street-level bureaucracy we know that state institutions serve as places where the population comes into direct contact with lived political structures and policies and that the experiences made there have an impact on political attitudes (Ariely 2013;Bruch and Soss 2018;Lipsky 1980;Shore and Tosun 2019).Individuals' treatment by authorities in these institutions affects their recognition of their own standing and value in society, subsequently influencing their long-term expectations of procedural justice by the state (Weiss and Parth 2023).Existing studies show that these interpersonal experiences in the sense of relational justice positively impact both liberal democratic orientation and trust in formal institutions (Resh and Sabbagh 2014;Pretsch and Ehrhardt-Madapathi 2018).
Experiencing fair treatment in school thus promotes liberal democratic attitudes among students, while unfair treatment can lead to the delegitimization of teachers and the school as a whole (Chory-Assad 2002; Chory-Assad and Paulsel 2004; Parth et al. 2020;Torney-Purta, Wilkenfeld, and Barber 2008).Through everyday experiences at school, young people learn a "hidden curriculum" of positions and power (Bruch and Soss 2018).Based on social learning theory, schools can thus be perceived as a kind of miniature society within which students learn sociopolitical processes on a small scale (Bandura 1977;Kiess 2022;Noack and Eckstein 2023; Wray-Lake 2019).Teachers thereby represent the primary agent of reward and punishment (Cherng 2017;Resh and Sabbagh 2014).Positive engagement with teachers, who serve as models of authority, can cultivate a sense of empowerment, civic duty, and respect for societal norm.Such constructive interactions potentially lay the foundation for adolescents to develop non-populist attitudes, fostering a sense of trust in established institutions and promoting nuanced, collaborative dialogue over simplistic, divisive rhetoric.In contrast, teachers who abuse their power can harm students' social development, particularly their attitudes towards reciprocity and society's institutions (Pretsch and Ehrhardt-Madapathi 2018).Previous research has shown that negative school experiences have detrimental effects.Bruch and Soss (2018) discovered that negative encounters with school authorities decrease political engagement and trust among young people.This highlights the influence of school experiences on shaping perceptions of democratic society (Resh and Sabbagh 2014).
Schools serve as tangible representations of the state, providing students with insights into how public institutions operate and how they can anticipate treatment from authorities.These formative experiences during childhood and early adolescence shape individuals' perceptions and can have long-lasting effects into adulthood (Bruch and Soss 2018).Therefore, equal treatment and interpersonal interactions in school are crucial for the development of civic identity (Resh and Sabbagh 2014).From research on adults, we know that populist attitudes arise from the interaction of the individual with the state (Hawkins et al. 2018;Hawkins, Rovira Kaltwasser, and Andreadis 2020).At this point, teachers represent the executive state and thus make the state tangible for young people, whereby, the behavior of individual teachers results in significant factors for political socialization (Koskimaa and Rapeli 2015;Quintelier 2015).Unfair treatment by teachers, respectively by the state, could thus foster anti-elitist attitudes.Therefore, we assume that if children perceive unfair treatment by their teachers, they may be more inclined to develop populist attitudes: H3: A negative teacher-child relationship is associated with higher levels of populist attitudes.

