Young and Prone to Populist Ideas? –Facets and Correlates of Populist Attitudes in Middle Adolescence
Abstract
Although susceptibility to populist sentiments among adolescents is viewed with concern, research focusing on the nature of populism in youth is still rare. This two-wave study (approx. 10 month) among 9th grade German adolescents (T1: N = 1,205; Mage = 14.4, SDage = 0.6; 52.1% female) examined populist attitudes as a multi-dimensional construct covering the facets of popular sovereignty, anti-elitism, and homogeneity of people. Adopting an inventory originally developed for adult samples, analyses supported the concept’s multi-dimensionality among adolescents. Cross-lagged panel models (CLPM) showed that these dimensions were linked to sociodemographic correlates (i.e., gender, region, school track, and migration background). Additional CLPM analyses examined reciprocal associations with conceptually related political attitudes (i.e., political trust, satisfaction with democracy, intolerance). The results suggested that anti-elitism, in particular, undermines political trust and satisfaction with democracy and give rise to intolerant attitudes. The methodological implications of measuring populism as multifaceted construct among youth are discussed.
Across many established liberal democracies there are reports about growing levels of distrust in political institutions and dissatisfaction with the functioning of democratic processes. At the same time, many countries experience an increasing societal polarization and a public discourse that is characterized by populist rhetoric. Susceptibilities to populist convictions are therefore viewed with concern, especially when observed among young people (Foa et al., 2020). Although adolescence is considered a critical period of political development and scholars increasingly draw attention to populism in general, little is presently known about the development of populist views among youth (Noack & Eckstein, 2023). The present study thus sets out to gain a better understanding of populist attitudes and its correlates during this period in life.
Populism as a Multi-Dimensional Construct
From a theoretical stance, populism is mainly described as a multi-dimensional construct that is characterized by three core elements: (1) The elites, who are generally perceived as corrupt, disloyal, or evil, (2) the ‘common’ people, who are seen as pure, good, and virtuous, as well as the (3) general will (of the people) that should govern (Mudde, 2004; Mudde & Rovira Kaltwasser, 2017). While there is a strong polarization between ‘the people’ and ‘the elite’ (i.e., manichaeism), each group is considered to be homogeneous in itself. With ‘the people’ this homogeneity further entails an anti-pluralistic stance that excludes everyone not considered to belong to this circle. Both, the strong distrust in elites (‘anti-elitism’) and the belief in the homogeneity and virtuousness of ‘the people’ have been introduced as characterizing dimensions of populism (Mudde & Rovira Kaltwasser, 2017; Schulz et al., 2018). Another dimension attributed to populism is the call for an unrestricted enforcement of the people’s will (popular sovereignty). According to this conceptualization, populist sentiments differ from other constructs, which have been more frequently considered in previous research (e.g., political trust, satisfaction with democracy or civic knowledge), by questioning key features of democratic processes, such as representative forms of government and multi-party political systems. Moreover, a populist rhetoric appeals to the importance of the common-sense that excels rational explanations or scientific evidence and that promises simple solutions to complex problems. Although, right-wing populism is most prominent in public discourse and empirical research, populism is not limited to specific political topics, currents, or ideologies, and can consequently be found on both the left and the right side of the political spectrum (Mudde, 2004). Likewise, populist attitudes may co-occur with other right-leaning attitudes such as right-wing authoritarianism, prejudice, or intolerance toward minorities, while still forming a construct in its own right (e.g. Akkerman et al., 2017; Castanho Silva et al., 2020).
Populism in Adolescence
The adolescent years are suggested to be “impressionable years” in the course of civic development (Neundorf & Smets, 2017; Sears & Levy, 2003). This is primarily explained by the numerous changes that characterize this period in life. In the course of an intensified searching for a sense of identity, questions about oneself but also the world surrounding oneself become more present (i.e., civic identity; Yates & Youniss, 1999). Rapid cognitive changes further characterize the adolescent years and allow young people to grasp societal and political issues and questions in more abstract, holistic, and complex ways (for an overview of socio-cognitive changes in adolescence, see Metzger & Smetana, 2010). Moreover, public and political issues might move closer into adolescents’ everyday life due to experiences in broader social surroundings (e.g., school, neighborhood). In line with this, political attitudes (e.g., Rekker et al., 2015) and behaviors (e.g., Eckstein et al., 2012) were found to increasingly stabilize throughout adolescence.
Developmental theorizing typically explains pathways and processes underlying specific political attitudes and behaviors while drawing on individual, social, and societal (risk) factors (e.g., model of radicalization; Beelmann, 2020). Although basic developmental assumptions can be certainly applied to the formation of populist attitudes, specific models are still lacking, as is corresponding research. Consequently, the available literature on populism in youth is primarily descriptive (rather than theory-driven) to date. At least, previous findings point to a considerable susceptibility toward populist and anti-democratic sentiments among youth. According to a German youth survey, a quarter of young people aged 12 to 25 years are leaning toward populist statements and a little less than 10 percent support a right-wing populist rhetoric (Shell Deutschland Holding, 2019). On a similar note, almost a quarter of US youth between the age of 16 to 24 was found to reject democracy as a way of governing (Foa & Mounk, 2016). A comprehensive integration of large-scale data sets further showed that young people’s satisfaction with democracy has been declining worldwide (Foa et al., 2020). However, given the lack of research on the development of populist attitudes during adolescence, it is even uncertain whether the multi-dimensional nature of populism that has been found among adults can be replicated at this period in life.
