Keywords

1 Introduction

There is little doubt that democracy is under stress (Carothers and O’Donohue 2019; Foa and Mounk 2016) and, consequently, that civic and citizenship education needs to undergo a strategic reorientation (Kennedy 2019). In terms of civic education research, there has been significant focus on young people’s civic involvement, including different forms of political participation and political engagement. Comparatively, however, there has been less interest in how youth understand democracy, how such understanding develops, and how it could be enhanced.

Part of the reason for this is the assumption that “active citizenship” should be the aim of civic education. This is certainly a worthwhile goal adopted by supranational bodies such as the European Union (European Economic and Social Committee 2015) and advocated by many civic educators. Such an approach is very much a reflection of a “steady state” in global, regional, national, and local affairs where democracy and its values remain mostly unquestioned. In this context, the purpose is to urge all citizens, but young people in particular, to exercise their rights in the interests of creating fair and just societies. Nevertheless, the rise of populism has shown that citizens can be “active” in multiple ways, including being socially destructive and deliberately undermining democratic institutions (Kennedy 2019). Being “active,” therefore, is no longer enough: any call for action must be based on a clear understanding of what democracy is and unequivocal support for democratic principles.

In this chapter, we assess Asian student’s understanding of democracy, using measures present in the International Civic and Citizenship Education Study (ICCS) 2016 from the International Association for the Evaluation of Educational Achievement (IEA) (Schulz et al. 2018a). In particular, we focus on students from Chinese Taipei, Hong Kong, and Korea. Using the presented measures from ICCS 2016 as dependent variables, we explore different forms of learning opportunities as predictors in order to identify significant associations that may help to explain how democratic understanding is best supported. Finally, we assess the implications of the study for theory, policy, and practice in civic and citizenship education.

2 Understanding Democracy—The Missing Element in Research on Civic and Citizenship Education

Within the literature of civic and citizenship education, in the past two decades, there have been three broad focuses of research: civic engagement, civic values, and civic knowledge. The emphases of past research, however, have been somewhat skewed. Civic engagement has undoubtedly received more attention than the other two areas (Kennedy 2019). Moreover, across all areas, previous literature has emphasized a “civic deficit,” especially in relation to young people.

A major report in Australia in the 1990s, for example, provided the basis for a revival of civic education, arguing that young people needed to know more about the political system (Civic Education Group 1994). A similar argument was present in McCabe and Kennedy (2014), concerning a lack of civic knowledge by students in the United States. While in England this phenomenon was referred to as the “democratic deficit” (Kerr 2003, p. 3). The perception that there is a gap in young people’s civic knowledge and that more knowledge automatically produced better citizens was widespread. This assumption had a consequence since policymakers saw a key role for civic education as a tool to introduce such knowledge to students.

Nevertheless, not all researchers saw the civic deficit in this light. Perez-Diaz (2004), for example, saw the civic deficit as more related to civic engagement and a lack of understanding of the civic duties required of democratic citizens, especially in the European context. This view also found expression in the United States (Galston 2004; Westheimer and Kahne 2004; Esser and de Vreese 2007). In Australia, where voting is compulsory, the Australian Electoral Commission registered concern about declining youth participation that, on average, was much lower than that of the general population (Print et al. 2004). Subsequent international studies across different societies showed that adolescents showed little appetite for conventional political participation, such as joining a political party or running for office. In contrast, what these studies did show was that youth were eager for social participation (e.g., volunteering, supporting environmental issues) (Torney-Purta et al. 2001; Schulz et al. 2010, 2018a). Thus, the original discourse of “civic deficit” turned into a “deficit of civic participation,” and much effort was put into improving youth participation in politics and civic life.

Other researchers, however, were more focused on what might be called a “values deficit” (Kennedy 2019) and how values were reflected in different forms of civic knowledge. Under this light, civic knowledge is more than civic literacy, the knowledge of the underlying political structures that characterize democratic societies. It has been conceptualized as a more sophisticated form of knowledge. For example, Osler and Starkey (2006) argued for an expanded view of civic knowledge taking into account the social, economic, and cultural aspects of citizenship. Also, they advocated for a broader conception of global citizenship to meet the needs of an increasingly globalized society. This broader conception had direct implications for the kind of civic knowledge that should form part of civic education and, in particular, knowledge and understandings that reflected a more diverse society underpinned by a diverse set of values. Osler and Starkey’s (2006) view won some support, especially from academics, but encountered a contrary response from the British government.

The British government’s initiative supported what it is called Fundamental British Values in the school curriculum (Department of Education 2014). This represented a narrowing of civic literacy highlighting how “deficit” is defined from ideological perspectives concerning the nature of civic knowledge. In general, the civic knowledge debate consists of conflicting views about the nature of civic knowledge, and what students should know. Often this debate is based on a particular conception of the ideal citizen. These ideal citizens are either the cosmopolitan citizen of Osler and Starkey (2006) or the obedient or responsible citizens firmly supporting the nation state.

