The rise of engaged citizenship: The evolution of citizenship norms among adolescents in 21 countries between 1999 and 2009

Abstract

Various authors claim that citizenship norms are changing rapidly in advanced democracies, leading to a stronger emphasis on self-expressive engagement and a decline of notions of civic duty. In this article, we compare results from two comparative surveys of adolescents: the 1999 Civic Education Study and the 2009 International Civic and Citizenship Education Study (ICCS). By using latent class analysis, we identify duty-based and engaged citizenship norms, both in 1999 and in 2009. As expected, the group supporting duty-based citizenship norms is clearly smaller in 2009 than in 1999, while the opposite is true for the group supporting engaged citizenship norms. In contrast to expectations, the empirical evidence also distinguishes additional normative concepts and shows that the distribution among countries is not according to the dynamics on value change as suggested in the literature, including a decline in engaged norms in Scandinavia and Western Europe.

Keywords

Citizenship norms duty-based citizenship engaged citizenship latent class analysis value change

Introduction

Within industrially advanced societies, political and social norms and attitudes are changing rapidly, especially as young age cohorts pay more attention to self-expressive values and to the importance of individual freedom (Welzel, 2013). This process of value change implies that these groups will relate differently to political systems than previous generations did. Traditional ideological cleavages are now less salient, while party membership and electoral politics are losing ground among younger age groups (Blais et al., 2004; Blais and Rubenson, 2013; Klingemann, 2015; Whiteley, 2011). Simultaneously, however, other forms of political and civic engagement are gaining ground, and these forms of engagement are less oriented toward electoral politics, but rather to various community, voluntary, and social movement oriented forms of participation (Flanagan, 2013; Inglehart and Catterberg, 2002; Zukin et al., 2006). These changes do not just signal different organizational preferences, but may be the result of changing citizenship concepts (Norris, 2011). More specifically, various authors argue that duty-based citizenship norms are being replaced by a more engaged or critical form of citizenship, whereby especially younger age groups place more emphasis on actively helping out in their community in a non-institutionalized manner (Dalton, 2008; Flanagan, 2013; Norris, 2011; Sloam, 2013). Previous studies indeed show that this engaged form of citizenship can be identified, and that it is clearly present in at least some Western societies (Copeland, 2014; Dalton, 2007, 2008; Martin, 2012; Raney and Berdahl, 2009; Shulman and Levine, 2012). It is assumed that especially in the Scandinavian countries, these norms will be strongly present (Dalton and Welzel, 2015).

While there seems to be a consensus in the sociological literature about the occurrence of this process of value change, opinions differ on the conceptualization of this process. Various authors use different terms to describe this newly emerging citizenship norm, including ‘self-actualizing’ (Bennett, 2012), ‘critical’ (Norris, 1999, 2011), ‘monitorial’ (Schudson, 1998), and ‘engaged’ (Dalton, 2007, 2008). Although these terms highlight different aspects of citizenship norms, these studies represent a growing consensus that contemporary young people are often highly engaged within their communities while they de-emphasize traditional duty-based concepts of citizenship. These changes lead to the expectation that young age groups will prefer social movement oriented forms of participation, rather than participation that is directed toward electoral politics (Torney-Purta, 2001). Indeed, empirical studies demonstrate such a shift in participation repertoires (Stolle and Hooghe, 2011), but while these studies are compatible with the broader thesis of value change, they cannot be considered as a final test of these theories. If we want to test the claim that civic value patterns are changing in Western societies, it is important to investigate the trends with regard to these attitudes and values in a more direct and comparative manner.

The theoretical literature on new citizenship norms emphasizes that value change takes place across industrial democracies. Most of the available empirical studies, however, are based on cross-sectional observations that do not allow us to test the claim that engaged citizenship has become more prevalent (e.g. Dalton, 2008) or they investigate over-time changes in a single country or region (e.g. Howe, 2010; Oser and Hooghe, 2013). In this article, we aim to investigate in a comparative manner changes over time in citizenship norms. We do so using two large surveys among adolescents that were conducted in 1999 and 2009. Both the 1999 Civic Education Study (CivEd) and the 2009 International Civic and Citizenship Survey included a full battery of items on citizenship norms. This means we can compare the results for both studies for the 21 countries that participated in both data gathering efforts, and it also allows us to investigate differences between countries with regard to the prevalence of these norms. As far as we know, no other broad comparative dataset includes such a comprehensive battery on citizenship norms. Most previous studies on this topic opted for an item-based analysis of either single-item studies (e.g. Bolzendahl and Coffé, 2013) or attitudinal scales (e.g. Dalton, 2008). The theory does not only predict support for specific scales, but rather, it also assumes that actors will have very specific combinations of responses regarding what constitutes good citizenship. The theory thus predicts changes at the level of the actor, and in order to adequately test the theory, this should also be our unit of analysis. We opt therefore for an actor-centered analytical technique (latent class analysis (LCA)) that is explained more fully in the ‘Data and methods’ section of this article.

In the remainder of this article, we first briefly review the literature on changing citizenship norms among young age cohorts, before we present data and methods. We next ascertain the characteristics of citizenship norms in both surveys in order to investigate changes over time and to investigate differences between countries. In the concluding section, we reflect on what our findings tell us about the claim that citizenship norms are rapidly changing, especially among young age cohorts.

Changing citizenship norms

Just over a decade ago, the prevailing mood in the literature on the civic engagement of young age groups was rather pessimistic, as authors claimed that younger age groups are less inclined to participate in political and community life (Pharr and Putnam, 2000). Other authors, however, strongly countered this line of thinking by claiming that new generations do not abandon politics, but rather that they reshape the way politics is being performed (Bennett and Segerberg, 2013; Inglehart and Welzel, 2005; Norris, 2002; Zukin et al., 2006). These changes are well documented with regard to political behavior and participation, demonstrating that traditional, electoral forms of participation like voting or party membership are in decline among young age groups (Blais and Rubenson, 2013; Bolzendahl and Coffé, 2013; Fieldhouse et al., 2007), while these adolescents and young adults are much more active in non-institutionalized or elite-challenging forms of participation (Marien et al., 2010; Welzel and Deutsch, 2012).

In this literature on the development of participation patterns, a number of authors highlight that changing citizenship norms are the main driving force for this process of social change (Bennett, 2012; Dalton, 2007, 2008; Norris, 1999, 2002; Schudson, 1998). Inglehart and Welzel (2005) claim that younger age groups are more critical toward political institutions and organizations as they prefer an individualized form of engagement that allows them to express their own individual preferences and inner convictions. According to their ‘human development theory’, the fact that education levels and cognitive sophistication are rising among this younger age group allows them to develop a far more individualized pattern of political preferences in comparison to previous generations. Bennett and Segerberg (2013) build on the work of Castells (2000) to argue that young age groups avoid traditional political institutions, as they prefer loose and more informal networks. As a result of the emergence of new forms of electronically mediated communication networks, an idiosyncratic network structure of interaction partners can be established. The ‘critical citizen’ concept (Norris, 1999, 2011) departs from the assumption that younger age cohorts no longer adopt a deferential attitude toward authorities and institutions, but instead develop a more critical attitude toward politics, while they are staunch supporters of basic democratic values. Among these theories about the nature of political value change, Dalton (2013) offers the most explicit claims about the consequences of these changes for political engagement. He argues that evolving citizenship norms are guiding political behavior. In his argument, the structural change takes place at the level of norms and attitudes, and not at the level of behavioral patterns. Following Dalton (2008) we define citizenship norms as ‘a shared set of expectations about the citizen’s role in politics’ (p. 78). The assumption is that these norms guide the behavior of citizens (John et al., 2011; Straughn and Andriot, 2011). As Dalton (2008) claims,