Correlates of Populism in Adolescence
The nature and development of populist attitudes in youth cannot be completely understood without taking social and personal characteristics into account helping to explain why some youth are more prone to develop populist attitudes than others. Drawing on an Australian sample, Heiss and Matthes (2017), for example, showed that young people aged 15 to 20 years who followed right-wing populist candidates or organizations on Facebook were less educated and more likely to be male. Likewise, populist attitudes were reported to be linked to lower levels of educational attainment among German youth (Shell Deutschland Holding, 2019). In the same study, more pronounced populist attitudes were found in East Germany than in West Germany (Shell Deutschland Holding, 2019). This pattern of regional differences was also reflected in the support of populist parties among young voters in recent German elections (Der Bundeswahlleiter, 2022). For other sociodemographic characteristics, such as age, gender, or migration background, and their links to populist attitudes the empirical evidence is less consistent or lacking. Considering gender, for example, only minor differences in populist attitudes were reported among German youth (indicating slightly higher approval ratings among male than female participants; Shell Deutschland Holding, 2019).
In order to gain a more comprehensive understanding, populist attitudes should also be considered in relation to conceptually linked constructs. For instance, Heiss and Matthes (2017) also showed that, in addition to demographic correlates, adolescent followers of right-wing populist candidates and organizations reported lower levels of political trust. Among adult samples, relations of populist views with political trust and related constructs also have been found (e.g., Castanho Silva et al., 2020). However, the reported effects might not only go in one direction and populist sentiments themselves may not only result from political distrust but also, reversely, undermine trust in political institutions or foster political or democratic discontent. Given the paucity of longitudinal studies on populism as a multi-dimensional construct in adolescence, a better understanding of associations with conceptually related political attitudes across time is needed.
The Present Study
The present study aims at arriving at a better understanding of populist attitudes and its correlates in youth. More precisely, we examined whether facets of populism, which have been previously established in adult samples, can be identified among youth in middle adolescence. We adopted an inventory introduced by Schulz and colleagues (2018), which allows to differentiate three dimensions of populism, namely anti-elitism (AE), popular sovereignty (PS), and homogeneity of people (HP). In a first step, the construct’s dimensionality as well as the stability of its dimensions across time were examined in a sample of German 9th grade students who were surveyed twice at the beginning and at the end of a school year (Research Question 1). We expected that we could replicate the three-dimensional structure of populist attitudes both at T1 and T2. The stability of each dimension was assessed by drawing on two indicators: First, mean-level stability, which reflects the average amount of increase or decrease within the sample and, second, rank-order stability (correlational stability), which provides information on the stability of interindividual differences in a person’s relative position across time. Given the dearth of previous research, we took an exploratory approach to the extent of both forms of stability at this age.
In a second step, sociodemographic correlates were considered to gain a deepened understanding of characteristics accompanying populist attitudes. Based on research that focused on related civic outcomes we considered school track, gender, age, region, and immigrant background (Research Question 2). We expected populist attitudes to be more pronounced among adolescents attending lower-track schools (i.e., vocational schools) and among Thuringian adolescents compared to their age-mates. As previous findings concerning age, gender, and migration background are less consistent or lacking, we chose an exploratory approach.
Finally, we focused on associations of populist attitudes with related but conceptually distinct political attitudes. Drawing on previous findings, we included measures of political discontent (i.e., political trust, satisfaction with democracy) and exclusionary attitudes against societal minorities (i.e., intolerance). We expected that the three dimensions of populist attitudes would be related to less political trust, satisfaction with democracy, and higher levels of intolerance at T1. We further expected reciprocal longitudinal associations between the dimensions of populist attitudes and the examined political outcomes (Research Question 3).
Method
Participants and Procedure
Data were collected during the school year 2021/2022 in 31 schools (90 classes) from two federal states of Germany - Thuringia in the Eastern and North Rhine-Westphalia in the Western part of Germany. Most of the participating Thuringian schools were located in rather rural areas, while the North Rhine-Westphalian sample included schools from mostly urban regions. Students filled in a questionnaire covering various political topics at the beginning (NT1 = 1,206) and at the end (NT2 = 1,096) of 9th grade, with NT1 - 2 = 1,045 students participating in both waves (response rate of 86.7%). Participation was voluntary with permission of parents and school authorities and without individual compensation, but all classes received 100 Euros for their class fund after the data collection was finished (irrespective of the number of participants per class to avoid peer pressure). Data were collected at 16 single track academic schools (German: Gymnasium; 58.3% of students), eight vocational schools (German: Regelschule; 16.3% of students) and seven comprehensive schools (German: Gemeinschaftsschule; 25.4% of students). Students’ mean age at T1 was M = 14.4 years (SD = 0.6 years, range: 13-17 years), 52.1% identified as female (47.1% male; 0.8% diverse), and 59.5% students came from Thuringia. Moreover, 25.1% of the T1 sample reported having a migration background (i.e., at least one parent not born in Germany). In accordance with official population figures (Federal Institute for Population Research, 2023) the percentage of youth with and without migration background differed markedly between the two federal states with 42.9% students of immigrant descent in North-Rhine Westphalia and 11.9% in Thuringia. Comparing students who participated at both measurement points to students who participated only at T1 with regard to all study variables showed that the former were on average slightly younger (F (1, 1203) = 13.91, p < .001, η2 = .01), reported higher levels of political trust (F (1, 1194) = 13.09, p < .001, η2 = .01) and satisfaction with democracy (F (1, 1185) = 6.75, p = .010, η2 = .01), and indicated lower levels of intolerance (F (1, 1192) = 6.66, p = .010, η2 = .01). However, differences in mean levels were small and attrition was not related to populist attitudes or any sociodemographic characteristics (except for age).