In these multiple discourses concerning “civic deficits,” there is a singular lack of attention to students’ explicit understanding of democracy itself. This gap is also present for the most part with the IEA civic and citizenship education studies (Schulz et al. 2010, 2018a), although Torney-Purta et al. (2001) seems to have been an exception. There has certainly been an assumption in the IEA studies that civic knowledge, civic participation, and civic values are all part of a broader democratic enterprise. The specific nature of that enterprise, however, rarely seems to be addressed, especially in later studies. It might be that the civic knowledge scale used in these studies can provide some insight into this issue, but since, for the most part, the test items that make up the scale are confidential, little is known about them. As mentioned earlier, the exception was the attempt made by Torney-Purta et al. (2001), asking students to define what is “good” and “bad” for democracy using a set of 25 different items. In Torney-Purta et al. (2001), these different items were not scaled, and only descriptive statistics were produced for each participating society. Some further insight, however, was provided into these items by a subsequent secondary analysis (Husfeldt and Nikolova 2003).

These researchers explained that in the IEA Civic Education Study (CIVED) (Torney-Purta et al. 2001), there was a specific interest in students’ understanding of democracy as a distinct construct from areas such as national identity and social cohesion. This set of items on democracy was not repeated in ICCS 2009 (Schulz et al. 2010), yet, it reappeared in ICCS 2016 (Schulz et al. 2018a). The international report of ICCS 2016 does not include an explanation as to why the emphasis on democracy was reinstated in 2016. We speculate, however, that the threats to democracy that appeared in the second decade of the 21st century (Foa and Mounk 2016; Kennedy 2019) may have influenced the decision to make democracy a renewed focus for the assessment of citizenship knowledge and understanding.

In their study, Husfeldt and Nikolova (2003) presented “civic knowledge and skills” and “understanding democracy” as two different constructs. They explained that the 25 different items presented in the Torney-Purta et al. (2001) study, conducted in 1999, were developed to measure students’ understanding of different forms of democratic models. These included the generic rule of law, the liberalism model, the pluralism model, the participatory model, the communitarian model, the social welfare model, and the elitism model. This set of items presented different situations that threaten or strengthen democracy. Students had to judge if these different situations were “good” or “bad” for democracy. Students’ responses for the 1999 study did not conform to the expected structure of each democratic model. Thus, Husfeldt and Nikolova (2003) used exploratory and confirmatory techniques in secondary data analysis and proposed a three-factor structure. The proposed factors were Rights and Opportunities, Limited Government, and Threats to Democracy. Yet no attention was paid to explaining students understanding of democracy. The present chapter aims to bridge this gap.

The three proposed factors presented a good fit to the data both for lower secondary and upper secondary students’ samples, present in the 1999 study (Husfeldt and Nikolova 2003). The items from Rights and Opportunities and Threats to Democracy were included in ICCS 2016. However, they were not scaled, and descriptive statistics were reported (Schulz et al. 2018a, p. 110). In the present chapter, following the work of Husfeldt and Nikolova (2003), we specified a two-factor latent structure to the reinstated items of understanding democracy. By using the Threats to Democracy and Rights of Opportunities item responses, we analyzed the influences on the understanding of democracy that students present. We selected different factors based on political socialization theory (Torney-Purta et al. 2001, pp. 20–22). The latter model assumed that participation was the outcome variable. Our model (see Fig. 1), on the other hand, made understanding of democracy the dependent variable, associated with students’ demonstrated civic knowledge, their citizenship norms endorsement, their civic engagement, both in an out of school, and their social experiences at school. Using these different factors, we sought to identify the most relevant factors that present a significant relationship with the students’ understanding of democracy.

Fig. 1
figure 1

Identifying predictors of student understanding of democracy: a proposed model

Full size image

3 Method

Data. The present chapter made use of secondary data from ICCS 2016 (Schulz et al. 2018a). This international study includes measures of students’ civic knowledge, civic attitudes, values, participation experiences, and intentions. Twenty-four societies took part, and among them, there were three societies from Asia: Chinese Taipei, Hong Kong, and Korea. The data from these societies were used in the current study. The nominal sample includes 9207 students: 2653 Hong Kong students, 51.7% boys (N = 1371); 2601 Korea students, 54.4% boys (N = 1414); 3953 Chinese Taipei students, 51.6% boys (N = 2040) (Schulz et al. 2018b, p. 47).