… the norms of citizenship are vital to understanding the political behavior of the American public […] My central premise is that the social and political modernization of the United States – and other advanced industrial democracies – over the past several decades has systematically altered the distribution of citizenship norms in significant ways. (p. 77)

The basic idea in this line of the literature is that citizenship norms held by young age groups still stress the importance of an active involvement in social life (Eliasoph, 2013; Flanagan, 2013). The difference with previous generations, however, is that norms now depart from different motivations. In the past, a sense of civic and moral duty often served as the guiding principle (Blais, 2000). Even without an intrinsic motivation, civic engagement was simply seen as the duty of a citizen because of external pressures, the message by political elites, or simple tradition. Partly as a result of cognitive mobilization and higher education levels, this unquestioning acceptance of the duty argument, however, has become obsolete as young citizens are more critical (Bobek et al., 2009; Norris, 2011). Duty-based norms are being replaced by what Dalton (2008) labels ‘engaged citizenship’, that is, an action-based form of citizenship where activism in the local community is valued for its own sake. This form of intrinsic engagement can also express itself in network and internationally oriented forms of engagement (Bennett and Segerberg, 2013). These emerging citizenship norms expect less from state institutions, as the focus is clearly on what citizens themselves can do without being dependent on state intervention. Theoretically, this is congruent with the notion of lifestyle activism as it was developed in the work of Giddens (1991).

The literature on changing citizenship norms relates to broader processes of social change. Dalton (2008: 81) argues that the rise of an engaged citizenship norm is a consequence of a post-modernization process that leads to the spread of self-expressive values (Inglehart and Welzel, 2005; Welzel, 2013). These values, in turn, lead to the observation that ‘engaged citizens’ will be more sensitive to issues like human rights, but they will also be less deferential toward political authorities and political elites.

Despite this well-developed literature on changes in citizenship norms, comparative data over time are scarce. In their reviews of the citizenship norm literature, Bennett (2012) and Van Deth (2007) note that comparative empirical research on citizenship norms is limited, especially outside North America and Europe. Although the literature about the behavioral effect of changing citizenship norms relates to changes over time, it is striking to note that the empirical support that Dalton marshals for this thesis is based largely on cross-sectional observations from the United States (Dalton, 2007, 2008). Beyond the United States, engaged or critical citizenship norms have been identified in a number of cross-sectional studies, including single-country studies of Canada (Raney and Berdahl, 2009) and Australia (Martin, 2012), Dalton’s (2007: 138–160) analysis of 19 established democracies and Hooghe et al.’s (2015) study of adolescent norms in 38 countries. Yet, these studies do not inform us about whether engaged citizenship is actually on the rise. Given the general literature on value change and cognitive mobilization, this might be expected, especially in countries that traditionally score high on post-materialist indicators (Inglehart, 1997), but there is no conclusive evidence yet in the literature. While most studies focus on Northern and Western Europe, there is far less information available about the spread of the attitudes in other geographical regions.

The aim of the current article is to address the question of over-time change by analyzing strictly comparative data from 1999 and 2009. If value change in recent years is indeed such a powerful and structural process as Bennett, Dalton, Inglehart, Norris, Welzel, and other authors argue, a decade should be sufficient to observe significant changes in adolescents’ citizenship norms.

We depart from three main research questions. First, we need to ascertain whether these two forms of citizenship norms indeed can be empirically distinguished, both in the 1999 and the 2009 data. If this is the case, the second research question is whether we observe a decline of more traditional, institution- and duty-based citizenship norms and a corresponding rise in engaged or critical citizenship norms. The third question is whether the engaged citizenship norms are most prevalent in the most affluent and stable democracies, as post-modernization theory argues. The 1999 and 2009 data can be used to address these questions, since both surveys were conducted among the same age group, used the same methodology, and reached comparable levels of representativeness.

Both surveys were conducted among adolescents, which is ideal for our purpose since the literature states that new citizenship concepts will emerge first among younger age groups (Eckstein et al., 2012; Hooghe and Dassonneville, 2013). Although we cannot yet adequately investigate participation among adolescents since they are still excluded from electoral participation, research shows that citizenship norms can be studied in a meaningful and valid manner among adolescents (Van Deth et al., 2011). Yet, we also need to acknowledge that the democratic ideals of adolescents are still in a developing phase (Flanagan, 2013), so it remains to be investigated how these norms will change once they enter early adulthood. We address this limitation more fully in the ‘Discussion’ section of this article.

Data and methods

If we want to use survey data to investigate changes over time, it is important that research design and data gathering methods are strictly comparable. For this reason, we selected two surveys, both conducted in a similar manner by the International Association for the Evaluation of Educational Achievement (IEA). This organization is responsible for comparative evaluation studies of education systems, and in 1999 and 2009, civic education was the focus of such an effort, just as in the past IEA conducted tests of students’ language and mathematical skills. While these data have not been routinely used in sociological or political science research, the school-based approach of the fieldwork ensures a high level of representativeness for this age group.

Both surveys were conducted in the same manner, as educational authorities in the participating countries were responsible for fieldwork and quality control. The survey was conducted in school classes, and in most countries, educational authorities implemented the survey themselves. This design resulted in high response rates and representative samples for the school population of 14-year-olds. In 1999, the IEA CivEd was conducted among 90,000 respondents in 28 countries (Torney-Purta et al., 2001). Ten years later, a similar design was followed for the International Civic and Citizenship Education Study (ICCS) (Schulz et al., 2011), this time including information from 140,000 respondents in 38 countries.

Both surveys can be directly compared as they use the same methodology and target the same age group. In total, 21 countries participated in both surveys, the majority of which are in Europe, but with also a few countries in Asia and Latin America. In order to compare the 1999 and the 2009 data in a reliable manner, we limit the analysis to the following 21 countries for which we have comparable data: Bulgaria, Switzerland, Chile, Columbia, Cyprus, Czech Republic, Denmark, England, Estonia, Finland, Greece, Hong Kong, Italy, Lithuania, Latvia, Norway, Poland, Russia, Slovakia, Slovenia, and Sweden.¹

The main and unique appeal of these two datasets is that they include a comprehensive battery on citizenship norms. A rather straightforward way to analyze this survey would be to construct attitudinal scales (e.g. based on a factor or principal component analysis), and to use these as the main dependent variable. Such an approach would be completely valid as a research design. However, in the literature on changing citizenship norms, it is clear that the main concern is about individuals, not about attitudes as such. The expectation is that groups of individuals will strongly emphasize one set of values or preferences, while they disregard another set. Exactly, this combination of preferences renders them ‘engaged’, ‘critical’, or ‘self-expressive’. If we applied traditional attitude-focused regression analysis, it would be impossible to detect this specific combination. For example, with a traditional analysis of attitude scales or regression analysis, we could arrive at the conclusion that girls score higher on civic duty and lower on critical engagement, but these kinds of findings would not mean that there is actually a group of girls that make this very distinctive combination. In order to test the theory in a valid manner, therefore, rather than focus on the prevalence of an attitude, we examine combinational patterns made by specific groups of individuals. This kind of actor-centered approach traditionally used cluster analysis, but given the important problems with regard to the identification of specific clusters of groups, this kind of approach by now has been largely abandoned. In the more recent literature, actor-centered analyses mainly relied on LCA or related techniques, as these allow us to identify latent classes on the basis of reliable goodness of fit statistics (Collins and Lanza, 2010). While LCA is not yet routinely used in the study of political attitudes, it is by now widely applied, for example, in medical sociology, and in this article, we use this technique to investigate changing citizenship concepts (Hagenaars and Halman, 1989; Magidson and Vermunt, 2004; Yuan et al., 2014).