Measures
Populist Attitudes
To assess populist attitudes among students, we used an inventory that was originally developed for adult samples (Schulz et al., 2018). Nine items from the original scale (12 items) were used to assess the facets of Anti-Elitism (AE; e.g., “The differences between citizens and the ruling elite are much greater than differences between citizens”), Popular Sovereignty (PS; e.g., “The citizens, not the politicians, should make our most important political decisions.”), and Homogeneity of People (HP; e.g., “Ordinary people are of good an honest character”). The assessment was complemented by two further items as an initial set of exploratory factor analyses proved that both were a suitable addition. One item complemented the dimension of anti-elitism (“A few important groups of people determine the fate of millions of people”; Imhoff & Bruder, 2014), while the second item fit into the dimension of sovereignty of the people (“Politicians should listen more to the people”; see Appendix A for all item wordings). Items were rated on a 5-pt rating scale (1 = totally agree; 5 = totally disagree). Consistency measures of subscales are reported as part of the result section (RQ 1).
Attitudinal Correlates
Political Trust was assessed in terms of students’ trust in political institutions and processes. Students were asked to indicate how much they trust “1) . . . the federal government”, “2) . . . political parties”, and “3) . . . elections and election results” (Schulz et al., 2016; 1 = not at all; 5 = completely; ωT1 = .75, ωT2 = .78). Intolerance was measured with four items reflecting students’ negative attitudes toward refugees and immigrants (Gniewosz & Noack, 2008; e.g., “Refugees and newly migrated people come here to exploit our welfare state“; 1 = totally disagree; 5 = totally agree; ωT1 = .87, ωT2 = .88). Satisfaction with democracy was measured with a single item (“All in all, how satisfied or unsatisfied are you with democracy as it exists in Germany?”, 1 = not at all satisfied; 5 = very satisfied; Shell Deutschland Holding, 2019).
Sociodemographic Correlates
Sociodemographic correlates comprised Age, Gender (1 = female versus 0 = male and diverse), Region (1 = North Rhine-Westphalia versus 0 = Thuringia), School Type (two dummy indicators; 1 = vocational schools and comprehensive schools, respectively versus 0 = single-track academic schools) and Migration Background (1 = at least one parent born abroad versus 0 = both parents born in Germany).
Analytical Approach
All analyses were run with Mplus 8.7 (Muthén & Muthén 2010-2021) using a full information maximum likelihood (FIML) estimation for continuous variables. In a first step, we employed a confirmatory factor analysis (CFA) to examine the factor structure of all populism items and to test whether the three underlying dimensions of anti-elitism (AE), popular sovereignty (PS), and homogeneity of the people (HP) can be identified in our adolescent sample. In doing so, three latent factors were specified simultaneously at T1 and T2. The three-factor-model was further compared to a single-factor model. We then tested for measurement invariance across time. Model fit was evaluated based on the χ2-statistic and standard values of common goodness-of-fit indices (Goodness of Fit Index (GFI)≥.95, Root Mean Square Error of Approximation (RMSEA)≤.06, Standardized Root Mean Square Residual (SRMR)≤.08; Hu & Bentler, 1999). Next, a cross-lagged panel model (CLPM) was specified to capture rank-order stabilities within each dimension of populism as well as bidirectional relations between dimensions across time. Mean level stability was examined using Wald Chi-Square Test, testing significance of T1 and T2 level differences.
In a second set of CLPM-analyses, sociodemographic correlates were examined by regressing the latent factors of PS, AE, and HP at T1 and T2 on age, gender, region, school track, and immigrant background. In doing so, the effect of each covariate was controlled for the effects of the other covariates in the model. In a third set of analyses, we focused on longitudinal associations between the three dimensions of populism and political trust, intolerance, and satisfaction with democracy. Again, CLPM-analyses were run and cross-lagged effects were examined separately for political trust, satisfaction with democracy, and intolerance to avoid multicollinearity. The sociodemographic variables tested in step 2 were added as covariates at T1 and T2.
Data were collected at the classroom level, which leads to a clustered structure of the data. We computed intraclass correlations (ICCs) for the manifest variables that were less than.10 for populist attitudes and less than.17 for attitudinal correlates. Although in case of such ICCs the regression estimates may not be biased, standard errors still can be underestimated (McNeish et al., 2016). We therefore used Mplus’ Type = Complex-option with class as clustering variable and the implemented sandwich estimator to adjust standard errors in all CLMP-analyses. Due to the number of parameters in relation to the number of clusters and number of independent observations, respectively, we also decided to run CLMP-analyses of steps two and three separately for PS, AE, and HP to ensure trustworthy standard errors.