3.1 Variables

Dependent variables. Students were asked to judge nine situations related to democracy, indicating whether these were “good” or “bad” for democracy, using a three-option response scale of “Good for Democracy,” “Neither good nor bad for democracy,” and “Bad for democracy.” Following Husfeldt and Nikolova (2003), we specified a two-factor structure. The Threats for Democracy factor includes the items “Political leaders give government jobs to their family members;” “One company or the government owns all newspapers in a country;” “The police have the right to hold people suspected of threatening national security in jail without trial;” “The government influences decisions by courts of justice.” We scored the responses to these items as “Good for democracy” = 1, “Neither good nor bad for democracy” = 2, and “Bad for democracy” = 3. The remaining items represent the factor of Rights and Opportunities. This factor includes the following items: “People are allowed to publicly criticize the government;” “All adult citizens have the right to elect their political leaders;” “People are able to protest if they think a law is unfair;” “Differences in income between poor and rich people are small;” “All <ethnic/racial> groups in the country have the same rights.” They were scaled: “Good for democracy” = 3, “Neither good nor bad for democracy” = 2, and “Bad for democracy” = 1.

Independent variables. As independent variables we include demographics factors, civic knowledge and citizenship norms profiles, measures of student of civic engagement, measures of civic learning opportunities at school, and measures of the school experience. These different variables are described in the following sections.

Demographic variables. We included the sex of the students, and coded boys as 0, and girls as 1. The national index of students’ socioeconomic background (NISB) was included. This measure has a mean of 0 and a standard deviation of 1 for each included country. NISB is derived from three indices: the highest occupational status of parents, highest educational level of parents, and the number of books at home.

Civic knowledge is a continuous measure represented by five different plausible values generated from an item response theory model. These scores have a mean of 500 and standard deviation of 100 for the international pooled sample. The overall reliability of the scale was 0.84 (Schulz et al. 2018b, p. 133).

Citizenship norms. Five citizenship norms profiles summarize how students endorsed 12 different citizenship norms. This is a nominal variable, generated as latent class realization of a mixture model comparing all countries included in ICCS 2016 (see Chap. 3). This nominal variable classified students as socially engaged if they mostly endorsed norms relative to the participation in the community, the protection of the environment, and the promotion of human rights. In contrast, duty-based students mostly endorsed norms relative to following the rule of law, working hard, and voting in national elections. Students were classified as monitorial if they disregarded joining political parties and engaging in political discussion, while simultaneously endorsing the participation in protest against unjust laws. Anomic students disregard all presented citizenship norms. Finally, students in the comprehensive profile, were likely to deem all citizenship norms as important. In the present study, four dummy variables were created, with the comprehensive group used as a reference group for analysis.

Civic learning asked students to what extent they have learned about seven topics at school, this includes for example how laws are introduced in their country, and how citizens’ rights are protected. Students’ response categories were “to a large extent,” “to a moderate extent,” “to a small extent,” and “not at all.” The Cronbach alpha for this scale in the international sample is 0.98. For Chinese Taipei, Hong Kong, and Korea, it is 0.76, 0.81, and 0.80 respectively.

Open classroom discussion items asked students how frequently (“never,” “rarely,” “sometimes,” “often”) different situations happened during regular lessons when discussing of political and social issues. These items ask if teachers encourage students to make up their minds and express their opinions during this discussion. The Cronbach alpha for this scale for the international sample is 0.89. For Chinese Taipei, Hong Kong, and Korea it is 0.84, 0.87, and 0.90 respectively.

School participation items asked students if they had participated in seven different civic-related activities at school either “within the last twelve months,” “more than a year ago,” or “never.” This includes voting for school representatives, participating in school assemblies, and participating as a candidate. The Cronbach alpha for this scale for the international sample is 0.76. For Chinese Taipei, Hong Kong, and Korea, it is 0.72, 0.75, and 0.82 respectively.

Community participation asked whether students had participated in 10 different organizations, clubs, or groups in the wider community either “within the last 12 months,” “more than a year ago,” or “never.” The Cronbach alpha for this scale for the international sample is 0.89. For Chinese Taipei, Hong Kong, and Korea, it is 0.84, 0.87, and 0.90 respectively.

Political discussion required students to indicate their level of involvement in a series of activities outside of school using the response categories “never or hardly ever,” “monthly (at least once a month),” “weekly (at least once a week),” and “daily or almost daily.” The Cronbach alpha for this scale for the international sample is 0.78. For Chinese Taipei, Hong Kong, and Korea it is 0.76, 0.81, and 0.80 respectively.

Social media items asked students how often they were involved in three kinds of activities. This includes using the internet to find information, post, or comment political and social issues on the internet. The response categories were “Never or hardly ever,” “Monthly (at least once a month),” “Weekly (at least once a week),” and “Daily or almost daily.” The Cronbach alpha for this scale for the international sample is 0.69. For Chinese Taipei, Hong Kong, and Korea it is 0.64, 0.78, and 0.70 respectively.