In recent years, the application of LCA as a probabilistic form of cluster analysis became widespread in various fields in the social sciences. The technique allows for the clear identification of subgroups within a population, combining a similar set of characteristics (Keel et al., 2004; Wolf et al., 2012). The main difference between LCA (and related techniques) and factor analysis is that the grouping of observed data for factor analysis is based on the correlation between indicators, whereas for LCA, the grouping is based on an actor-centered approach that identifies respondents who share similar patterns of answers in a survey. In terms of the research interest in this article of identifying groups with specific sets of citizenship norms, factor analysis could help us to identify how variables group together into separate ‘dimensions of citizenship’ (Dalton, 2008: 80–81), whereas LCA can directly assess the theory that distinctive groups of people share specific citizenship norms. Model selection in LCA is guided by goodness of fit criteria and stability statistics, and this is an advantage compared with traditional cluster analysis where the number of clusters often depends on rather arbitrary decision rules (Yang, 2006). We also use multilevel analysis to ascertain whether economic development and democratic stability of a country have an impact on the prevalence of specific citizenship norms.

Results

Our analysis proceeds in three steps. First, we ascertain whether it is possible to identify distinct citizenship concepts, both in the 1999 and the 2009 dataset. Subsequently, we investigate whether the trend over time corresponds to theoretical ideas about value change in contemporary society. Third, we analyze differences between countries.

Results: CivEd 1999

In both the 1999 and 2009 surveys, respondents were asked to explain what a good citizen is or does. Subsequently, the respondent received a list of 12 different activities and was requested to rank these activities on a 4-point scale of importance: two categories that rated the item as not important (‘not important at all’ and ‘not very important’), and two response categories that rated the item as important (‘quite important’ and ‘very important’).² In this article, we present LCA findings based on the more parsimonious two-category coding that makes a distinction between behaviors that respondents consider to be not important (0) and behaviors that respondents consider important (1). Although this two-category coding scheme entails a loss of variance, an important advantage of this coding is that we can be confident of cross-country metric equivalence (Agresti, 2007; Kankaraš et al., 2010; Kankaraš and Vermunt, 2014).

Most of the activities in the scale receive high scores (Table 1). A vast majority of respondents consider obeying the law important for a good citizen, but protecting the environment, defending human rights, or voting in elections are also generally seen as important. Discussing politics and party membership, on the other hand, are seen as important by only a minority of respondents.

Table 1.

Frequency of Good Citizenship Acts in CivEd 1999. In this section, there are some statements that could be used to explain what a good adult citizen is or what a good adult citizen does: Obeys the law; votes in every election; joins a political party; works hard; would participate in a peaceful protest against a law believed to be unjust; knows about the country’s history; follows the political issues in the newspaper, on the radio, or on TV; participates in activities to benefit people in the community; shows respect for government representatives; takes part in activities to promote human rights; engages in political discussions and in activities to protect the environment.

	Obey	Rights	Local	Work	Envir	Vote	History	Respect	News	Protest	Discuss	Party
Mean	.95	.84	.83	.81	.81	.80	.73	.72	.72	.68	.44	.31
n	70,451	67,854	68,617	68,822	69,054	69,879	69,026	67,765	69,078	65,140	66,166	66,651
Standard deviation	.212	.363	.375	.394	.389	.400	.446	.449	.450	.467	.496	.462
Order	1	2	3	4	5	6	7	8	9	10	11	12

Source: CivEd 1999, 21 countries, N = 72,543.

CivEd: Civic Education Study.

Entries are proportion of the respondents considering this act as (very) important for a good citizen.

A first step in the analysis is to determine the number of latent classes that should be distinguished. A crucial advantage of LCA is that the Bayesian Information Criterion (BIC) allows researchers to identify the model (including the number of groups) that matches most closely with the structure of the dataset. The BIC can be used to identify a unique best model, which obviously was not the case in more traditional cluster analysis (Nylund et al., 2007). Although one model selection approach is to opt for the absolute lowest BIC value, in research on complex attitudinal concepts, it is common to base model selection on a declining BIC, along with the percent change in the likelihood ratio chi-square statistic and the classification error (Magidson and Vermunt, 2004: 176–177). Using these model selection criteria suggests that five latent classes provide an optimal fit to these data (Table 2). Models where we try to identify two, three, four, or six latent groups do not offer a better match with the structure of the data.³ A test for measurement equivalence in both observation points confirmed the cross-national metric equivalence of the latent classes, which means that the citizenship concepts can be validly compared across countries.⁴ In the online appendix, we document the LCA fit statistics and conditional probabilities findings for all countries separately in both observation points. These findings show that fit statistics support the model choice considerations in the pooled sample: individual country analyses tend to show a major improvement in fit up through the five-cluster model, after which point an increased number of classes provides limited improvement in model fit. Regarding the conditional probabilities, the findings show an overall stability in the findings based on the two observations. For all countries, versions of the five main citizenship norms are present, even though individual countries sometimes have idiosyncratic emphases. This analysis therefore confirms the cross-national and cross-time measurement equivalence of our key theoretical constructs.

Table 2.

LCA model fit statistics for citizenship norms in 1999.

	BIC (LL)	L²	% change L²	Class. Err.
1-Class	639,326	52,627		0.00
2-Class	603,265	16,425	−0.69	0.11
3-Class	598,406	11,424	−0.78	0.18
4-Class	596,635	9512	−0.82	0.23
5-Class	595,191	7927	−0.85	0.26
6-Class	594,125	6719	−0.87	0.29

Source: CivEd 1999.

LCA: latent class analysis; BIC: Bayesian Information Criterion; LL: log likelihood; L²: likelihood ratio chi-square statistic; AIC: Akaike Information Criterion; AIC3: Akaike Information Criterion with 3 as penalizing factor; CAIC3: consistent Akaike Information Criterion with 3 as penalizing factor.

LCA findings using Latent Gold 4.5 software. Entries are test statistics for latent class models identifying one and more clusters of respondents. Additional commonly used test statistics were estimated (AIC, AIC3, CAIC3) and were all in agreement with the BIC and the percent change in L² regarding the preferred model (in bold).

Three of these latent classes are represented in Figure 1 by gray markers, where the markers represent the probability that respondents belonging to that specific latent group consider each activity to be (very) important for what an adult good citizen should do. The latent class probabilities can be compared with the population mean for each indicator, which is listed in parentheses underneath the x-axis. This comparison shows that each one of the classes has a unique pattern of combining specific preferences.

Figure 1.

Citizenship norms in 1999, highlighting ‘Engaged’ and ‘Duty-based’.