Results
Dimensional Structure of Populist Attitudes
The CFA with three latent factors for AE, PS, and HP at T1 and T2 yielded an acceptable model fit (χ2(183) = 496.46, p < .001, RMSEA = .04, CFI = .94, SRMR = .04). All three dimensions were substantially related, ranging from rAE,HP = .32 to rPS,AE = .60 at T1 and rAE,HP = .40 to rPS,AE = .62 at T2. Comparing this three-factor model to a single-factor model (χ2(197) = 1204.50, p < .001, RMSEA = .06, CFI = .81, SRMR = .07) showed that the former fit the data significantly better than the latter (Δ χ2(14) = 708.06, p < .001), supporting the assumption that the three dimensions of populist attitudes are related, yet distinct.
Focusing on the consistency of measurements, however, showed that particularly for HP some of the standardized loadings were lower than.50 (see Appendix A). Moreover, the Omega of scales remained below the recommended threshold of .70 (Hayes & Coutts, 2020), ranging from ωT1 = .69 for PS to ωT1 = .58 for AE and ωT1 = .53 for HP at T1. At T2, standardized factor loadings and Omega measures were higher and quite acceptable for PS (ωPS-T2 = .76) and AE (ωPS-T2 = .68) given the small number of items, but omega for HP remained rather low (ωPS-T2 = .63).
Analyses of measurement invariance across time supported the existence of full metric invariance (i.e., equal factor loadings; Δ χ2(8) = 7.06, p = .530) and partial scalar invariance (i.e., equal intercepts; Δ χ2(8) = 38.22, p < .001). Since the overall model fit was still acceptable (χ2(199) = 541.73, p < .001, RMSEA = .04, CFI = .94, SRMR = .04), both forms of invariance were kept in all further analyses.
Stability analyses (Wald Test) indicated that, except for PS, there were no average mean-level changes across time (AE: W(1) = -1.42, p = .156; HP: W(1) = 0.15, p = .884). PS showed a small, but significant decrease between T1 and T2 (ΔM = 0.06, W(1) = 2.57, p = .010). Further, CLMP-analyses pointed to a substantial rank-order stability of all three dimensions across time ranging from b = .56 for PS to b = .69 for HP. Although levels at T1 and T2 were correlated, there were no significant cross-lagged effects between the three dimensions (the final three-factor model is depicted in Figure 1).
Figure 1
Final Three-Factor-Model of Populist Attitudes
Note. (χ2 (199) = 541.73, p < .001, RMSEA = .04, CFI = .94, SRMR = .04; Standardized Estimates; Error Terms Across Time were Allowed to Correlate; for Reasons of Clarity non-significant Paths are not shown)
![Final Three-Factor-Model of Populist AttitudesNote. (χ2 (199) = 541.73, p < .001, RMSEA = .04, CFI = .94, SRMR = .04; Standardized Estimates; Error Terms Across Time were Allowed to Correlate; for Reasons of Clarity non-significant Paths are not shown)](https://content.iospress.com:443/media/dev/2023/17-1-3/dev-17-1-3-dev230344/dev-17-dev230344-g001.jpg)
Sociodemographic Correlates of Populist Attitudes
To examine sociodemographic correlates, the latent factors of AE, HP, and PS at T1 and T2 were regressed on age, gender, region, school track, and immigrant background. CLMP-analyses were run separately for each dimension. As part of the longitudinal design, effects on AE, HP, and PS at T2 were controlled for the outcome’s initial level at T1.
The findings are summarized in Table 1 and revealed a similar pattern for PS and AE at T1. Both dimensions were significantly related to region, with lower levels observed among students from North-Rhine Westphalia than among adolescents from Thuringia. Moreover, young females scored higher than their male age-mates at T1. For HP, school type and immigrant background showed significant relationships, suggesting that HP at T1 was lower among students from single-track academic schools and students with no migration background as compared to their counterparts. We found no effects across time, indicating that none of the considered sociodemographic factors explained relative changes in the considered dimensions of populism between T1 and T2.
Table 1
Results of CLMP-Analyses for Sociodemographic Correlates
Popular Sovereignty | Anti-Elitism | Anti-Pluralism | ||||||||||
T1 | T2 | T1 | T2 | T1 | T2 | |||||||
B (SE) | p | B (SE) | p | B (SE) | p | B (SE) | p | B (SE) | p | B (SE) | p | |
Outcome T1 | - | - | .68 (.06) | <.001 | - | - | .73 (.06) | <.001 | - | - | .69 (.09) | <.001 |
Being Femalea | .09 (.04) | .036 | -.01 (.05) | . 853 | .08 (.04) | .048 | .00 (.04) | .980 | .02 (.06) | .721 | -.06 (.05) | .288 |
Age | -.06 (.05) | .184 | .03 (.04) | .472 | .02 (.03) | .530 | -.01 (.03) | .830 | .05 (.05) | .243 | .00 (.04) | .911 |
School type | ||||||||||||
Regular schoolsb | .01 (.08) | .929 | .11 (.06) | .086 | -.03 (.06) | .598 | .09 (.04) | .090 | .49 (.08) | <.001 | -.13 (.09) | .133 |
Comprehensive schoolsb | .00 (.07) | .999 | .01 (.06) | .801 | -.04 (.06) | .450 | -.00 (.03) | .952 | .21 (.07) | .003 | .09 (.07) | .232 |
North Rhine-Westphaliac | -.17 (.06) | .003 | -.10 (.06) | .096 | -.22 (.06) | <.001 | -.01 (.05) | .849 | .02 (.07) | .791 | .03 (.06) | .605 |
Having Migration Backgroundd | .08 (.06) | .225 | -.08 (.06) | .225 | .09 (.06) | .141 | .05 (.06) | .401 | .29 (.07) | <.001 | .02 (.08) | .785 |
Note. CLMP-analyses were run separately for each dimension. As part of the longitudinal design, effects on T2 were controlled for initial levels at T1. a 0 = male & diverse, b 0 = single-track academic schools, c 0 = Thuringia, d 0 = both parents born in Germany.