Student-teacher relations contained five items assessing the degree to which students agreed or disagreed with statements about relationships in their school. Students’ response options ranged from “strongly agree” to “strongly disagree.” The Cronbach alpha for this scale for the international sample is 0.89. For Chinese Taipei, Hong Kong, and Korea, it is 0.89, 0.89, and 0.89 respectively.

Student interactions were measured by three items asking students to rate their level of agreement with the statements about students’ interpersonal relations at their school (ranging from “strongly agree” to “strongly disagree”). The Cronbach alpha for this scale for the international sample is 0.85. For Chinese Taipei, Hong Kong, and Korea it is 0.83, 0.85, and 0.86 respectively.

Physical/verbal safety was measured by six items, which asked students how often they experienced different abusive situations at their school during the last three months (ranging from “not at all,” “once,” “2 to 4 times,” “5 times or more”). A higher score reflects lesser chances of physical and verbal abuse at school. The Cronbach alpha for this scale for the international sample is 0.75. For Chinese Taipei, Hong Kong, and Korea it is 0.72, 0.78, and 0.68 respectively.

A summary of descriptive statistics of both dependent and independent variables can be seen in Table 1.

Table 1 Descriptive statistics for independent and dependent variables (unweighted estimates)
Full size table

3.2 Analytic Techniques

The data were analyzed using Mplus 8.0. Confirmatory factor analysis was used on the ICCS 2016 data to test the factor structure of the understanding democracy items originally proposed by Husfeldt and Nikolova (2003). Structural equation modeling was used to assess how much each of the selected predictors contributes to explaining students’ understanding of democracy. All selected factors were included as latent variables with their original items, with the exception of students’ gender, students’ socioeconomic background, citizenship norms profiles, and civic knowledge which were included as imputed data. Estimates are quasi pseudo maximum likelihood estimates (Asparouhov 2005) using the WLSMV estimator and “TYPE = COMPLEX” option in Mplus software, while including the study complex sample design, specifying schools as cluster, jackzones as stratification factors, and students’ total weights as survey weights. In the present chapter, standardized estimates are reported. Model fit indexes were used to assess the adequacy of the specified model to the data. These fit indexes included the comparative fit index (CFI; Bentler 1990), the Tucker Lewis index (TLI; Tucker and Lewis 1973), and the root of mean square error approximation (RMSEA). Values of CFI and TLI of 0.90 are interpreted as reasonable fit, and values over 0.95 as good model fit (Bentler 1990; Hu and Bentler 1999). RMSEA values of less than 0.08 suggest adequate fit (Bentler 1990; Hu and Bentler 1999; Joreskog and Sorborn 1993). We have included the unweighted descriptive statistics (see Table 1) of all included latent factors present in the structural model as mean scores of their respective items, to diagnose their differing variances. Civic knowledge scores were included as imputed values in the fitted model (Rutkowski et al. 2010), and re-scaled by dividing the original score by 100, so the unstandardized results present similar variances to the rest of the variables. This last procedure assures avoiding ill scaled matrixes in structural equational models (Kline 2016, p. 81).

4 Results

Measurement model. The results of the confirmatory factor analysis for the understanding democracy items supported a two-factor structure. For Chinese Taipei, CFI was 0.95, TLI was 0.94, and RMSEA was 0.07; for Hong Kong CFI was 0.98, TLI was 0.97, RMSEA was 0.09; and for Korea, CFI was 0.96, TLI was 0.94, RMSEA was 0.08. Following Husfeldt and Nikolova (2003) Factor 1 was named “Threats to Democracy” and Factor 2 was named “Rights and Opportunities.”

Structural equation model (SEM) results. The specified models present an acceptable fit to the data for Chinese Taipei (CFI = 0.94, TLI = 0.93, RMSEA = 0.02); and good fit for Hong Kong (CFI = 0.96, TLI = 0.96, RMSEA = 0.02) and Korea (CFI = 0.97, TLI = 0.97, RMSEA = 0.02). The selected factors account for substantial portion of the variance. For Chinese Taipei R2 is of 0.50 and 0.51 for Threats and Rights; for Hong Kong R2 is of 0.31 for Threats and 0.36 for Rights; while in Korea obtained R2 is of 0.41 for Threats, and 0.38 for Rights.

To inspect the relative associations of the selected variables we used standardized estimates (see Table 2). Holding all selected variables constant, demographic variables presented low estimates across regions and studied factors, with absolute estimates sizes ranging from 0.01 to 0.14. In the present study, girls scored lower on the two factors than boys in Chinese Taipei (Threats β = −0.04, p < 0.01; Rights β = −0.14, p < 0.001). In Hong Kong, girls scored lower than boys on identifying threats to democracy (β = −0.08, p < 0.01); while in Korea, girls scored lower than boys on Rights and Opportunities (β = −0.07, p < 0.001). Students’ socioeconomic background was positively related to Threats to Democracy and Right and Opportunities in the three societies, with estimates ranging from 0.05 to 0.12 except for Threats to Democracy in Korea which was not significant.