The largest of these groups, including a quarter of the sample, is what we would label an ‘all-around’ citizen. This group is distinct because it considers almost every act as important. The probabilities are only somewhat lower for protesting, discussing politics, and joining a party, but they are still well above average. In contrast, the group we refer to as the ‘mainstream’ citizen (21% of all respondents) is largely on par with the sample mean on the indicators. Although this group has some deviations from the sample mean, such as a relatively high emphasis on work and news, and a relatively low emphasis on respect and party, this group largely reflects the average scores on indicators for good citizenship. Finally, in contrast to the ‘all-around’ citizens, we also find a sizeable group of respondents (16%) that has very low average scores. This group has a relatively high emphasis, however, on working hard and obeying the law, so in line with traditional political culture research, they can be referred to as ‘subjects’ (Almond and Verba, 1963).

The two theoretically most interesting groups are represented by black markers with connecting lines in Figure 1. As the figure indicates, the two latent classes of engaged and duty-based citizens have distinctive normative preferences. For the sake of clarity, we will refer to those groups as respectively engaged and duty-based citizens. ‘Engaged’ citizens (16%) are likely to consider the protection of the environment and human rights as important, and they also place emphasis on being active in the local community. At the same time, however, this group is less likely to consider traditional duty-bound acts to be important, such as working, voting, respecting government officials, and joining a party. This group therefore most closely represents Dalton’s concept of ‘engaged citizenship’. While their emphasis on human rights and the environment is in line with the theories of Inglehart and Welzel (2005), their focus on local level civic engagement resembles most closely the ideas put forward by Dalton (2008).

Likewise, the ‘duty-based’ group (22%) is close to the ideal type of traditional, civic duty-based citizenship. In contrast to the engaged citizens, this group emphasizes traditional acts like voting, showing respect, and watching the news. Even though the duty-based citizens have an active sense of traditional political duty, they do not highly value protecting the environment and human rights. As such, this group emphasizes the importance of institutionalized and electoral politics, while it does not pay much attention to typical post-modern sensitivities. Although the work of Norris (2011) does not refer explicitly to this rather traditional group, it is clear that their preferences are exactly the opposite from what she labels as ‘critical citizens’.

This analysis of the CivEd 1999 data confirms the empirical validity of two sets of citizenship norms that can be referred to as engaged and duty-based. It has to be noted, however, that the engaged and duty-based groups combined amount to only 38 percent of all respondents. Almost two-thirds of the respondents in 1999 do not hold these citizenship norms. Rather surprisingly, 16 percent of respondents adhere to very traditional citizenship norms, by emphasizing obeying the law and working hard.

Results: ICCS 2009

Subsequently, we proceed in the same manner, with the 2009 ICCS survey data. When looking at the frequencies (Table 3), a first observation is that there is not much difference in comparison to the 1999 results (Table 1). Obeying the law remains the most often mentioned item, while party membership remains lowest on the priority list. For all 12 items, the differences between the scores in 1999 and those in 2009 remain rather limited, and at first sight, a simple look at the population averages might suggest stability over this 10-year period. LCA, however, allows us to determine the underlying trends.

Table 3.

Frequency of Good Citizenship Acts in ICCS 2009. Report: How important are the following behaviors for being a good adult citizen? Joining a political party, learning about the country’s history, following political issues in the newspaper, on the radio, on TV, or on the Internet (Note: Internet was not mentioned in CivEd 1999), showing respect for government representatives, engaging in political discussions, participating in peaceful protest against laws believed to be unjust, participating in activities to benefit people in the local community, taking part in activities promoting human rights, taking part in activities to protect the environment, working hard, always obeying the law.

	Obey	Rights	Local	Work	Envir	Vote	History	Respect	News	Protest	Discuss	Party
Mean	.89	.82	.79	.80	.84	.79	.77	.77	.73	.62	.40	.28
Difference in 2009	−.06	−.02	−.04	−.01	+.03	−.01	+.04	+.05	+.01	−.06	−.04	−.03
n	73,421	73,210	73,214	73,320	73,284	73,711	73,153	73,411	73,463	73,174	73,279	73,430
Standard deviation	.314	.381	.405	.400	.367	.404	.422	.424	.444	.484	.490	.448
Order	1	3	6	4	2	5	7	8	9	10	11	12

Source: ICCS 2009, 21 countries, N = 75,020.

ICCS: International Civic and Citizenship Education Study.

Entries are proportion of the respondents indicating that this behavior is (very) important for a good citizen. Indicators ordered from left to right in the same order as the ascending CivEd means in Table 1; the ‘Order’ row ranks the indicators from most to least prevalent in the 2009 data.

As in our analysis of the ICCS 2009 data, we first determine the optimal number of latent classes (Table 4). Goodness-of-fit measures again indicate a clear preference for a five latent classes solution which allows for a valid comparison over time.

Table 4.

LCA model fit statistics for citizenship norms in 2009.

	BIC (LL)	L²	% change L²	Class. Err.
1-Class	894,867	91,113		0.00
2-Class	840,078	36,178	−0.60	0.10
3-Class	827,149	23,104	−0.75	0.15
4-Class	821,190	17,000	−0.81	0.20
5-Class	817,354	13,019	−0.86	0.22
6-Class	814,928	10,448	−0.89	0.26

Source: ICCS 2009.

LCA: latent class analysis; ICCS: International Civic and Citizenship Education Study; BIC: Bayesian Information Criterion; LL: log likelihood; L²: likelihood ratio chi-square statistics; AIC: Akaike Information Criterion; AIC3: Akaike Information Criterion with 3 as penalizing factor; CAIC3: consistent Akaike Information Criterion with 3 as penalizing factor.

In Figure 2, five latent classes are plotted for the 2009 data, which in many ways parallel the 1999 findings. Again, the ‘all-arounders’ are the largest group (35% of all respondents). As 10 years earlier, members of this group tend to answer that everything is important, and they have the highest probabilities to consider every single item to be important. It can be noted, however, that in 2009, this group places less emphasis on party membership and engaging in political discussions. While in 1999 65 percent of the group members considered party membership to be important, this was down to 51 percent in 2009.

Figure 2.

Citizenship norms in 2009, highlighting ‘Engaged’ and ‘Duty-based’.

While the ‘engaged’ group only amounted to 16 percent of all respondents in 1999, this group increased to 30 percent in 2009. This sharp rise supports the claim that engaged citizenship norms are on the rise in contemporary society. The profile of the engaged group is very similar in both surveys, with high scores for human rights, the environment, and being active in the local community, while this group remains skeptical about voting and joining a political party. The group of engaged citizens is thus very similar in both surveys, but it almost doubles in size compared to the 1999 results. As far as we know, this is the very first time the rise of engaged citizenship norms has been demonstrated in an unequivocal manner for such a broad range of countries.

A third group identified in the analysis is the duty-based citizens, and this group falls from 22 percent in 1999 to 15 percent in 2009. The group is again characterized by high scores for voting and respecting authority, but protecting rights and the environment and being active in the local community become even less important for this group. Again, in 2009, we see a ‘mainstream citizens’ cluster which roughly echoes the emphases of the population average, although this group is smaller in 2009. Finally, the percentage of ‘subjects’ declines from 16 to 6 percent, and they remain attached to obeying the law and working hard.