Associations Between Populist Attitudes and Political Correlates
The final set of CLPM-analyses separately examined the relations of PS, AE, and HP with intolerance, political trust, and satisfaction with democracy at T1 and T2. In addition, the effects of sociodemographic variables on all outcomes at T1 and T2 were considered. Results are depicted in Figure 2. Analyses revealed significant correlations between all three dimensions of populism and the examined outcomes at T1. All correlations were in the expected direction, showing that higher levels of AE, PS, and HP were negatively related with political trust and satisfaction with democracy, and positively related to intolerance.
Figure 2
CLMP-Analyses (Unstd./Std. Estimates) for Dimensions and Political Trust
Note. (a), Satisfaction with Democracy (b), and Intolerance (c). Analyses were run Separately for each Outcome and Dimension of Populist Attitudes (indicted in Parentheses). All Variables at T1 and T2 were Controlled for Sociodemographic Covariates (Age, Female, Region, School Type, and Migration Background).
![CLMP-Analyses (Unstd./Std. Estimates) for Dimensions and Political TrustNote. (a), Satisfaction with Democracy (b), and Intolerance (c). Analyses were run Separately for each Outcome and Dimension of Populist Attitudes (indicted in Parentheses). All Variables at T1 and T2 were Controlled for Sociodemographic Covariates (Age, Female, Region, School Type, and Migration Background).](https://content.iospress.com:443/media/dev/2023/17-1-3/dev-17-1-3-dev230344/dev-17-dev230344-g002.jpg)
Further, the results revealed significant cross-lagged effects on all attitudinal outcomes at T2. More precisely, the results showed that higher levels of AE, PS, and HP at T1 were related to lower levels of political trust at T2 above and beyond the level of T1 trust. Higher levels of AE and HP at T1 further predicted lower levels of political satisfaction at T2. Intolerance at T2, in contrast, was only predicted by levels of AE at T1. Reversed cross-lagged paths were found for political trust and AE only, suggesting that higher levels of trust at T1 were associated with lower levels of AE at T2.1
Discussion
Given the paucity of longitudinal research on populism in adolescence, the present study aimed at getting a better understanding of the nature and development of populist attitudes in middle adolescence. In particular, we examined the multi-dimensional structure of populist attitudes based on a sample of German adolescents from two federal states, who were surveyed at the beginning and at the end of the 9th grade. As in adult samples, the dimensions of popular sovereignty, anti-elitism, and homogeneity of the people as part of populist attitudes could be identified. Yet, these dimensions seemed to be less consistent among adolescents than among adults. We found differences in the level of popular sovereignty and anti-elitism depending on region and (small effects of) gender, and differences in the level of homogeneity of the people depending on school type and migration background. The pattern of reciprocal associations with conceptually close constructs varied depending on which dimension and outcome was examined. Mutually reinforcing relations across time were found only for political trust and anti-elitism. Otherwise, lagged effects from populist sentiments (especially anti-elitism) to political outcomes were more pronounced than vice versa.
Most of researchers agree that populism can be understand as a multi-dimensional construct (e.g., Mudde, 2004). Since to the best of our knowledge there was no established youth-specific measure of populism available in the field, we adopted an instrument that had originally been developed for adult samples (Schulz et al., 2018). As expected, we could identify three distinct dimensions representing the facets of popular sovereignty, anti-elitism, and homogeneity of the people as also found in adult samples. The overall fit of a three-factor model across both measurement time points was good and each item loaded on the expected dimension. Moreover, the three-factor model clearly exceeded the fit of a single-factor model. However, indices of internal consistency and standardized factor loadings were lower than those reported for adult samples (e.g., Castanho Silva et al., 2020; Schulz et al. 2018). This applied particularly to the first time point and to the assessment of youth’s ratings on the subdimension “homogeneity of the people”. One potential explanation of this lower consistency might be provided by the quite abstract wording of the original items. Although, we have adapted selected terms to make items more suitable for our sample, we did not introduce any youth-specific reformulations. Thus, some phrases might still have been difficult to understand for adolescents, resulting in more noise within our measurement.