Research Design
We conducted a survey of adolescents aged 12 to 18 (mean: 14.66 years, SD: 1.25) in the Lake Constance area, spanning schools in Austria (n=1,523), Germany (n=356), and Switzerland (n=1,244). 3All surveys were taken in class electronically and submitted anonymously by each student online through the LimeSurvey platform. 4 The fieldwork period ranged from fall 2019 to early March 2020 in Eastern Switzerland, from March to June 2020 in Western Austria (Vorarlberg), and from September to December 2020 in Southern Germany (Baden-Württemberg). Whereas the Austrian and Swiss samples were part of larger nation-wide studies with representative sampling strategies (Quenzel and Böheim-Galehr 2021;Beck and Ha 2018), the German sample was conducted independently and could not be stratified based on representative quotas due to the ongoing COVID-19 pandemic.Overall, the sample is 52% female and 32% have a migration background. 5 Our main dependent variable, populist attitudes, was measured through a six-item version of the Castanho Silva et al. ( 2018) battery on five-point Likert scales (refer to online appendix A for question wording and summary statistics).The question battery includes two items each on anti-elitism, people-centrism, and the perception of a Manichean worldview (table 1), which showed good internal coherence, cross-national validity, and external validity across countries (Castanho Silva et al. 2020).Since populist attitudes are considered a non-compensatory concept (Wuttke, Schimpf, and Schoen 2020) whereby individuals need to hold high values on all three dimensions (i.e., anti-elitism and people-centrism and Manichean worldview) at the same time to be regarded as populist, an additive index is inadequate for operationalization.Instead, we first calculated a mean index for each dimension which ranged from zero (do not agree at all) to four (fully agree) and then multiplied all three dimensions with each other.This way, respondents who score high on populist attitudes have high values on all three dimensions (see also Jungkunz, Fahey, and Hino 2021).For robustness checks, we also provide fit statistics from confirmatory factor analyses in the online appendix (table A.2).The results indicate a good to very good fit in all countries.Furthermore, we reran all analyses using alternative aggregation methods (mean index and minimum value across dimensions) in online appendix A. The results are basically similar to the ones presented in the main text.
As for independent variables, we use the relationship with different socialization agents: parents, peers, and teachers.For parental support we use one item asking "All in all, how much of a say do you have at home?" with response options on a five-point scale from very little to very much.Since this item is somewhat less precise in capturing the concrete experiences at home, we use two items as additional robustness checks that asked about whether "My parents don't care how I do in school" and "My parents don't have time to care about my school" with response options on four-point scales from fully agree to fully disagree. 6Unfortunately, these two items were only asked in Austria and Germany which is why we present the results in the online appendices.For peer relationship, we use four items asking about how children are doing in class and their experiences with classmates: "I am treated badly by my classmates," "I am alone in the breaks," "When I make mistakes, I am made fun of by others," and "My classmates stand by me when it matters."We combined the responses from four-point Likert scales (fully agree to fully disagree) through a mean index (α = 0.653).For teachers, we use three items asking for how many of their teachers the following sentences were true: "I feel I am treated fairly," "I am graded fairly," and "Other students are treated better than I." The responses were recorded on four-point scales from (almost) none to (almost) all.We combined the three items through a mean index (α = 0.628).
Finally, we added further control variables to our models.Migration background was constructed based on either the children or their parents being born in a country outside Austria, Germany, Liechtenstein, and Switzerland.To account for different aspiration levels of schools across countries, we categorized the education level of children's schools into those that grant access to university and those that do not.Similarly, we measure parental education, as a dummy which is one if at least one of both parents has a university degree.For material deprivation we use the Family Affluence Scale (FAS III) from the Health Behaviour in School-Aged Children (HBSC) study of the World Health Organization (WHO), which is a sum score of family wealth based on six items asking about the number of cars, computers, and bathrooms a family owns, the child having their own room, the family possessing a dishwasher, and the number of vacations abroad during the past year (Inchley et al. 2020).The total score ranges from zero (low wealth) to ten (high wealth).We reversed the total score so that higher scores indicate higher deprivation.We further control for sex, age, and country-level differences (using country dummies).Finally, since data collection was impacted by the COVID-19 pandemic, we add a dummy if the interview took place after the first lockdown in the respective country.
Analytically, we perform a series of multilevel regression models to adjust for the nested data structure (students nested in classes).In addition, we can investigate the association of class level averages with individual level attitudes.Indeed, we may assume that little support between classmates or highly negative teacher perception in general is linked to populist attitudes.To account for these possibilities, we calculated class averages of teacher evaluation and peer evaluation and added them to the models.All models are estimated as random slope models, that is, we allow the intercepts of the dependent variable (populist attitudes) and the effects of the independent variables of interest on populist attitudes to vary between classes.To ease interpretation, we centered age around the mean and rescaled all other continuous variables to a scale from zero to ten in all models.
Finally, we emphasize our findings by using panel data from the British Household Panel Study (BHPS) and the UK Household Longitudinal Study (UKHLS) (University of Essex, Institute for Social and Economic Research 2023).We explain the research design, question wording, and operationalization in greater detail in online appendix C and we report a summary of the findings in the section on robustness checks.The data allow us to track young people's attitudes during adolescence up into adulthood.To do so, we link children's answers from the youth questionnaire (age 13 to 15) with

PPL1
Politicians should always listen closely to the problems of the people.PPL2 The will of the people should be the highest principle in this country's politics.