In sum, both in 1999 and in 2009, distinctive groups of engaged and duty-based citizens could be identified, together with three additional groups (i.e. the all-around, mainstream, and subject groups). At first sight, these findings confirm our claim that it was appropriate to use an actor-based form of analysis. However, the question remains, did the use of LCA lead to theoretically superior results, compared to a more standard attitude-based analysis? To assess this, we conducted an additional analysis that relies purely on a factor analysis. Factor analyses based on the data from 1999 and 2009 show that all items load primarily on one factor. In both 1999 and 2009, the main component has a very large eigenvalue (2.612 in 1999 and 2.856 in 2009), while the other two components have eigenvalues close to 1.0. In both observations, the first factor explains more than twice the variance as the next largest factor. These findings corroborate Bolzendahl and Coffé’s (2013: 69) report that a factor analysis of the 25 Eastern and Western European countries in the International Social Survey Programme in 2004 led to just one factor which lumped together various citizenship items. A closer look at the LCA findings shows why a traditional factor analysis is so inconclusive. The huge percentage of all-around and mainstream group respondents (almost half of the total sample in both measurements) attributes high values to a wide range of items, and this leads to one large and theoretically unsatisfactory factor where all items are lumped together. The factor analysis is completely dominated by this group, so that it does not allow us to observe what goes on in the other smaller groups that are of theoretical interest in this article. The LCA, on the other hand, is much more precise in this crucial aspect, and it allows us to gain new insights in this process of value change among specific groups of the population.

Results: Comparison 1999–2009

The previous results indicated that duty-based and the engaged citizenship norms can be empirically distinguished, both in 1999 and 2009. We have also seen that engaged citizenship norms became more prevalent, while the opposite trend has been observed for duty-based citizenship norms. We emphasize, however, that in both observation periods, more than half of the respondents do not fit into either of these two categories. While much of the theoretical literature is focused only on engaged and duty-based groups, it is important to remember that in a wide variety of industrialized democracies, young people continue to adhere to other norms, including very traditional ones.

As we focus on comparing the engaged and duty-based groups in 1999 and 2009, Figures 3 and 4 show that these norms are largely similar in both observations, but that we can also observe some subtle changes. For the engaged citizens (Figure 3), human rights and the environment remain of paramount importance, but we can observe that discussing politics is rated slightly lower in 2009. Knowing the history of one’s country, on the other hand, receives a higher rating than in 2009. The main defining elements of the engaged citizenship norm are almost identical in both observations, strengthening the argument that this concept can be compared over time.

Figure 3.

Comparing the ‘Engaged’ citizenship norm in 1999 and 2009.

Figure 4.

Comparing the ‘Duty-based’ citizenship norm in 1999 and 2009.

In Figure 4, we compare the characteristics of the duty-based norm. In general, this group rates the various acts lower in 2009 than it did in 1999, so it seems that even among this group, the idea of a ‘duty’ to perform these acts gets weaker. Defining characteristics of this group endure, however, such as respecting authority and working hard (which has an even higher likelihood in 2009). So while the group of duty-based citizens becomes smaller, it develops a more distinct normative profile, with stronger differences between acts that are considered important and those that are not.

In sum, our first research question whether engaged and duty-based citizenship norms can be empirically identified was confirmed. Both in 1999 and in 2009, the LCA indicated that both of these norms are present, with rather small differences regarding the relative emphases of these norms in both observations. The second research question also can be answered in an affirmative manner (Figure 5): the proportion of engaged citizens rises (from 15% to 30%), while the duty-based citizenship norms lose support (from 22% to 15%). These findings support the claim that the decline of duty-based citizenship and the accompanying rise of engaged citizenship are present across industrialized democracies.

Figure 5.

Distribution of citizenship norms, 1999 and 2009.

Results: Comparative analysis

Thus far, we have presented analyses for a wide range of 21 different societies, but these trends should also be investigated in a comparative manner. It is relevant to determine whether similar trends can be found in all of these countries separately. First, we present an overview of our findings that compares geographical regions. In most of the literature on value change in contemporary society, it is assumed that especially the Scandinavian countries and Western Europe will be frontrunners in the trend toward new normative sensitivities, such as support for human rights or the environment. The figures in Table 5 show that this was indeed still the case in 1999: the engaged citizenship concept was most strongly present in the Scandinavian countries (with 37% of all respondents) and in Western Europe (30% of all respondents), while less than 10 percent of respondents adhered to this norm in other regions, including Mediterranean Europe, Central and Eastern Europe, and South America and Asia. The 1999 results thus clearly follow the pattern we would expect based on the literature with regard to post-materialism. For 2009, however, it can be observed that this trend reversed, with a decline of engaged citizenship in the Scandinavian countries and Western Europe and a simultaneous rise of engaged citizenship in all other regions. This distribution clarifies that some of the conventional wisdom about more contemporary value orientations being present mainly in Northern and Western Europe no longer seems to be valid in the most recently available figures. Especially, the countries in Asia and Latin America seem to have experienced a rapid transition toward more engaged citizenship norms among adolescents, although it has to be acknowledged that the number of participating countries in those remains limited.

Table 5.

Trends in citizenship norms across regions of the world.

	Scandinavia	Western Europe	Mediterranean Europe	Central and Eastern Europe	South America and Asia
Prevalence of engaged norms
1999	0.37	0.30	0.08	0.09	0.07
2009	0.28	0.18	0.14	0.34	0.42
Prevalence of duty-based norms
1999	0.03	0.10	0.34	0.19	0.34
2009	0.28	0.28	0.08	0.12	0.08

Source: CivEd 1999; ICCS 2009.

ICCS: International Civic and Citizenship Education Study; LCA: latent class analysis.

Percentage of all respondents adhering to one of the norm groups, for 1999 and 2009, based on LCA findings. Scandinavia: Denmark, Finland, Norway, and Sweden; Western Europe: Switzerland and England; Mediterranean Europe: Cyprus, Greece and Italy; Central Europe: Bulgaria and Czech Republic; Eastern Europe: Lithuania, Latvia, Poland, Russia, Slovak Republic, and Slovenia; South America: Chile and Columbia; Asia: Hong Kong.

The country-specific findings challenge an assumption often made in this line of research that changes in citizenship norms and political attitudes follow a typical pattern where Nordic countries take the lead, while newly emerging democracies are slower to adapt to the new trends (Inglehart, 1997). In Appendix 1, we list the country figures for the distribution of the citizenship norms, both in 1999 and in 2009 (Tables 7 and 8). These findings show that in the CivEd study in 1999, the assumption about the geographic distribution of this trend is still rather valid as we find the highest proportion of engaged citizens in Nordic countries like Denmark, Norway, and Sweden. In 2009, however, these citizenship norms can equally be found in Latin America or the Baltic states, and the Nordic countries are no longer exceptional. The theoretical relevance is that while Northern and Western Europe traditionally have been depicted as frontrunners in the trend toward new citizenship norms, at least among adolescents, their position clearly is no longer exceptional.

In order to test these differences between countries in a more rigorous manner, we conducted a multilevel regression analysis, including both individual and country level independent variables. On the individual level, we keep the information to a minimum, by including for only gender and socio-economic status. In this kind of data, it is usually not possible to obtain reliable information from the adolescent respondents about family income or professional status of the parents. Therefore, a routine solution is to ask adolescents about the number of books at home as a proxy for socio-economic status. On the country level, we include the years of stable democracy and gross domestic product (GDP) per capita (Inglehart and Catterberg, 2002). Given the fact that the number of observations on the country level is limited, both indicators had to be included one by one. In line with the literature, our hypothesis is that both democratic stability and economic status will have a positive effect on engaged citizenship and a negative one on duty-based citizenship.