Yet, the finding that the consistency of scales increased across time and that the intercorrelations between the three dimensions were higher at T2 might also point an increasing crystallization of populist attitudes. In adolescence, young people achieve the necessary maturity and (cognitive) abilities to deal with more abstract concepts and ideas (Metzger & Smetana, 2010). As a result, political views and ideas might become more consistent and coherent with growing age. For instance, perceptions on ethnocentrism and economic egalitarianism were shown to stabilize significantly throughout adolescence and young adulthood (Rekker et al., 2015). The fact that adolescents’ views become more consistent across time does, however, not necessarily mean that their level of agreement increases as well. As our results showed, on average, mean levels did not change across the school year, except for a small but significant decrease of popular sovereignty. At the same time, there was a substantial inter-individual stability of all three dimensions across time.
Apart from the structure and changes of populist attitudes, we also examined possible associations with sociodemographic characteristics. The results revealed that Thuringian adolescents scored higher on popular sovereignty and anti-elitism than did students from North-Rhine Westphalia. This pattern is consistent with findings from another recent German youth study (Shell Deutschland Holding, 2019) as well as regional differences in voting behavior (Der Bundeswahlleiter, 2022). Furthermore, school track was related to the dimension of homogeneity of people, with higher levels of agreement among students from comprehensive and vocational tracks than among students from single-track academic schools. Quite unexpectedly, approval ratings for popular sovereignty and anti-elitism were also higher among female adolescents and for homogeneity of people among adolescents with immigrant background than among their respective age-mates. Yet, similar to other studies gender differences were rather small. A possible explanation for the differences depending on migration background may be provided by the wording of one particular item highlighting the concept of Germanness (i.e., “Although the Germans are very different from each other, when it comes down to it they all think the same”). Higher approval may mean different things among youth with and without migration background. For German adolescents, the wording may strike at the core element of ingroup homogeneity with the good and virtuous ‘common people’ versus the ‘disloyal elite’ (Mudde, 2004). For adolescence with immigration background, in contrast, the wording might rather reflect a kind of outgroup homogeneity effect (e.g., Judd et al., 1995).
Finally, we examined whether populist views and discontent with politics and democracy mutually reinforce each other across time. In line with our expectations, all three dimensions of populist attitudes were negatively related to political trust and satisfaction with democracy, and positively linked to intolerance at the first time point. This further speaks in favor of the external validity of the scale as the correlational pattern showed that the dimensions were linked, yet distinct from conceptually close constructs (Castanho Silva et al., 2020). We also found evidence that populist sentiments may foster political distrust and dissatisfaction with democracy as well as intolerance across time. In particular, anti-elitism predicted a relative decrease of political trust and satisfaction with democracy as well as a relative increase of intolerance during the school year. For popular sovereignty and homogeneity of people effects were less consistent and less pronounced. With regard to the latter, this might be probably due to its lower reliability. Yet, the pattern of relationships could also point to the particularly polarizing effect of anti-elitism. While popular sovereignty still has some kind of democratic core, the sense that elites (disloyal and looking out for the own advantage) have lost touch to the people undermines trust in political institutions, its incumbents, and democracy itself (Noack & Eckstein, 2023). The loss of political trust, in turn, seems to foster anti-elite sentiments as shown by reversed effect of political trust at the first time point on anti-elitism at the second time point.
This research is one of the few studies that focused on populist attitudes among youth and therefore provides a rare view on populist attitudes and its correlates during this period in life. However, there are some limitation that have to be considered when interpreting the results. As the scale was part of a multi-themed survey, first, the number of items was restricted and the facets were measured less consistent among our student sample than among adults. Some of the effects might therefore not have been detected due to the noise in these measurements (i.e., measurement error). Accordingly, a broader assessment with more youth-specific item wordings would have been helpful. Furthermore, another follow-up survey one year or half a year later would make it possible to rule out that the increasing consistency of scales is not primarily due to repeated measurement but may reflect a substantive process. Second, although we could rely on longitudinal data and used CLMP with covariates to control for initial levels as well as possible third variables, it is not possible to claim causality of the identified relationships (Hamaker et al., 2015). Third, the focus of the present study was on individual- rather than class-level effects. To still avoid an underestimation of standard errors as a result of the classroom-based data collection, we used sandwich estimator to account for the nested data structure. Ultimately, however, only two-level modeling allows for a decomposition and a more sophisticated analysis of class-level versus individual-level effects (McNeish et al., 2017).
Finally, the relations found with sociodemographic as well as attitudinal correlates are not particularly impressive. This could be due to the limited consistency of measures, but may also indicate more complex relationship patterns. In particular, several authors claim that looking at different dimensions is helpful to better understand the structure of populist sentiments, but that populism as such is characterized by the simultaneous presence of all three domains (Castanho Silva et al., 2018). Future studies might therefore focus more strongly on the interplay of the facets of popular sovereignty, anti-elitism, and homogeneity of the people. Moreover, future studies might also consider further possible predictors of populist views, such as personality traits (e.g., Fatke, 2019), and account for contextual influences, particularly effects of family and school (e.g., Noack & Eckstein, 2023). We thus consider our work as an initial step and hope to stimulate a further inquiry into the development of populist beliefs in young people. After all, a stable and healthy democracy cannot persist without the support of its younger generation.