Table 2 Predictors of Asian students’ understanding of democracy
Full size table

Students’ civic knowledge was the highest predictor of all included variables. Its estimates were positively related to Threats to Democracy and Rights and Opportunities in the three societies, with estimates ranging from 0.52 to 0.69 across countries and understanding democracy domains. Citizenship norms profiles add additional explanatory power only among the Chinese Taipei students. In this society, relative to the comprehensive students, socially-engaged (β = 0.06, p < 0.001), duty-based (β = 0.05, p < 0.01), and anomic (β = 0.06, p < 0.01) students scored higher on Threats to Democracy. Conversely, socially-engaged (β = −0.09, p < 0.001), duty-based students (β = −0.09, p < 0.001), monitorial (β = −0.04, p < 0.01), and anomic (β = −0.12, p < 0.001) students scored lower on Rights and Opportunities, in comparison to the comprehensive students.

The different school activities included in the study yielded small absolute estimates, ranging from 0.01 to 0.11 across the studied societies and understanding democracy domains. Civic learning was positively related to Rights and Opportunities in the three societies (Chinese Taipei β = 0.06, p < 0.05; Hong Kong, β = 0.11, p < 0.01; Korea β = 0.08, p < 0.05). Meanwhile, open classroom discussion was positively related to Threats to Democracy (β = 0.06, p < 0.05) and Rights and Opportunities (β = 0.08, p < 0.05) in Hong Kong, but negatively associated to Threats to Democracy in Korea (β = −0.07, p < 0.01). Students’ school participation presents a small negative association to Threats to Democracy among Chinese Taipei students (β = −0.04, p < 0.05).

Among the included out of school activities of students, we observed small absolute estimates, ranging from 0.00 to 0.15. Community participation was negatively associated to Threats to Democracy in Korea (β = −0.09, p < 0.01). While political discussion was positively related to Threats to Democracy only (Chinese Taipei β = 0.07, p < 0.01; Hong Kong β = 0.12, p < 0.01; Korea β = 0.15, p < 0.01) with non-observable effects on Rights and Opportunities. Conversely, political social media use among students is negatively associated to Threats to Democracy (Chinese Taipei β = −0.10, p < 0.001; Hong Kong β = −0.12, p < 0.01; Korea β = −0.10, p < 0.05).

Finally, the different factors included here to study the relationship between understanding of democracy and students’ school experience, presented similar absolute sizes to the school activities factors, ranging from 0.00 to 0.15. Student-teacher relations negatively predicts Threats to Democracy (β = −0.10, p < 0.05) and is positively related to Rights and Opportunities (β = 0.12, p < 0.01) in Hong Kong. However, no other relations were observed among the rest of the societies for this factor. Student interaction was negatively related to Threats to Democracy in Chinese Taipei (β = −0.10, p < 0.01) and Korea (β = −0.15, p < 0.01). Students who report less frequent experience of physical/verbal abuse at school was positively related to Rights and Opportunities in Chinese Taipei (β = 0.06, p < 0.05) while negatively related to Threats to Democracy in Korea (β = −0.08, p < 0.01). Conversely, students who experienced more physical/verbal abuse at school would report lower scores on Rights and Opportunities in Chinese Taipei and higher score on Threats to Democracy in Korea.

5 Discussion

In this study, we used structural equation modeling (SEM) to explore the predictors related to Asian students’ understanding of democracy. The context for the study was the repeated claims by scholars (Fukuyama 1992; Tu 1996) that within Asia, there are competing ideological perspectives including local traditions such as Confucianism that, in some respects, run contrary to democratic values and understandings. Additional research has shown that Asian adolescents are subject to multiple political influences. However, they appear to manage and balance them as part of growing up in the region (see Kennedy 2021 for a review). For this reason, it seems crucial to understand how young people’s understanding of democracy can be developed and enhanced. The result of this study has provided some possible directions that have implications for both theory and practice.

In what follows, we shall focus on groups of variables that the study showed had differential effects on students’ understanding of democracy. These included demographics, civic knowledge, citizenship norms, and multiple learning opportunities available to students. The influence of these different factors was explored concerning the different domains of “Threats to Democracy and “Rights and Opportunities.”

It was mentioned earlier that apart from IEA’s CIVED (Torney-Purta et al. 2001), successive IEA large-scale civic and citizenship assessments have not distinguished understanding of democracy from civic knowledge. In the current study, we have drawn on items that were first used in Torney-Purta et al. (2001) as part of the Democracy domain, some of which were subsequently present in ICCS 2016 (Schulz et al. 2018a). The results showed that these items reflected two factors, Threats to Democracy and Rights and Opportunities, just as they had done in the previous secondary analysis of the original items (Husfeldt and Nikolova 2003). Nevertheless, there was a significant difference in the results of the present study.