ANT1
The government is pretty much run by a few big interests looking out for themselves.

ANT2
Quite a few of the people running the government are crooked.MAN1 You can tell if a person is good or bad if you know their politics.MAN2 The people I disagree with politically are just misinformed.
answers from when children move into the adult questionnaire (at age 16).This setup also allows us to include parental information by linking children to parents.However, since our study is the first to measure populist attitudes among adolescents, we use external political efficacy as dependent variable-which has been shown to be associated with populist attitudes (Bene and Boda 2023;Geurkink et al. 2020;Spruyt, Keppers, and Van Droogenbroeck 2016) and populist party support (Krause and Wagner 2021).Similarly, we have to rely on more general satisfaction with family and friends instead of specific measures of relationships.

Results
To give an impression about the prevalence of populist attitudes among adolescents, figure 1 displays the overall distribution of populist attitudes and the mean values by age in the pooled sample.In general, the degree of populist attitudes among adolescents is rather low; three-quarters of that population hold values below two on a scale from one to five (Panel A).Furthermore, there are no significant age differences, as the mean level of populist attitudes fluctuates around the value of two between age 12 and 18 (Panel B).The are further no differences in distributions between countries (refer to figure A.1 in the online appendix).Finally, the mean values and distributions of populist attitudes are also quite similar compared to adult samples in Austria and Germany as we show in online appendix B. 7  Whereas adolescents have a mean level of populist attitudes of 1.80 (SD=0.86,refer to table A.1), adults hold a mean level of 1.70 (SD=0.70,refer to table B.1). 8  Table 2 presents the main findings of our study. 9As we can see in the first model, possibilities of codetermination, that is, having a say at home, show no correlation with populist attitudes (b=-0.025,95%-CI [-0.072; 0.021]).This also holds when we use a more detailed index about the parent-child relationship that is only available in Austria and Germany (refer to table A.3, b=0.051, 95%-CI [-0.005; 0.107]).Hence, the parent-child relationship seems to be less well connected to populist attitudes.In Model 2, we then find that a bad peer relationship, e.g.being treated badly by classmates or feeling alone at school, is positively associated with populist attitudes (b=0.075,95%-CI [0.019; 0.132]).Similarly but even stronger, we find in Model 3 that a negative teacher-child relationship is also positively associated with populist attitudes (b=0.121,95%-CI [0.076; 0.166]).Perceived unfair treatment in class and in school is thus substantially connected to populist attitudes among adolescents.
In Model 4, we then include the relationship perceptions with all three socialization agents at the same time.As we can see, the association of peer evaluation with populist attitudes becomes weaker (b=0.057,95%-CI [0.001; 0.112]), whereas the association of teacher evaluation remains largely the same (b=0.114,95%-CI [0.070; 0.159]).The coefficient for parental relationship remains insignificant.
Furthermore, we tested in Model 5 whether the average level of unfair teacher treatment is associated with adolescents' populist attitudes, too.This does not seem to be the case, though (b=0.037,95%-CI [-0.092; 0.166]).Since we could argue that teacher perceptions work differently when they are shared with others, we also interacted the individual teacher evaluation with the class-level average evaluation of teachers.As we show in table A.4 in the online appendix, such a connection is not significant.There is also no significant interaction term for bad peer-relationships on a class level (table A.4).However, we observe that the association of peer relationship perception with populist attitudes becomes non-significant in those models.Thus, we tentatively conclude that teacher relationships seem to hold an important connection with populist attitudes.Finally, we find that girls and adolescents from schools with a higher aspiration level have on average a somewhat lower level of populist attitudes, whereas adolescents with migration background and higher family affluence are associated with stronger populist attitudes.There are no differences across age, parental education, countries, and whether the survey was carried out before or after the lockdown.