References
1 | Akkerman, A , Zaslove, A , Spruyt, B. ((2017) ). ‘We the people’ or ‘we the peoples’? A comparison of support for the populist radical right and populist radical left in the Netherlands, Swiss Political Science Review 23: (4), 377–403. https://doi.org/10.1111/spsr.12275 |
2 | Beelmann, A. ((2020) ). Asocial-developmental model of radicalization:Asystematic integration of existing theories and empirical research, International Journal of Conflict and Violence 14: , 1–14. https://doi.org/10.1016/j.jsurg.2016.07.008 |
3 | Castanho Silva, B , Andreadis, I , Eva, A , Blanusa, N , Morlet Corti, Y , Delfino, G , Rico Ruth, G. S. P , Spruyt, B , Steenbergen, M , Littvay, L. (2018). Public opinion surveys: A new scale. In A. Hawkins, K. A., Carlin, R. E., Littvay, L.&Rovira Kaltwasser, C. (Eds.), The ideational approach to populism. Concept, theory, and analysis (pp. 150-171). Routledge. https://repositorio.uca.edu.ar/handle/123456789/8613 |
4 | Castanho Silva B , Jungkunz, S , Helbling, M , Littvay, L. ((2020) ). An empirical comparison of seven populist attitudes scales, Political Research Quarterly 73: (2), 409–424. https://doi.org/10.1177/1065912919833176 |
5 | Der Bundeswahlleiter (2022).Wahl zum 20. Deutschen Bundestag am 26. September, 2021 [Election to the 20th German Bun destag on September 26, 2021]. Retrieved August 29, 2022, from https://www.bundeswahlleiter.de/dam/jcr/8ad0ca1f-a037-48f8-b9f4-b599dd380f02/btw21_heft4.pdf |
6 | Eckstein, K , Noack, P , Gniewosz, B. ((2012) ). Attitudes toward political engagement and willingness to participate in politics: Trajectories throughout adolescence, Journal of Adolescence 35: (3), 485–495. https://doi.org/10.1016/j.adolescence.2011.07.002 |
7 | Fatke, M. ((2019) ). The personality of populists: How the Big Five traits relate to populist attitudes, Personality and Individual Differences 139: , 138–151. https://doi.org/10.1016/j.paid.2018.11.018 |
8 | Federal Institute for Population Research. (2023). Bevölkerung mit Migrationshintergrund in den Bundesländern (2020) [Population with a migration background in the federal states (2020)]. Retrieved May 15, 2023, from https://www.bib.bund.de/Permalink.html?cmspermaid=1217678 |
9 | Foa, R. S , Klassen, A , Wenger, D , Rand, A , Slade, M. (2020). Youth and satisfaction with democracy: Reversing the democratic disconnect? Centre for the Future of Democracy. Retrieved May 15, 2023, from https://www.cam.ac.uk/system/files/youth_and_satisfaction_with_democracy.pdf |
10 | Foa, R. S , Mounk, Y. ((2016) ). The danger of deconsolidation: The democratic disconnect, Journal of Democracy 27: (3), 5–17. https://doi.org/10.1353/jod.2016.0049 |
11 | Gniewosz, B , Noack, P. ((2008) ). Classroom climate indicators and attitudes towards foreigners, Journal of Adolescence 31: (5), 609–624. https://doi.org/10.1016/j.adolescence.2007.10.006 |
12 | Hamaker, E. L , Kuiper, R. M , Grasman, R. P. P. P. ((2015) ). Acritique of the cross-lagged panel model, Psychological Methods 20: (1), 102–116. https://doi.org/10.1037/a0038889 |
13 | Hayes, A. F , Coutts, J.J. ((2020) ). Use Omega rather than Cronbach’s Alpha for estimating reliability, But. . . . Communication Methods and Measures 14: , 1–24. https://doi.org/10.1080/19312458.2020.1718629 |
14 | Heiss, R. , Matthes, J. ((2017) ). Who ‘likes’ populists? Characteristics of adolescents following right-wing populist actors on Facebook. Information, Communication&Society, 20: (9), https://doi.org/10.1080/1369118X.2017.1328524 |
15 | Hu, L.-T , Bentler, P. M. ((1999) ). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives, Structural Equation Modeling 6: (1), 1–55. https://doi.org/10.1080/10705519909540118 |
16 | Imhoff, R , Bruder, M. ((2014) ). Speaking (un-)truth to power: Conspiracy mentality as a generalised political attitude, European Journal of Personality 28: (1), 25–43. https://doi.org/10.1002/per.1930 |
17 | Judd, C. M , Park, B , Ryan, C. S , Brauer, M , Kraus, S. ((1995) ). Stereotypes and ethnocentrism: Diverging interethnic perceptions of African American and White American youth, Journal of Personality and Social Psychology 69: (3), 460–481. https://doi.org/https://doi.org/10.1037/0022-3514.69.3.460 |
18 | McNeish, D , Stapleton, L. M , Silverman, R. D. ((2017) ). On the unnecessary ubiquity of hierarchical linear modeling, Psychological Methods 22: (1), 114–140. https://doi.org/10.1037/met0000078 |
19 | Metzger, A , Smetana, J. G. (2010). Social cognitive approaches to civic engagement. In L. R. Sherrod, J. Torney-Purta&C. A. Flanagan (Eds.), Handbook of research on civic engagement in youth (pp. 221-248). Wiley. |
20 | Mudde, C. ((2004) ). The populist Zeitgeist, Government and Opposition 39: (4), 541–563. https://doi.org/10.1111/j.1477-7053.2004.00135.x |
21 | Mudde, C , Rovira Kaltwasser, C. (2017). Populism:Avery short introduction. Oxford University Press. |
22 | Muthén, B. , Muthén, L. (2010-2021). Mplus. Los Angeles, CA: Muthén&Muthén. |
23 | Neundorf, A , Smets, K. (2017). Political socialization and the making of citizens. In Oxford Handbooks Online. Oxford University Press. https://10.1093/oxfordhb/9780199935307.013.98 |
24 | Noack, P. , Eckstein, K. ((2023) ). Populism in youth: Do experiences in school matter, Child Development Perspectives 17: (2), 90–96. https://doi.org/10.1111/cdep.12481 |
25 | Rekker, R , Keijsers, L , Branje, S , Meeus, W. ((2015) ). Political attitudes in adolescence and emerging adulthood: Developmental changes in mean level, polarization, rank-order stability, and correlates, Journal of Adolescence 41: , 136–147. https://doi.org/10.1016/j.adolescence.2015.03.011 |
26 | Schulz, W , Ainley, J , Fraillon, J , Losito, B , Agrusti, G. (2016). IEA International Civic and Citizenship Education Study 2016 Assessment Framework. International Association for the Evaluation of Educational Achievement (IEA). Springer https://doi.org/10.1007/978-3-319-39357-5 |
27 | Schulz, A , Müller, P , Schemer, C , Wirz, D. S , Wettstein, M , Wirth, W. ((2018) ). Measuring populist attitudes on three dimensions, International Journal of Public Opinion Research 30: (2), 316–326. https://doi.org/10.1093/ijpor/edw037 |
28 | Sears, D. O , Levy, S. (2003). Childhood and adult political development. In D. O. Sears, L. Huddy,&R. Jervis (Eds.), Oxford handbook of political psychology (pp. 60-109). Oxford University Press. |
29 | Shell Deutschland Holding (2019). Jugend 2019. Eine Generation meldet sich zu Wort [Youth 2019. A generation has a say]. Beltz. |
30 | Yates, M , Youniss, J. (1999). Roots of civic identity: International perspectives on community service and activism in youth. Cambridge University Press. |
Bio Sketches
Astrid Körner is a postdoctoral researcher at the Department of Educational Psychology at Friedrich Schiller University Jena, Germany. Her research interests include civic participation in adolescence and adulthood, contextual influences (school, media, peers, family) on young people’s EU/European orientations and educational processes in adulthood.
Katharina Eckstein is a postdoctoral researcher at the Department of Educational Psychology at Friedrich Schiller University Jena, Germany. Her research interests include the development of civic attitudes and behaviors during adolescence/young adulthood, the role of family and school contextual influences in youth’s civic development, as well as structural equation modeling and multilevel analysis techniques.
Peter Noack is Professor for Educational Psychology at Friedrich Schiller University Jena, Germany. His research interests include youth research, development and socialization in family and friendship relations, attitudes, learning behavior, and school achievement of students, as well as civic socialization.
Appendix A
Item Wordings and CFA Factor Loadings
Table 1
Item Wordings and Standardized Factor Loadings of CFA with Three Latent Factors for Popular Sovereignty, Anti-elitism, and Homogeneity of the People at T1 and T2
No. | Item | Standardized Loadings | |
Popular Sovereignty (PS) | T1 | T2 | |
PS1 | The citizens, not the politicians, should make our most important policy decisions. | .62 | .69 |
PS2 | The citizens should be asked whenever important decisions are taken. | .64 | .69 |
PS3 | The people should have the final say on the most important political issues by voting on them directly in referendums. | .53 | .63 |
PS4 | Politicians should listen more to the people. | .62 | .66 |
Anti-Elistism (AE) | |||
AE1 | The differences between citizens and the ruling elite are much greater than the differences between citizens. | .55 | .57 |
AE2 | A few important groups of people determine the fate of millions of people. | .59 | .68 |
AE3 | Members of the Bundestag very quickly lose touch with citizens. | .52 | .60 |
AE4 | People like me have no influence on what the government does. | .40 | .51 |
Homogeneity of the people (HP) | |||
HP1 | Ordinary people all pull together. | .64 | .67 |
HP2 | Although the Germans are very different from each other, when it comes down to it they all think the same. | .48 | .53 |
HP3 | Ordinary people are of good and honest character. | .46 | .59 |
Note. χ2(183) = 496.46, p < .001, RMSEA = .04, CFI = .94, SMRS = .04.
Notes
1 To test whether cross-lagged effects of populist attitudes remain stable when the mutual effects of all facets were controlled, we ran a further CLMP analysis including PS, AE, and HP simultaneously. Results showed that effects of PS and HP at T1 on political trust at T2 decreased and were no longer significant (PS: B = - .02, SE = 0.05, p = .733; HP: B = - .07, SE = 0.05, p = .165). As pointed out earlier, standard errors and significance levels of these models should be interpreted with caution due to the ratio of parameters to independent observations following the clustered data structure. Still, the results indicate that effects of PS and HP on trust might be substantially driven by the variations that both dimensions share with AE.