Husfeldt and Nikolova (2003) found that while the resulting factors fit well for the upper secondary students, the fit of the models was less convincing for the lower secondary student responses. Thus, they reported results for upper secondary students. In the current study, however, the samples used from Chinese Taipei, Hong Kong, and Korea were for lower secondary students around 15 years old. The model fit in each of these societies ranged from acceptable for Chinese Taipei to good for Hong Kong and Korea. The obtained results in the present study suggest that for these Asian lower secondary students, unlike the Husfeldt and Nikolova (2003) samples, conceptions of democracy were reasonably well developed. There was a distinction, however, in their endorsement of the respective scales. They endorsed Rights and Opportunities more strongly than Threats to Democracy (see Table 1), suggesting greater confidence in understanding the former compared to the latter.

This result could reflect the fact that in all of these societies, democracy, local tradition, and authoritarianism have all been present as part of their social and political history. Thus, past historical experience may serve students to further their understanding of what strengthens or weakens democratic systems. It may be, therefore, that such experiences sharpen democratic understanding rather than blunt it. We present this idea as a tentative conclusion that requires further investigation.

Since the development of citizenship norms are an essential part of the current research project present in this book, these were included as independent variables in the fitted model. The results may seem disappointing since there were no observable effects with the exception of Chinese Taipei. Among Chinese Taipei students, engaged, duty-based, and anomic citizens scored lower than the comprehensive group on Threats to Democracy, and all citizenship norms profiles scored lower than the comprehensive profile on Rights and Opportunities. How is it possible to explain these results?

The citizenship norms and the groups they have generated reflect mainly the actions and character of different types of “good citizen.” They range from the duty-based citizen who is committed to upholding the status quo to socially-engaged citizens who are active in a broad range of public affairs to anomic citizens who deem no citizenship norm as important for adult citizenship. While monitorial citizens disregard political discussion, yet value the participation in protest against unjust laws. Finally, in the comprehensive profile, students endorse all citizenship norms, including conventional and non-conventional forms of civic participation. In these different forms of citizenship endorsement, there is no assumption in these norms that “good citizens” know anything about democracy, even though it is assumed that they are all participants in a democratic society. If this is an accurate depiction, it raises some important issues.

There is a conceptual distinction between participating in a democratic society and understanding the principles of a democratic society. This difference is irrespective of the nature of the participation and involvement itself. As discussed earlier, this focus on participation has been one of the main areas of research in the field of civics and citizenship education in the past two decades, and enhancing democratic participation is often seen as a policy priority. This study has shown that Asian students know about democracy, both the threats to it and the opportunities it provides, and they participate in different ways. An interesting question is whether there is a relationship between participating and knowing?

If students are aware of threats to democracy will this encourage them to participate in more active ways and if they know about the rights and opportunities democracy provides will they be more active in defending democracy? That is, is participation fueled by knowledge or is it an outcome of other processes of political socialization? The present study cannot answer this question, but it is an important question. Participation cannot be an end in itself since we see quite recently that participation in anti-democratic groups is on the rise (Kennedy 2019). More work is needed to develop the links between participating in democratic processes and knowing about democracy. In particular, does knowing more about democracy enhance democratic participation?

A final comment on the issue of both participating in and knowing about democracy is whether any conception of the “good citizen” needs to be expanded to include a knowledge component? In one sense, this is the exact role of much civic education in various countries: to support students in becoming more knowledgeable. However, it is probably the most criticized aspect of civic education since teacher-driven knowledge-based lessons are bound to deter most students from becoming engaged. In the following sections, we shall suggest some strategies, based on the results of this study, which might help to overcome the problem of lack of student engagement while at the same time enhancing knowledge and understanding.

In terms of the relationship of demographic variables with Asian students’ understanding of democracy, the results were mixed. Not surprisingly, socioeconomic status exerted a positive effect in all societies, except for a small non-significant effect on Korean students’ understanding of Threats to Democracy. These estimates were not large, but they were significant. These relations most likely reflect the associations of family cultural and social capital that, on the one hand, encourage learning, discussion, and a commitment to the status quo where parents benefit from the existing democratic system. On the other hand, where the benefits of democracy are not experienced, primarily through benefits and rewards form the economic system, then families are likely to pass on these negative perceptions. The present gap is a significant result about which teachers should be aware. What students bring to school with them in terms of values and understanding needs requires consideration in any civic education program.