Robustness Checks
We performed a series of robustness checks in the online appendices to corroborate our findings.First, we reran all main models using different operationalizations for populist attitudes (tables A.5 and A.6). Regardless of operationalization, our results are robust and we basically find the same patterns as presented in the text.Furthermore, we tested whether the relationships behave differently in the three countries (tables A.7 through A.9).Although we find no major differences between Austria, Germany, and Switzerland, we have to acknowledge that the association of negative teacher relationship with populist attitudes is only significant at the 10% level in Switzerland (p=0.097).However, given the sample size (n=639), we still believe that this is in line with our main findings.The coefficients of peer relationship perceptions are, however, not significant in any of the country-specific models.
Since political socialization could potentially work differently for adolescents with migration background (for example, more negative teacher relationship or less responsive parental education), we also reran our main models excluding all adolescents with migration background.The findings show, however, that our results from the main models hold even if we exclude migrants from the models (table A.11).Furthermore, including parental political interest (as reported by the child) does not change our results in a meaningful way (table A.12).We also tested whether the three subdimensions of populist attitudes are connected differently to (negative) relationship experiences with socialization agents (table A.13).The results show that negative relationships with teachers (and to a somewhat lower degree with peers) are strongly correlated with antielitism and Manicheanism.In turn, negative relationships with parents are not related to any of the three dimensions.
Finally, we checked upon the heterogeneity of associations of socialization agent perceptions with populist attitudes by age groups through additional interaction models (table A.10).While we find no significant interaction term in general, the results are suggestive that the association of teacher evaluation with populist attitudes might increase as adolescents get older (figure 2).Since the majority of our sample falls between age 14 to 16 however, imprecision increases at the lower and upper end of the age range.Future research could, therefore, investigate this relationship further.We also find no differences in associations by age for peer and parental evaluation.While both models show a negative trend across age groups, that is, a lower association strength, it is not significant (refer to figures A.4 and A.5).
Taken together, we believe that our results are robust across various specifications. 10While we treated the potential heterogeneity across age in exploratory fashion, it is in line with previous research on the socialization of children and early adolescents, which finds that the role of teachers and schools increases over time, whereas the influence of parents diminishes (see also Bacovsky and Fitzgerald 2023;Bruch and Soss 2018).
One potential downside of our study is its crosssectional nature and the risk of unobserved confounders (see also Cinelli, Forney, and Pearl 2022).To underscore our findings, we ran additional analyses using panel data from the BHPS and UKHLS (refer to online appendix C). 11 Table 3 shows the association of relationships with socialization agents in adolescence with political efficacy in adulthood-a potential mediator in the causal chain to populist attitudes.The results mainly confirm the findings from our own study, showing a significant negative association of a bad teacher relationship in adolescence with the level of external political efficacy in early adulthood (b=-0.077,95%-CI [-0.151; -0.003]) and vice versa for good relationship (b=0.101,95%-CI [0.029; 0.174]).In turn, we find no significant associations for satisfaction with family or friends.Models including additional control variables do not change the results substantially (table C.3 in online appendix C).Finally, we investigated the connection of negative relationships with socialization agents on political involvement more generally.Using fixed-effects models, we tested whether changes in relationship perceptions are associated with changes in political interest.The results from fixed-effects linear probability models in table C.4 in online appendix C show that a one point higher perception that the "the teachers are always getting at me" (on a scale from zero to ten) is associated with holding 1.3 percentage points lower level of political interest (and 2.3 percentage points higher level for positive teacher perception).In turn, there are no significant results for satisfaction with family or friends.