The results of the multilevel regression analysis (Table 6) confirm expectations for the 1999 data (see Appendix 1, Table 9, for descriptive statistics). For duty-based citizenship, there is a significantly negative effect of both the years of stable democracy and the GDP per capita, while we find exactly the opposite relation for engaged citizenship in 1999. When we subsequently conduct the same analysis for the 2009 data, we obtain exactly the opposite result, with less engaged citizenship in countries with a long history of stable democracy and high economic development level. Our finding that the Scandinavian and established Western European countries have lost their front runner position in this regard is thus confirmed by the results of this comprehensive multilevel regression test. At least among adolescents, the spread of new citizenship norms seems to have become a more global phenomenon that is no longer concentrated in specific geographic regions.

Table 6.

Multilevel regression model, determinants of duty-based and engaged norms.

	Duty I	Duty II	Duty III	Engaged I	Engaged II	Engaged III
1999
Individual level
Female	0.143*** (0.034)	0.150*** (0.016)	0.154*** (0.015)	0.247*** (0.023)	0.218*** (0.016)	0.214*** (0.015)
SES proxy ^a
101–200 books	−1.003*** (0.045)	0.091*** (0.021)	0.076*** (0.021)	0.577*** (0.030)	0.005 (0.021)	−0.003 (0.021)
201 books+	−0.943*** (0.041)	0.109*** (0.020)	0.086*** (0.020)	0.647*** (0.028)	−0.032 (0.020)	−0.047** (0.019)
Country level
Democracy (years stable)		−0.057** (0.028)			0.040*** (0.015)
GDP per capita			−0.013* (0.007)			0.010*** (0.003)
Constant	−3.486*** (0.028)	−4.159*** (0.861)	−4.181*** (0.831)	−4.134*** (0.019)	−3.727*** (0.446)	−3.750*** (0.426)
Observations	71,613	66,724	71,613	71,613	66,724	71,613
Number of groups		20	21		20	21
2009
Individual level
Female	0.184*** (0.023)	0.196*** (0.019)	0.204*** (0.019)	0.582*** (0.025)	0.570*** (0.023)	0.564*** (0.022)
SES proxy ^a
101–200 books	0.298*** (0.029)	0.057** (0.025)	0.052** (0.025)	−0.026 (0.032)	0.209*** (0.029)	0.205*** (0.029)
201 books+	0.254*** (0.029)	−0.049* (0.025)	−0.055** (0.025)	−0.245*** (0.032)	0.098*** (0.030)	0.085*** (0.029)
Country level
Democracy (years stable)		0.024** (0.011)			−0.017* (0.010)
GDP per capita			0.003** (0.001)			−0.002* (0.001)
Constant	−4.013*** (0.018)	−3.944*** (0.330)	−3.896*** (0.321)	−2.386*** (0.020)	−2.537*** (0.315)	−2.597*** (0.307)v
Observations	73,890	71,111	73,890	73,890	71,111	73,890
Number of groups		20	21		20	21

Source: CivEd 1999; ICCS 2009.

SES: socio-economic status; GDP: gross domestic product; ICCS: International Civic and Citizenship Education Study.

Entries are results of multilevel regression analysis with standard errors in parentheses. Reference category is 0–100 books.

Dependent variables are conditional probability of duty-based or engaged membership.

p < 0.05. **p < 0.01. ***p < 0.001.

Discussion

In this study, we conduct a major comparative analysis of trends in citizenship norms in 21 countries, exploiting the availability of new survey data on citizenship norms among adolescents. One of the advantages of working with these data is that they allow for a rigorous research design that maps trends over time in an empirically valid manner.

The results of the study largely confirm the claims put forward in the literature on value change. First, both engaged and duty-based citizenship norms can be identified in an empirically valid manner, and second, we can observe that engaged citizenship is on the rise while duty-based citizenship is declining. Even more strongly than the literature expects, these findings indicate a very general trend that is also present in Latin American and in Central and Eastern Europe. Value change thus indeed proceeds in the way authors like Inglehart, Norris, Schudson, and Dalton suggested. Simultaneously, however, a very large part of the sample does not fit into this theoretical framework, and respondents in the largest of these groups have the tendency to state that everything is important. This ‘all-around’ group is even so large that in a factor-based regression analysis, it overwhelms the entire pattern. An actor-centered analytical technique like LCA is required to detect the occurrence of the smaller groups that are central in the theoretical literature. Our findings show that while these ‘pioneer’ groups may be highly relevant from a theoretical point of view, they do not constitute a majority among this age group.

Our findings deviate from Dalton’s claim, however, in a different manner. What we can observe from a direct comparison between these data in 1999 and 2009 is that there is very little change in the overall importance attached to the various items. Adolescents in 2009 state that voting and obeying the law are of the utmost important for being a good citizen, just like their counterparts did in 1999. The importance of these ‘traditional’ acts of good citizenship did not diminish. What does differ, however, is the way adolescents consider these traditional acts to be important when examined as part of a broader citizenship norm that simultaneously takes other aspects of good citizenship into account. When we analyze these data, we observe that the group that emphasizes the importance of voting and obeying the law becomes a smaller proportion of the population, in favor of groups that have different priorities for good citizenship. So, while our analysis supports the trend toward the increased prevalence of ‘engaged citizenship’, we can also see that more traditional and institutionalized acts of citizenship participation, like voting, are still considered as important and will not be easily replaced in the near future.

The evidence for the geographical distribution of citizenship norms across countries is rather mixed. For the 1999 data, expectations were confirmed, as we did find the highest prevalence of engaged citizenship norms in Scandinavia, and in other countries with a stable democratic tradition and a high economic development level. Ten years later, however, this pattern is reversed. This questions the theoretical assumption that the trend toward post-materialist value orientations can be regarded as a purely linear process. Apparently, other countries caught up with the high levels that were already present in the Scandinavian countries more than a decade ago. Tentatively, this could indicate that at least among adolescents, the Scandinavian countries have lost their exceptional status, and that the spread of these new norms has become a much more global phenomenon. It has to be remembered, however, that we have few countries from these regions in our dataset, so that broader data are necessary to arrive at a more comprehensive assessment of these trends. An alternative plausible explanation for this trend might be that the questions also probe into what adolescents consider to be most salient. For adolescents in Scandinavia, protecting the environment or human rights might not be considered as highly salient, because the governments in those countries have already taken numerous initiatives to incorporate these sensitivities in their policies. For the adolescents in Latin America, on the other hand, this might seem a highly salient issue, exactly because adequate government policy is missing. To what extent this is not just a plausible but also a likely explanation, however, still needs to be investigated.

A number of caveats are in order. First, while the identified citizenship norms are remarkably similar in both surveys, they are obviously not identical. The most characteristic items, however, do remain firmly in place, thereby confirming that a comparison over time is warranted. The analysis suggests, however, not only that duty-based citizenship becomes less prevalent over the observation period but also that the specific element of ‘duty’ becomes diluted as members of this group are less likely in 2009 to consider a whole range of duty-based acts as important. In addition, it is worth emphasizing that both surveys were conducted among adolescents, and we wish to elaborate upon our prior observation that there are different positions in the literature about the theoretical status of political attitudes among adolescents. The remarkable stability of the 1999 and 2009 LCA solutions suggests that these adolescents did not respond in a random manner, but that there is a distinct pattern in their answers. In line with previous research, these findings indicate that citizenship norms can be measured in a valid manner among adolescents.

A question remains, however, regarding the relationship between normative beliefs of adolescents and their adult counterparts. On the one hand, some authors make the point that there tends to be a strong pattern of stability between adolescence and early adulthood (Quintelier and Hooghe, 2012). Following this assumption would allow us to predict that the adolescent respondents in these surveys to a large extent will continue to have the same political preferences as they mature into adulthood. On the other hand, authors in the field of developmental psychology suggest that adolescents are still in developmental phases in the lifecycle, during which actors continue to develop their moral and cognitive concepts and identities (Flanagan, 2013). This line of the literature would suggest that there will be important differences between the citizenship concepts of adults and adolescents. For example, the large proportion of respondents who highly value obeying the law is consistent with traditional findings in developmental psychology that authority beliefs are strong among this age group (Smetana and Turiel, 2003). Since we are unaware of any comparative dataset among adults that includes a similar over-time battery of citizenship concepts, it remained beyond the scope of this article to investigate or explain differences between adolescents and adults. In future research, it would be highly relevant to make such a direct comparison between citizenship concepts among adolescents and adults, as only such a test provides the necessary information to ascertain whether our finding could be generalized toward a larger population.

Our findings generate a number of new research questions. In line with Dalton’s change thesis, we can observe that the group that finds it important to defend human rights and to work in the local community has become larger over time. Yet, it remains to be seen whether and how this preference for local and non-institutionalized participation enables the establishment of a meaningful linkage mechanism between citizens and the political system. Various authors have already documented how political decision-making tends to shift toward a higher geographic scale as a reaction to economic and ecological demands (Stiglitz, 2007). Simultaneously, however, the citizens of tomorrow, as we might label these adolescents, increasingly seem to prefer local and non-institutionalized engagement. One cannot help but wonder how effective this form of engagement can be, and if in fact ‘engaged citizenship’ may be tantamount to citizens losing their clout in the process of political decision-making.

Footnotes

Appendix 1

Table 9.

Descriptive statistics of variables in multilevel regression analyses.

	n	Mean	Standard deviation	Min.	Max.
1999
Female	74,308	0.51	0.50	0	1
SES proxy^a	74,173	1.61	0.80	1	3
Democracy (years)^b	72,118	40.92	30.65	0	90
GDP per capita^c	75,020	313.87	241.40	51.74	1025.26
1999
Female	71,948	0.52	0.50	0	1
SES proxy	71,812	1.67	0.84	1	3
Democracy (years)	67,546	34.13	29.67	0	82
GDP per capita	72,543	143.92	120.40	16.23	393.04

Source: CivEd 1999 and ICCS 2009.

GDP: gross domestic product; ICCS: International Civic and Citizenship Education Study; SES: socioeconomic status.

SES proxy is the self-reported number of books at home that includes the following categories: (1) 0–100 books; (2) 101–200 books; (3) 201+ books.

Democracy (years): based on the Polity IV 2011 dataset for 1998 and 2008, and includes the total number of years that the country scored 8 or higher on the 10-point democracy index (Marshall et al., 2011). Hong Kong does not receive a score in the Polity IV data, and is therefore not included in the multilevel regression that controls for democratic stability.

Based on the International Monetary Fund’s World Economic Outlook Database, annual data for 1998 and 2008 (International Monetary Fund, 2013).

Acknowledgements

The authors thank Michael Delli Carpini and Miloš Kankaraš for comments on the research. We also appreciate the constructive comments from the anonymous reviewers and the editor of this journal.

Funding

The authors acknowledge the funding by the European Research Council, ERC Advanced Grant ‘Linkage between citizens and the state’.

Notes

References

Agresti

(2007) An Introduction to Categorical Data Analysis, 2nd edn. Hoboken, NJ: Wiley.

Almond

Verba

(1963) The Civic Culture. Princeton, NJ: Princeton University Press.

Bennett

(2012) The personalization of politics: Political identity, social media, and changing patterns of participation. Annals of the American Academy of Political and Social Science 644(1): 20–39.

Bennett

Segerberg

(2013) The Logic of Connective Action. Digital Media and the Personalization of Contentious Politics. Cambridge: Cambridge University Press.

Blais

(2000). To Vote or Not to Vote? The Merits and Limits of Rational Choice Theory. Pittsburgh: University of Pittsburgh Press.

Blais

Rubenson

(2013) The source of turnout decline: New values or new contexts? Comparative Political Studies 46(1): 95–117.

Blais

Gidengil

Nevitte

(2004) Where does turnout decline come from? European Journal of Political Research 43(2): 221–236.

Bobek

Zaff

. (2009) Cognitive, emotional, and behavioral components of civic action. Towards an integrated measure of civic engagement. Journal of Applied Developmental Psychology 30(5): 615–627.

Bolzendahl

Coffé

(2013) Are ‘Good’ citizens good’ participants? Testing citizenship norms and political participation across 25 nations. Political Studies 61(S1): 45–65.

10.

Castells

(2000) The Rise of the Network Society. Oxford: Blackwell.

11.

Collins

Lanza

(2010) Latent Class and Latent Transition Analysis: With Applications in the Social, Behavioral, and Health Sciences. Hoboken, NJ: Wiley.

12.

Copeland

(2014) Value change and political action: Postmaterialism, political consumerism, and political participation. American Politics Research 42(2): 257–282.

13.

Dalton

(2007) The Good Citizen. How a Younger Generation is Reshaping American Politics. Washington, DC: CQ Press.

14.

Dalton

(2008) Citizenship norms and the expansion of political participation. Political Studies 56(1): 76–98.

15.

Dalton

(2013) Citizen Politics: Public Opinion and Political Parties in Advanced Industrial Democracies, 6th edn. Washington, DC: CQ Press.

16.

Dalton

Welzel

(eds) (2015) The Civic Culture Transformed: From Allegiant to Assertive Citizens. Cambridge: Cambridge University Press.

17.

Eckstein

Noack

Gniewosz

(2012) Attitudes towards political engagement and willingness to participate in politics: Trajectories throughout adolescence. Journal of Adolescence 35(1): 1–11.

18.

Eliasoph

(2013) Making Volunteers: Civic Life after Welfare’s End. Princeton, NJ: Princeton University Press.

19.

Fieldhouse

Tranmer

Russell

(2007) Something about young people or something about elections? Electoral participation of young people in Europe: Evidence from a multilevel analysis of the European Social Survey. European Journal of Political Research 46(6): 797–822.

20.

Flanagan

(2013) Teenage Citizens. Cambridge, MA: Harvard University Press.

21.

Giddens

(1991) Modernity and Self-Identity. Cambridge: Polity Press.

22.

Hagenaars

Halman

(1989) Searching for ideal types: The potentialities of latent class analysis. European Sociological Review 5(1): 81–96.

23.

Hooghe

Dassonneville

(2013) Voters and candidates of the future. The intention for electoral participation among adolescents in 22 European countries. Young: Nordic Journal of Youth Research 21: 1–28.

24.

Hooghe

Oser

Marien

(2015) A comparative analysis of ‘Good Citizenship’: A latent class analysis of adolescents’ citizenship norms in 38 countries. International Political Science Review. Available at: http://ips.sagepub.com/content/early/2014/07/24/0192512114541562.abstract

25.

Howe

(2010) Citizens Adrift: The Democratic Disengagement of Young Canadians. Vancouver, BC, Canada: UBC Press.

26.

Inglehart

(1997) Modernization and Postmodernization. Cultural, Economic, and Political Change in 43 Societies. Princeton, NJ: Princeton University Press.

27.

Inglehart

Catterberg

(2002) Trends in political action: The developmental trends and the post-honeymoon decline. International Journal of Comparative Sociology 43: 300–317.

28.

Inglehart

Welzel

(2005) Modernization, Cultural Change, and Democracy. The Human Development Sequence. New York: Cambridge University Press.

29.

International Monetary Fund (2013) World economic and financial surveys: World economic outlook database. Available at: http://www.imf.org/external/pubs/ft/weo/2013/01/weodata/index.aspx (accessed 22 February 2015).

30.

John

Fieldhouse

Liu

(2011) How civic is the civic culture? Explaining community participation using the 2005 English citizenship survey. Political Studies 59(2): 230–252.

31.

Kankaraš

Vermunt

(2014) Simultaneous latent class analysis across groups. In: Michalos

(ed.) Encyclopedia of Quality of Life and Well-Being Research. Heidelberg: Springer, pp. 5969–5974.

32.

Kankaraš

Moors

Vermunt

(2010) Testing for measurement invariance with latent class analysis. In: Davidov

Schmidt

Billiet

(eds) Cross-Cultural Analysis: Methods and Applications. New York: Routledge, pp. 359–384.

33.

Keel

Fichter

Quadflieg

. (2004) Application of a latent class analysis to empirically define eating disorder phenotypes. Archives of General Psychiatry 61(2): 192–200.

34.

Klingemann

(2015) Dissatisfied democrats. Democratic maturation in old and new democracies. In: Dalton

Welzel

(eds) The Civic Culture Transformed: From Allegiant to Assertive Citizens. Cambridge: Cambridge University Press, pp. 116–157.

35.

Magidson

Vermunt

(2004) Latent class models. In: Kaplan

(ed.) The SAGE Handbook of Quantitative Methodology for the Social Sciences. London: SAGE, pp. 175–198.

36.

Marien

Hooghe

Quintelier

(2010) Inequalities in non-institutionalized forms of political participation. A multilevel analysis for 26 Countries. Political Studies 58(1): 187–213.

37.

Marshall

Gurr

Ted Robert

Jaggers

Keith

(2011) Polity IV Project: Political Regime Characteristics and Transitions, 1800-2013, Dataset Users’ Manual. Vienna: Center for Systemic Peace, Available at: http://www.systemicpeace.org/inscrdata.html (accessed 22 February 2015).

38.

Martin

(2012) Political participation among the young in Australia: Testing Dalton’s good citizen thesis. Australian Journal of Political Science 47(2): 221–226.

39.

Norris

(ed.) (1999) Critical Citizens. Oxford: Oxford University Press.

40.

Norris

(2002) Democratic Phoenix: Reinventing Political Activism. Cambridge: Cambridge University Press.

41.

Norris

(2011) Democratic Deficit: Critical Citizens Revisited. Cambridge: Cambridge University Press.

42.

Nylund

Asparoutiov

Muthen

(2007) Deciding on the number of classes in latent class analysis and growth mixture modeling: A Monte Carlo simulation study. Structural Equation Modeling 14(4): 535–569.

43.

Oser

Hooghe

(2013) The evolution of citizenship norms among Scandinavian adolescents, 1999–2009. Scandinavian Political Studies 36(4): 320–346.

44.

Pharr

Putnam

(eds) (2000) Disaffected Democracies: What is Troubling the Trilateral Countries? Princeton, NJ: Princeton University Press.

45.

Quintelier

Hooghe

(2012) Political attitudes and political participation: A panel study on socialization and self-selection effects among late adolescents. International Political Science Review 33(1): 63–81.

46.

Raney

Berdahl

(2009) Birds of a feather? Citizenship norms, group identity, and political participation in Western Canada. Canadian Journal of Political Science 42(1): 187–209.

47.

Schudson

(1998) The Good Citizen: A History of American Civic Life. New York: Martin Kessler Books.

48.

Schulz

Ainley

Fraillon

(eds) (2011) ICCS 2009 Technical Report. Amsterdam: IEA.

49.

Shulman

Levine

(2012) Exploring social norms as a group-level phenomenon: Do political participation norms exist and influence political participation on college campuses? Journal of Communication 62(3): 532–552.

50.

Sloam

(2013) ‘Voice and Equality’: Young people’s politics in the European Union. West European Politics 36(4): 836–858.

51.

Smetana

Turiel

(2003) Moral development during adolescence. In: Adams

Berzonsky

(eds) Blackwell Handbook of Adolescence. Malden, MA: Blackwell, pp. 247–268.

52.

Stiglitz

(2007) Making Globalization Work. New York: Norton.

53.

Stolle

Hooghe

(2011) Shifting inequalities. Patterns of exclusion and inclusion in emerging forms of political participation. European Societies 13(1): 119–142.

54.

Straughn

Andriot

(2011) Education, civic patriotism, and democratic citizenship: Unpacking the education effect on political involvement. Sociological Forum 26(3): 556–580.

55.

Torney-Purta

(2001) Civic knowledge, beliefs about democratic institutions, and civic engagement among fourteen year olds. Prospects 31(3): 279–292.

56.

Torney-Purta

Lehmann

Oswald

. (2001) Citizenship and Education in Twenty-eight Countries: Civic Knowledge and Engagement at Age Fourteen. Amsterdam: IEA.

57.

Van Deth

(2007) Norms of citizenship. In: Dalton

Klingemann

(eds) The Oxford Handbook of Political Behavior. Oxford: Oxford University Press, pp. 402–417.

58.

Van Deth

Abendschön

Vollmar

(2011) Children and politics: An empirical reassessment of early political socialization. Political Psychology 32(1): 147–174.

59.

Vermunt

(2010) Latent class modeling with covariates: Two improved three-step approaches. Political Analysis 18: 450–469.

60.

Welzel

(2013) Freedom Rising: Human Empowerment and the Quest for Emancipation. Cambridge: Cambridge University Press.

61.

Welzel

Deutsch

(2012) Emancipative values and non-violent protest: The importance of ‘Ecological’ effects. British Journal of Political Science 42(2): 465–479.

62.

Whiteley

(2011) Is the party over? The decline of party activism and membership across the democratic world. Party Politics 17(1): 21–44.

63.

Wolf

Miller

Reardon

. (2012) A latent class analysis of dissociation and posttraumatic stress disorder evidence for a dissociative subtype. Archives of General Psychiatry 69(7): 698–705.

64.

Yang

(2006) Evaluating latent class analysis models in qualitative phenotype identification. Computational Statistics and Data Analysis 50(4): 1090–1104.

65.

Yuan

Wei

Wang

. (2014) Self-efficacy difference among patients with cancer with different socioeconomic status: Application of latent class analysis and standardization and decomposition analysis. Cancer Epidemiology 38(3): 298–306.

66.

Zukin

Keeter

Andolina

. (2006) A New Engagement? Political Participation, Civic Life, and the Changing American Citizen. New York: Oxford University Press.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.51 MB

0.53 MB