The gender effects were noticeable. In general, holding all factors constant, boys seem to have a better understanding of democracy. In Chinese Taipei, this is true for both Threats to Democracy and Rights and Opportunities. In Hong Kong boys score higher on Threats to Democracy and in Korea on Rights and Opportunities. It is difficult to account for the differences across societies. These results are surprising, as adolescent males have previously also been found to be on the margins of civic learning in selected Asian societies (Kuang and Kennedy 2018). Thus, the present findings require further research. Considerable attention is needed to these issues and civic education programs need to be tailored accordingly. Civic education programs are not often thought of as being gendered, but engaging equally female and male students is essential, not just for their own sake but for the sake of society. For a variety of reasons, civic disengagement may well start in schools, but it should not be thought of as natural. Both males and females need to be knowledgeable about all aspects of democracy.

As shown below, there were multiple measures of student learning in the study, and the resulting associations suggested how the understanding of democracy might be enhanced. Nevertheless, not all forms of learning were equally effective, and some were effective in one place, but not another, and others were not effective at all. We need to understand both what works in promoting democratic understanding and what does not.

The civic knowledge (CK) scale used in ICCS 2016 (Schulz et al. 2018a) was a significant and robust predictor of understanding democracy in the three societies. While CK was designed “to measure a single trait labeled civic knowledge” (Schulz et al. 2018a, p. 10), the present results suggest that at least part of that trait is related to the understanding of democracy. The assessment framework for ICCS 2016 did not refer explicitly to democratic understanding (unlike IEA’s CIVED study in 1999 where one of the assessment domains was Democracy). Two of the domains, however, Civic Society and Systems and Civic Principles, that between them cover over 70% of the items in the cognitive test from which CK is constructed (Schulz et al. 2018b, p. 19), certainly can test student understanding democracy. Based on a description of these domains (p. 4), however, it seems they are loaded towards rights and opportunities provided by democracy but may not test any understanding about threats to democracy. The ability of students to identify factors that strengthen and weaken democracy is an essential issue for future ICCS studies because, given the times in which we live in, students need to be aware of threats that are manifesting themselves worldwide.

Civic learning in school (CLS) was a positive and significant predictor of Rights and Opportunities in the three societies, suggesting that school-based civic education programs are equipping students with some understanding of democracy. However, the association of CLS and Threats to Democracy was not significant in any of the societies. These previous results suggest that just as CK was not assessing student understanding of Threats to Democracy, so too CLS may not be preparing students to understand the threats facing democracy. We considered the present results as evidence of a new “civic deficit:” a lack of teaching students’ about the threats of democracy in modern democracies. It is no longer enough to know the strengths of democracy: there must also be a better understanding of the threats confronting democracy and how these can be resisted (Foa and Mounk 2016; Kennedy 2019).

When it comes to classroom learning processes, again, the results are mixed. Open classroom discussion (OCC) exerted non-significant effects across each of the societies, except for Rights and Opportunities in Hong Kong (a positive association) and Threats to Democracy in Hong Kong (positive association) and Korea (where the association was negative). These patterns of results suggest that OCC may not be the best way to promote understanding of democracy, where the learning is of a particular type. This result seems to contradict other research (Kuang et al. 2018), where OCC was shown to be a powerful classroom process for engaging students. However, being engaged in learning is not the same thing as learning specific content, so for the present purposes, OCC does not seem to help in understanding different aspects of democracy.

On the other hand, political discussion (PD), which focused on discussion outside of school, is positively and significantly related to Threats to Democracy in the three societies. These results suggest that such informal interactions play an important role in political socialization when it comes to understanding democracy. Nevertheless, this was not the case with Rights and Opportunities where there were no significant associations in any society, suggesting that informal political discussion may not have encompassed this perspective.

The result for Threats to Democracy and students’ traditional media use has implications for classrooms. The news media cover different examples of threats to democracy. Take, for example, news articles about Brexit, Donald Trump, right wing ascendency in Europe, and of course China’s long-standing authoritarianism. These different articles in all likelihood would dwell on different threats to democracy and it may be these that are picked up in political discussion outside the classroom. Yet such informal learning processes need to be reinforced by more deliberate classroom strategies in civic education. For example, students can be asked to initiate the discussion with their parents and friends on issues related to democracy and then teachers can arrange for ongoing classroom discussions to share their results. Integrating formal and informal learning is a crucial way to make the most of the multiple learning opportunities available to students (Calvo de Mora and Kennedy 2020).

Social media (SM) items were especially included in ICCS 2016 since it is now such a common form of engagement for young people. Yet in the present study, such involvement was not a significant predictor of understanding democracy, except for negative associations on Threats to Democracy for Chinese Taipei, Hong Kong, and Korean students. Although SM has been hailed as the new form of youth civic engagement (Kahne et al. 2016), such involvement does not guarantee that young people will have a better understanding of democracy. Indeed if the results are any indication, the effects of SM are adverse since students appear to have underestimated threats to democracy.

The links between SM, civic engagement, and understanding democracy need to be further investigated. The democratic promise of SM does not seem to have eventuated, and in many cases, the results of SM engagement turn out to be negative (Kennedy 2019). However, there is little doubting the attraction of SM in the digital age for both younger and older citizens. Since this study shows that in general, SM does not seem to support students’ understanding of democracy, the question is, what does it support? As in the case of participation (to be discussed below), simple engagement in SM is not a cognitively oriented activity, but it does have effects, particularly on young people. More needs to be known about those effects. What do students learn (since, on the whole, they are not learning about democracy from SM), and how can this learning be used positively to enhance not just engagement of democratic processes but the understanding of democracy itself. These are key questions for a future research agenda.

When it comes to student participation, both in schools and the community, there is very little pay off in terms of contributing to a deeper understanding of democracy. For school participation (SP), there were no significant associations across the three societies and for community participation (CP), only a small negative but significant association for students from Chinese Taipei on Threats to Democracy. In one sense, this should not be surprising since participation is an experiential learning strategy that has outcomes other than cognitive outcomes. Some studies, however, have shown an association between SP and CK (Kennedy et al. 2014). CK may be a single trait, as argued by Schulz et al. (2018b, p. 10), but it contains multiple domains that probably account for the association with SP. When it comes to what might be regarded as “harder” content such as understanding democracy, the association is washed out.

The results regarding different forms of participation by students are consistent with results reported earlier for the citizenship norms (see Table 2). These are participation-based norms and, like SP and CP, thus showed few associations with understanding democracy.

Focusing only on providing opportunities to participate is not a substitute for ensuring young people gain access to specific forms of democratic knowledge. Participation and understanding of democracy are distinct and uncorrelated dimensions, and this pattern is an important finding for civic education policymakers and teachers.

Students’ social school experiences (student-teacher relations, student interactions, and physical/verbal safety) are adapted from social learning theory that focuses on the development of individuals’ prosocial behavior (Grusec and Davidov 2015). The assumption is that young people can learn from the modeling of such behaviors. In the present study, however, the results showed an inconsistent pattern of associations across the three societies. Hong Kong was the exception where student-teacher relationships (REL) showed a positive association with an understanding of Rights and Opportunities and a negative association with Threats to Democracy. The remainder of the results showed no significant associations.

Modeling behavior provides an important opportunity for teachers (as well as peers and parents) to demonstrate expected positive behaviors for young people. What is clear from these results, however, is that such modeling has a limited impact on cognitive understandings, except in the case of Hong Kong and, to some extent Korea. It is difficult to explain why modeling appears to be effective in these places compared to others. What this suggests, however, is that more needs to be known about social learning and how it might be used more deliberately. As with participation experiences, social learning experiences do not have cognitive understanding as a major outcome. It may be that in the future, however, more needs to be known about social learning and especially ways in which it might interfere with student learning as well as enhance it.

6 Conclusion

Civic learning and political discussion outside the school emerged as the main predictors of understanding of democracy. Civic learning presents standardized coefficients ranging from 0.06 to 0.10, presenting itself as an enhancing factor of students’ understanding of rights across the three societies. Students’ engagement in political discussion presented similar estimates, ranging from 0.07 to 0.15 across the three societies, as an enhancing factor of students’ understanding of threats to democracy. These two factors together can be promoted by schools to enhance students’ understanding of democracy.

Nevertheless, some adjustments need to be made. The results of this study suggest that these learning processes may focus more on the rights and opportunities provided by democracy, with possibly less emphasis on the threats to democracy. What is more, other learning opportunities, both within and outside classrooms, seem to offer little support for learning about or understanding democracy. Therefore, if Asian students are to be fully knowledgeable about democracy, opportunities might need to be found to help students understand possible threats to democracy. The ability to understand what weakens and strengthens democracy is a crucial agenda for the future amidst a challenging global environment that often poses threats to democratic development.

While Asian students, as represented in this study, demonstrated a good understanding of democracy, it seems that many learning opportunities that could supplement this understanding are missed in schools. Keeping in mind that Asian students are confronted with multiple ideological influences on their civic learning, it is crucial to take as many opportunities as possible to reinforce democratic understanding. Schools continue to play an important role in developing and supporting this understanding relying on formal school-based civic education programs. Formal learning experiences, however, are not enough and can be supplemented by informal learning opportunities such as social learning, learning from participation, and constructive learning from social media engagement. The successful combination of formal civic learning opportunities and informal civic learning opportunities may require further research. Learning for democratic understanding is an essential goal for the future. This study offered insights on both the nature of that understanding and possible practices that enhance Asian students’ understanding. It remains an essential and ongoing agenda both for the students and the societies they represent.