Discussion and Conclusion
In this study, we highlighted the prevalence of populist attitudes among adolescents using unique representative surveys in three countries and we investigated the role of socialization agents for the development of populist attitudes.Our results demonstrate a consistent association between negative teacher relationships and higher levels of populist attitudes.Thus, children who feel they are treated unfairly by their teachers are more likely to hold higher levels of populist attitudes already at young age.Potentially, this association is also likely to increase with age but needs further research.In turn, a negative peerrelationship is less strongly associated with populist attitudes and there is consistently no correlation between parental relationship perception and populist attitudes.These findings hold for a variety of specifications and the theorized mechanisms and the long-term connection of teacher relationship perceptions are supported through additional findings from panel data in the UK.One reason-maybe the main reason-for this might be the nature of the relationship with teachers as opposed to the one children have with their parents and peers.Teachers are a (state) authority outside the household in which children grow up.At the same time, children only interact with teachers in a given setting in school, that is, teachers "play" one specific role, which makes it easier for adolescents to draw conclusions from good or bad relationships.For parents and peers this is somewhat different, as they might fulfill different roles at different times.For instance, parents serve as authority figures, but they are also an emotional safe haven in times of trouble from the outside world.Thus, it becomes difficult for children to relate the negative experience with this kind of authority figure to potential anti-elitist sentiments from populists.Preliminary evidence can be drawn from correlations between relationship perceptions and the three different subdimensions of populist attitudes, but further research is needed.
Our study is the first to measure populist attitudes among a representative sample of adolescents in multiple countries.The findings present important implications for the development of populist attitudes and their consequences for democracy.Most importantly, populist attitudes are not only formed by the time adolescents start to interact with the political system, that is, around voting age.Rather, our results are more in line with the "impressionable years" hypothesis (Dinas 2013;Stoker and Jennings 2008), suggesting that populist attitudes might develop already before adolescents engage with the political process, that is, during the teenage years.In a developmental perspective, the struggle between individual self-development and the need for group attachment can leave a void for adolescents that can attract them to populist rhetoric that portrays the world in structured black-and-white terms.This further highlights the importance of fostering political trust from early onwards, for example, through citizenship education in school but also extracurricular activities and positive learning environments that stimulate a participatory culture in class (see Ott, Meusburger, and Quenzel 2023).
However, our study has some limitations.First, the field time of our study in Austria and Germany coincided with the COVID-19 pandemic.While we controlled for postlockdown interviews in our models, there may be other processes at hand that cannot be captured by such a dummy variable.Second, although the Austrian and Swiss samples are representative for their respective regions, the German sample is not.Third, our analyses are mostly based on subjective perceptions of relationships with socialization agents.While we believe that the internalization of experiences are more relevant for populist attitudes, further research could investigate the role of objective indicators.Fourth, our main study is based on crosssectional data, which means that we cannot make causal claims.Although the addition of the UK panel data allows us to further study the long-term connection between relationship experiences in adolescence with political efficacy in adulthood, future research could investigate the (enduring) causal impact on populist attitudes more directly when new data becomes available.
While we believe that we make a valuable contribution to the field of populism and political socialization research, more work is also needed to address the specific timing of the development of populist attitudes.This would, however, require a multidisciplinary effort to integrate populist attitudes scales into existing panel data sets in the fields of sociology or educational sciences.Doing so would then allow us to investigate, for instance, whether schools can work against other conditions like deprivation and negative emotions that contribute towards the development of populist attitudes.Experimental work could further increase our understanding about mitigating populist attitudes already at the onset of political socialization.In addition, our results revealed further gaps to which in-depth research should be devoted.In the wake of the debate surrounding the increasing malaise of boys, the findings that girls seem to be less populist and also that higher educated children seem to hold lower levels of populist attitudes, the role of gender presents itself as a interesting venue for further research.
Finally, more research is also needed about children's and adolescents' understanding of the concept of "populism."While it has been shown that even children in the first year of primary school can hold structured political orientations (Deth, Abendschön, and Vollmar 2011), we still know too little about what adolescents conceive of concepts like "the elite" or "the people"something that applies to adults, too.In sum, our study provided first insights into a new and hopefully ongoing area of populism research.relationship is by far the highest of all predictors, it is considered the least sensitive coefficient in the model.11 For a similar procedure, see Jungkunz and Marx (2024).
Figure 1 Distribution of populist attitudes Figure 2Marginal effect of teacher evaluation on populist attitudes by age

Table 1
Question wording for populist attitudes Unstandardized estimates from linear multilevel regression models with standard errors in parentheses.Models include randomslopes for evaluations of relationships with parents, peers, and teachers (if possible).Age has been centered."Education high" refers to being in a school that grants university entrance.CH and DE are country dummies for Switzerland and Germany.Higher values on relationship variables indicate a bad relationship.All continuous variables range from zero to ten. Notes:

Table 3
Regression models of external political efficacy in early adulthood Unstandardized estimates from linear regression models with standard errors in parentheses.Data from BHPS and UKHLS.Shown are associations of mean values of predictors measured in adolescence (age 13 to 15) with the mean value of external political efficacy in early adulthood (age 18 to 21).All models control additionally for labor force status (in early adulthood), self-esteem (in adolescence), the highest level of education ever achieved, parental education, sex, and migration background.All continuous variables range from zero to ten.Full models are reported in table C.2 in online appendix C. Note: