Keywords

1 Introduction

Different conceptions of citizenship involve different kinds of expected behavior from citizens. Liberal conceptions are minimalist in nature, based on the assumption that citizens should be free from restraints, meaning that nothing much beyond voting is expected. Republican conceptions, on the other hand, expect much greater participation by citizens as a means of ensuring personal freedoms and liberty. Communitarians also advocate for citizen participation in the civic life of the nation. Their focus is on ensuring the health and well-being of the community, which is seen to be more important than the concerns of individuals. Civic engagement is a common value in each of these conceptions, but its extent, purposes, and outcomes differ depending on the theoretical lens used to understand the role of citizens in a democratic society.

Despite the different emphases in these theories, each includes a role for civil disobedience and dissent (Rawls 1999; Vatter 2005; Pickett 2008). There are debates about whether freedoms allowed for individual dissent are adequate compared to the regime supporting processes embedded in each of the theories. Sparks (1997) argued, for example, that not enough attention has been paid in democratic theory to what she called “dissident citizenship.” Writing from a critical feminist perspective, she articulated in specific terms the oppositional politics of dissent, which she saw as:

…the practices of marginalized citizens who publicly contest prevailing arrangements of power by means of oppositional democratic practices that augment or replace institutionalized channels of democratic opposition when those channels are inadequate or unavailable. Instead of voting, lobbying, or petitioning, dissident citizens constitute alternative public spaces through practices such as marches, protests, and picket lines; sit-ins, slow-downs, and cleanups; speeches, strikes, and street theatre (p. 75).

Despite the acknowledged role of dissent in democratic theory, there has been little attention paid to dissent or protest as part of citizenship education. Social movement literature, on the other hand, places dissent at its core (Laschever 2017; Savyasaachi and Kumar 2014). Since many young people may end up participating in such movements, it seems vital to understand how dissent figures in their thinking about future civic engagement. It is particularly important in Asian contexts where protest activities and dissent from the status quo have been features of recent political activity in Chinese Taipei’s Sunflower Movement (Yang and Kang 2017), Hong Kong’s Umbrella Movement (Kwong 2018) and Korea’s Candlelight Revolution (Kim 2018). Hong Kong’s recently developed pro-independence movement (Ng and Kennedy 2019) suggests that unconventional forms of civic engagement may have a continuing role in the region’s political development.

The purpose of this chapter, therefore, is to explore the ways in which young people in Hong Kong, Chinese Taipei, and Korea view their future civic engagement, and the extent to which they are willing to consider legal and illegal forms of protest activity. First, a brief review of the literature will map current understandings of unconventional forms of civic engagement. An empirical study will then be discussed involving these three societies. Finally, following a summary of the results, the implications for understanding Asian students’ civic engagement intentions for the future will be discussed.

2 Literature Review

While “active” citizenship has been an important theme guiding much of citizenship education policy and practice over the past three decades, the focus has been on conventional civic engagement. This focus was highlighted in two studies that have played an important role in seeking to articulate the nature of active citizenship. Hoskins and Mascherini (2009), using data from the European Values Survey, developed indicators of active citizenship to measure progress by European countries in making active citizenship a priority. They used four item domains, one of which was Protest and Social Change. The protest items included “participating in a lawful demonstration, signing a petition, boycotting products and deliberately buying certain products for political, ethical or environmental reasons (ethical consumption)” (p. 465). In addition, there were items related to social participation, such as volunteering, community engagement, etc. There were not, however, items connected to illegal forms of protest with the implication that such civic engagement was not seen in European contexts to be an aspect of “active citizenship.”

In a second study, Hoskins et al. (2011) developed another set of indicators, this time using data from the Civic Education Study (CIVED) conducted by the International Association for the Evaluation of Educational Achievement (IEA) (Torney-Purta et al. 2001). In this study, the domains did not include any reference to protest activities of any kind. Moreover, “actions that were deemed against the law, (such as blocking traffic) … were not included in our final list, due to a lack of consensus within Europe as to whether they were reflective of civic competence” (p. 93). As Hoskins et al. (2011) point out, such activities could be understood as challenging the social cohesion that was seen to be an important aspect of the European Union’s approach to developing a conception of European citizenship.

Despite the reluctance of the indicator studies to include items related to illegal protest, successive large-scale assessments of civic and citizenship education have continued to include items that require students to indicate whether they would consider engagement in illegal protest as a form of future civic engagement (Schulz et al. 2010, 2018a). Following the approach of Torney-Purta et al. (2001), these studies included three items related to blocking traffic, occupying buildings, and writing graffiti. These items were scaled in all three studies as Illegal Protest and have been analyzed as part of IEA’s International Civic and Citizenship Education Study (ICCS) international reports, national reports, and subsequent secondary analyses. The picture that emerges from all three is naturally quite similar, even though researchers have focused on different aspects of the results.

In ICCS 2016 (Schulz et al. 2018a), there was little difference in the scale scores of students who indicated they would participate in legal protest and those who indicated they would participate in illegal protest activities. There were, however, large differences in gender (boys more likely to list illegal protest as a future civic activity) and those with lower levels of civic knowledge were more likely to indicate their intentions to participate in illegal activities (p. 98). In a secondary analysis of ICCS 2009 data (Schulz et al. 2010), the overall results suggested that, “in all countries, the average student did not intend to get involved in any of these forms of protest” (Schulz et al. 2010, p. 140). Otherwise, similar results were found regarding gender (boys rather than girls) and civic knowledge (lower rather than higher). Additionally, however, it was found that students with higher scores on “citizenship efficacy” were more likely to engage in illegal protest activities, while those with lower scores on “trust in institutions” were more likely to indicate they would be involved in illegal protest activities.

Compared to other forms of civic engagement, the intention to engage in illegal activities always ranks relatively lower than other forms of engagement. This was shown graphically in a recent study of youth in six European countries by Hoskins and Janmaat (2019). The preference for future civic engagement in these countries was: voting, legal protest, formal participation, and illegal protest, with mean scores ranging from just above 7 (for voting) to just below 3 (for illegal protest). There is little doubt that preferences for future civic engagement are with more formal political structures. One issue that is of interest here, both practically and theoretically, is how many young people will end up opting for illegal forms of civic engagement. It is an important question given Hoskins and Janmaat’s (2019) findings that showed different levels of support for illegal engagement across countries.

Chow and Kennedy (2014), drawing on ICCS 2009 data from the participating Asian societies (Shultz et al. 2010), used cluster analysis (Aldenderfer and Blashfield 1984) to identify groups that shared similar characteristics in terms of future civic participation. Using this person-centered analytic approach, they discovered that 29% of the sample would use illegal protest as a form of future civic engagement, ranging from 23.1% in Hong Kong to 39.2% in Korea. This did not mean that these individuals would not also participate in other ways, such as voting and legal protest, but it meant that they were willing to endorse illegal protest. Similar to the earlier studies discussed above, these supporters of illegal protest were characterized by gender (boys rather than girls) and level of civic knowledge (low rather than high).

Using latent class analysis, Kuang (2016) found a similar distribution of young people in Latin America and Europe to that of the Asian sample studied by Chow and Kennedy (2014). This suggests that intention to participate in illegal protest is a considered option across cultures and societies at least among a minority of students. This finding was supported by a recent study of 15-year-olds in China. While most Chinese adolescents showed that they were regime supporting in terms of their intended political trust and participation, 10% nevertheless indicated that they would consider engaging in illegal protest (Wang 2019). This is a surprising result given China’s authoritarian political system, yet it indicates that resistance is a disposition that some young people across political systems appear to value. Hoskins and Janmaat (2019) have raised two questions about those students who indicate they prefer illegal forms of civic engagement: “Are such students more engaged in general… or are they a specific group dismayed by the accepted, mainstream and turning to alternative ways to express their voice”? (p. 110). To this we would add: What is associated with students’ adoption of illegal protest as a form of civic engagement and what are the implications for policy, theory, and practice? We shall attempt to address these questions in the remainder of the chapter.

3 Method

3.1 Participants

The data used in this study were retrieved from ICCS 2016 (Schulz et al. 2018a), which measured students’ civic knowledge, attitudes, values, participation experiences, and intentions. It included three societies from Asia: Chinese Taipei, Hong Kong, and Korea. The sample included 9207 students: 2653 from Hong Kong (51.7% male, N = 1371); 2,601 from Korea (54.4% boys, N = 1414); and 3953 Chinese Taipei students (51.6% boys, N = 2040) (Schulz et al. 2018b, p. 47).

3.2 Measures

Future civic engagement scales for latent classes identification. Students’ expected participation in future legal protest (LEGACT) was measured by six items which asked students to express their opinions related to legal activities: “I would certainly do this-1,” “I would probably do this-2,” “I would probably not do this-3,” and “I would certainly not do this-4.” The activities include, for example, “Writing a letter to a newspaper” and “Taking part in a peaceful march or rally.” The items were recoded, thus higher values reflect greater likelihood of participation in related activities.

Students’ expected participation in future illegal protest (ILLACT), which asked students to express their opinions on three activities: spray-painting protest slogans on walls, blocking traffic, and occupying public buildings. Student responses are: “I would certainly do this-1,” “I would probably do this-2,” “I would probably not do this-3,” and “I would certainly not do this-4.” The items are recoded, thus higher values reflect greater likelihood of participation in related activities.

Students’ expected adult electoral participation (ELECPART), measured by three items, asked students to state what they thought they would do as adults: vote in local elections, vote in national elections, and get information about candidates before voting in an election. Student responses are: “I would certainly do this-1,” “I would probably do this-2,” “I would probably not do this-3,” and “I would certainly not do this-4.” The items are recoded, thus higher values reflect greater likelihood of participation in related activities.

Students’ expected adult participation in political activities (POLPART), measured by five items, asked students to state what they thought they would do as adults: help a candidate or party during an election campaign, join a political party, join a trade union, stand as a candidate in local elections, or join an organization for a political or social cause. Student responses are: “I would certainly do this-1,” “I would probably do this-2,” “I would probably not do this-3,” and “I would certainly not do this-4.” The items were recoded, thus higher values reflect greater likelihood of participation in related activities.

Using item response theory, these scales are transformed into weighted likelihood estimates with an average of 50 and a standard deviation of 10.

3.3 Other Measures

Socioeconomic background (SES) and gender are directly related to different forms of future participation (Kuang and Kennedy 2020). Students’ civic learning experiences, participation experiences, and other relevant civic values were identified as important variables in ICCS 2016 (Schulz et al. 2018b, pp. 151–176) and used as predictors in previous studies (Chow and Kennedy 2014; Kuang and Kennedy 2020). The current study, therefore, included these variables in the analyses.

Socio demographics. Gender: Boys were coded as 0; Girls were coded as 1. National index of students’ socioeconomic background (NISB) had a mean of 0 and a standard deviation of 1.

Participation experiences and attitudes. Discussion of political and social issues outside of school (POLDISC); Civic participation in the wider community (PARTCOM); Perceptions of openness in classroom discussions (OPDISC); Students’ engagement with social media for political use (SOCMED); Students’ perceptions of the value of participation at school (VALPARTS); Students’ willingness to participate in school activities (SCACT).

Civic beliefs and values. Students’ citizenship self-efficacy (CITEFF); Students’ attitudes towards gender equality (GENEQL); Students’ attitudes towards equal rights for all ethnic/racial groups (ETHRGHT); Students’ positive attitudes toward their country of residence (CNTATT); Students’ perceptions of the importance of personal responsibility for citizenship (CITRESP); Students’ trust in civic institutions (INTRUST).

Citizenship norms. Students’ citizenship norms endorsement were explored using item response theory (IRT) scores and using its latent class counterpart, citizenship norms profiles, develop in the present book (see Chap. 3). The IRT scores are: Student perceptions of the importance of conventional citizenship (CITCON) and Student perceptions of the importance of social movement related citizenship (CITSOC). The citizenship norms profiles consist of a nominal variable including comprehensive (students who present higher support for all citizenship norms), duty-based (students who support mainly traditional norms), socially-engaged (students who mainly support norms oriented to provide help in the community), monitorial (students who show mid-lower support for all norms), and anomic (students who express very low support for all citizenship norms).

School learning, experiences, and relations. Students’ experiences of physical and verbal abuse at school (S_ABUSE); Students’ perceptions of student-teacher relations at school (STUTREL); Students’ civic learning at school (CIVLRN); Civic knowledge (CK) was transformed into a metric with a mean of 500 and a standard deviation of 100. The overall reliability of CK was 0.84 (Schulz et al. 2018b, p. 133).

For more detailed information on these measures see Schulz et al. (2018b).

3.4 Analytic Techniques

Latent profile analysis (LPA) (Masyn 2013) was used on four future civic engagement scales to identify classes. As a model-based technique, it classified individuals and groups according to their probabilities. To get corrected standard errors for the stratified two-stage probability sample design of ICCS 2016, Taylor Series Linearization was used, where stratification indicators, primary sampling units, and student weights are used to get design based standard errors (Asparouhov and Muthén 2010). Students’ survey weights were scaled, so each country contributed equally to the estimations (Gonzalez 2012).

To determine the optimal number of classes in the analyzed observations, LPA offers a principled way of evaluating the optimal number of groups using selection criteria, for example, the Lo-Mendell-Rubin adjusted likelihood ratio test (LMRT) (Lo et al. 2001); Akaike information criteria (AIC) (Akaike 1973, 1974); Bayesian information criterion (BIC) (Schwarz 1978); sample size-adjusted Bayesian information criteria (sBIC) (Hix-Small et al. 2004; Schwarz 1978); and entropy value (Hix-Small et al. 2004). For LMRT, a significant p value suggested that the k cluster model improves the fit over the model with k-1 clusters. The smaller the value of AIC, BIC, or sBIC, the better was the model. Relative entropy was an indication of clear delineation of clusters, the closer that this value was to 1, the better of classification (Celeux and Soromenho 1996). Jung and Wickrama (2008) suggested values above 0.70 indicate acceptable classification accuracy.

In order to enhance confidence in the three-class solution, the study further tested the association between each class using a range of civic variables not used in developing that solution (for example, civic learning, civic values, civic participation experiences, student-teacher relations, among other measures). Beckstead (2002) pointed out the purpose was to “aid in substantive interpretation of the clustering solution and to provide validating support for the distinctiveness of the… clusters” (p. 316). This approach provided a measure of external validity for the chosen cluster solution.

The expected means of the civic variables were estimated using Taylor Series Linearization (TSL) for equally weighted countries, and the 95% confidence intervals (CI95) were retrieved for each latent profile realization for Hong Kong, Chinese Taipei, and Korea. These intervals were used to infer if there were real differences between classes in relation to the selected civic variables. If these intervals do not overlap, it indicated that there are mean differences between the classes, above the sampling error (Lumley 2010). This approach is more demanding than a t-test and other similar tests for mean comparison. As such, it works as a more robust option for mean difference tests (Goldstein and Healy 1995).

Multinomial/logistic regression (MLR) analysis is useful in predicting a categorical response variable using continuous and/or categorical explanatory variables. Generalized from binary logistic regression, the MLR model is appropriate for more than two levels of categories, which can be used to differentiate groups based on certain factors (Tansey et al. 1996). The MLR model can be used to determine the percent of variance in the dependent variable explained by the explanatory variables, to rank the relative importance of independents variables, to assess interaction effects, and to understand the relative importance of covariate control variables, and allows for comparison of more than one contrast simultaneously (El-Habil 2012). In the present study MLR was used to predict students’ future preferences of political participation profiles.

4 Results

LPA Classification. The AIC, BIC, sBIC, and LMRT fit indexes suggest models with more classes (see Table 1). As the number of classes increased, the BIC and sBIC decreased, although the improvement between the 6-class and 7-class models was small. A significant LMRT indicated that, in each case, adding a class improved the fit. For 2–6 class models, p-values were statistically significant, which indicated that the fit improved as classes were added. The LMRT for 7-class model was not statistically significant, which means the 7-class solution was no better than the six-class solution. Regarding classification utility, relative entropy values were different for different latent class solutions. Combining those indices, the study selected the three-class solution with a simpler structure and the highest relative entropy values (0.97). The three-class solution expected means is used in the following section, to describe the selected class solution (see Fig. 1).

Fig. 1
figure 1

Notes ELECTPART = Students’ expected adult electoral participation, POLPART = Students’ expected adult participation in political activities, LEGACT = Students’ expected participation in future legal protest, ILLACT = Students’ expected participation in future illegal protest

Mean values of three classes on future civic engagement.

Full size image
Table 1 Fit indexes for 2–7 latent class solution
Full size table

Class descriptions. Class 1: The “Moderates” were made up of 2872 Asian students (31.2%). Moderates had relatively high values on illegal and legal protest and political participation, with their lowest score being electoral participation. It seems that for Moderates, the first option would be protest, either legal or illegal. But, unlike “Radicals” (see below), their endorsement of other forms of political participation is more moderate.

Class 2: The “Rationals” were made up of 4910 Asian students (53.9%). Rationals had higher values for electoral participation and approached average scores for political participation and legal protest, with their lowest score on illegal protest. They are civically engaged, but their preference is for legal forms of engagement, especially electoral participation. Nevertheless, in general their endorsement of all forms of civic engagement is lower than both the Moderates and Radicals.

Class 3: The “Radicals” were made up of 1335 Asian students (14.6%). Radicals endorsed illegal protests more strongly and well above the average for other forms of civic engagement, including legal protest. This suggest that Radicals view illegal protest as the dominant form of such engagement although they do not neglect other forms of engagement that they endorse generally more highly than the other classes.

4.1 External Validity of Three Classes

The results of the comparisons between the selected civic variables and each of the classes indicated significant differences between classes on those variables (see Appendix D). The differences revealed substantive variation in the associations between the classes and the civic variables (see Fig. 2) and reinforced the substantive nature of the classes themselves.

Fig. 2
figure 2

Notes CIVLRN = Student reports on civic learning at school, CITCON = Students’ perception of the importance of conventional citizenship, CITSOC = Students’ perception of the importance of social movement related citizenship, CITRESP = Students’ perception of the importance of personal responsibility for citizenship, CITEFF = Students’ citizenship self-efficacy, OPDISC = Students’ perception of openness in classroom discussions, POLDISC = Students’ discussion of political and social issues outside school, SOCMED = Students’ engagement with social media, COMPART = Students’ participation in the wider community, SCHPART = Students’ participation at school, CNTATT = Students’ positive attitudes toward their country of residence, INTRUST = Students’ trust in civic institutions, ETHRGHT = Students’ endorsement of equal rights for all ethnic/racial groups, GENEQL = Students’ endorsement of gender equality, INTACT = Students’ perceptions of student interaction at school, SCACT = Students’ willingness to participate in school activities, STUTREL = Students’ perception of student-teacher relations at school, VALPARTS = Students’ perception of the value of participation at school, ABUSE = Students’ experiences of physical and verbal abuse at school

Comparison of the means for civic values and behavior by class.

Full size image

Rationals scored higher on civic learning, personal responsibility for citizenship, openness in classroom discussions, equal rights for all ethnic/racial groups, student interaction at school, and the value of participation at school than Radicals, who in turn scored higher than the Moderates. Rationals also scored higher on student-teacher relationship than Radicals and Moderates.

Radicals scored higher on the importance of conventional citizenship and social movement related citizenship, social media engagement, attitudes toward their country of residence, trust in civic institutions, and physical and verbal abuse at school than Rationals, who in turn scored higher than the Moderates.

Rationals and Radicals scored higher on the school participation and discussion of political and social issues outside school than the Moderates. There were no significant differences between Rationals and Radicals.

Radicals scored higher on citizenship self-efficacy, community participation, and willingness to participate in school activities than the Moderates, who in turn scored higher than the Rationals.

Rationals scored higher on gender equality than the Moderates, who in turn scored higher than the Radicals. Rationals also scored higher on student-teacher relationship than radicals and moderates.

Radicals scored higher on student-teacher relations at school than the Moderates and the Radicals. There were no differences between Moderates and Radicals.

4.2 Predicting Class Membership Using Multinomial Logistic Regression (MLR)

MLR was used to predict class membership using the civic related variables referred above and the citizenship norms profiles identified in this book. Three classes were used as outcome variables, with the Rational class defined as the reference group. Three societies were used as control variables by using dummy coding, treating Korea as the reference group (see Table 2).

Table 2 Predictors for class membership using multinomial logistic regression
Full size table

For students from Chinese Taipei (β = −0.72, OR = 0.48) and Hong Kong (β = −1.01, OR = 0.36), the odds of belonging to the Moderates relative to Rationals decreased by 52% and 64%, respectively.

For students who are girls (β = −0.11, OR = 0.89), who had higher SES (NISB, β = −0.09, OR = 0.91), who had higher scores on civic learning (CIVLRN, β = −0.10, OR = 0.90), and who had higher scores on school participation (β = −0.06, OR = 0.94), and civic knowledge (β = −0.50, OR = 0.60), the odds of belonging to the Moderates relative to Rationals decreased by 11%, 9%, 10%, 6%, and 40% respectively. For students who are more willing to participate in school activities (SCACT, β = 0.18, OR = 1.20), and who engaged more in social media (SOCMED, β = 0.09, OR = 1.09), the odds of belonging to the Moderates relative to Rationals increased by 20% and 9%, respectively.

For students from Chinese Taipei (β = −0.62, OR = 0.54) and Hong Kong (β = −1.04, OR = 0.35), the odds of belonging to the Radicals relative to Rationals decreased by 46% and 65%, respectively.

For students who belonged to the socially-engaged (β = −0.45, OR = 0.64) and duty-based (β = −0.44, OR = 0.64) groups, the odds of belonging to the Radicals relative to Rationals decreased by 36% and 36%, respectively. Other civic norms did not significantly predict future civic engagement as defined in this study. For girls (gender, β = −0.50, OR = 0.61), students who had higher scores on political discussion outside school (POLDISC, β = −0.12, OR = 0.89), school participation (SCHPART, β = −0.10, OR = 0.90), and civic knowledge (β = −1.13, OR = 0.32), the odds of belonging to the relative Radicals to Rationals decreased by 39% and 11%, 10%, and 68%, respectively.

For students who had higher scores on community participation (COMPART, β = 0.08 OR = 1.08) and who had higher scores on civic efficacy (CITEFF, β = 0.29, OR = 1.34), who are more willing to participate in school activities (SCACT, β =0.50, OR = 1.65), and who engaged more in social media (SOCMED, β = 0.25, OR = 1.28), the odds of belonging to the Radicals relative to Rationals increased by 8%, 34%, 65%, and 28% respectively.

5 Discussion

In this study, we have used a sample of Asian students from ICCS 2016 (Schulz et al. 2018a) to explore the extent to which illegal protest is considered as a form of future civic engagement by students. We opted for a person-centered analysis of the data that yielded three latent classes we called Radicals, Moderates, and Rationals. As a validity check, these classes were also differentiated by a range of civic values and behaviors. Most of these were strongly and positively associated with Rationals and Radicals but were much less strongly associated with the Moderates. Class membership was associated with a range of variables, including gender, SES, civic knowledge, and civic efficacy, and students’ willingness to participate in school activities, social media, and political discussion.

As in previous studies, we have attempted to assess the importance and relevance of illegal protest as a form of civic engagement. We have shown that of the three groups identified, two consider illegal protest as a form of civic engagement while for the other it is clearly not a preferred option. This provides a more holistic picture of support for illegal protest, since such support seems to be embedded in a set of complex decisions judging the best or most effective way to secure civic goals. Thus, opting for illegal protest does not necessarily mean that other forms of civic engagement are ruled out.

Radicals, by endorsing illegal protest most strongly and electoral participation least strongly seem to indicate that they will consider all forms of civic engagement, but their preference is clear. Moderates, on the other hand, also consider illegal protest as an option for the future but not as strongly as the Radicals. The Moderates seem more pragmatic than the Radicals accessing forms of engagement that will most help them. This suggests a somewhat nuanced approach to understanding illegal protest as a form of civic engagement. It helps to answer Hoskins and Janmaat’s (2019) questions: “Are such students more engaged in general… or are they a specific group dismayed by the accepted, mainstream and turning to alternative ways to express their voice?” (p. 110).

We would argue that the Radicals are simply more engaged. This can be seen particularly by contrasting their engagement with that of the Rationals and Moderates whose strongest endorsement is for electoral participation but weaker endorsement for other forms and an outright rejection of illegal protest. Thus, the answer to the questions above is that different groups endorse illegal protest for different reasons, suggesting that illegal protest serves different purposes. For Rationals, it is not on their future agenda at all. Moderates will consider it alongside other strategies, but their support is not strong, while Radicals may consider it as their first line of engagement. Thus, illegal protest is not a strategy of last resort for Radicals. For Moderates, it is certainly an option. Yet their level of endorsement suggests that their engagement is perhaps more pragmatic than that of the Radicals. Moreover, there is evidence that previous civic experiences are associated in different ways with the proposed future actions of the different groups.

We noted earlier that girls, who had higher scores on civic knowledge, civic learning at school, and school participation, and who had higher SES, tended to be members of the Rationals. Those with lower scores on the other hand, tended to fall into the Moderates group. Also for girls, those who had higher scores on civic knowledge, political discussion, and school participation, and who had lower scores on efficacy, community participation, social media engagement, and willingness to participate in school activities, tended to be members of the Rationals. Those with the opposite scores on the other hand, tended to fall into the Radical group.

We cannot tell from this data which factors exerted the strongest relationship with students’ views of their future civic engagement. But it does seem that students who are currently engaged in civic activities seem to consider the broadest range of actions for future civic engagement have the most positive view of their intentions and this includes engaging in illegal protest activities.

A common result from previous studies is that boys are more inclined to engage in illegal protest than girls (Schulz et al. 2018a; Ainley and Schulz 2011). Our results indicated that girls tended to fall into the Rationals group compared to the Moderates and Radical groups and boys tended to fall into the Radical and Moderates groups. The results were consistent with Kuang and Kennedy’s (2020) study that found boys are more likely to be Radicals. It is also of interest to note that high SES students were more likely to be members of the Rationals rather than the Moderates.

The overall picture painted by the results is that civic engagement for these Asian students is bound on the one side by what might be called a “status quo” view of the world where limited engagement is valued but knowledge and values are valued more. This status quo group is the Rationals representing the majority of students (53.9%). They have high levels of civic learning, commitment to personal responsibility, and valuing of classroom open discussion and equality. These students appear to value participation, but they do not plan to be overly engaged. At the other boundary are the Radicals representing the minority of students (14.6%). Their values are associated with participation of all kinds—both currently and in the future. Ironically, they are students who trust the government and have positive attitudes towards the nation, but they will use every tool available to be engaged and involved as a means of influencing the world around them. Then there are the Moderates representing 31.2% of the students. They will be engaged but it seems without enthusiasm. They have relatively low levels of civic learning compared to the other groups and less commitment to democratic values. This diversity of group attitudes and intentions suggests that decision making about civic engagement is by no means simple. It is not just a simple binary, such as “to vote or not to vote.” Rather it is about choosing between a range of possibilities influenced by context and commitment. An issue for the future is to determine what influences these choices by young people have and how can they be assisted to make good choices that will benefit the whole of society.

Finally, how do students’ intention for future civic engagement relate to the citizenship norms profiles developed in this book? For all the norms there was a negative relationship when membership of the Rationals and Radicals was considered. Yet only two of these relationships were significant—the socially-engaged and duty-based groups were more likely to be members of the Rationals than the Radicals. This result reflects two key ideas central to this chapter.

First, the socially-engaged, duty-based groups, and the Radicals will be active in different ways. That is, these different groups of students endorse different forms of civic engagement. While the first endorses the engagement in the local, and voting in national elections, the second, endorses only the participation in national elections. In contrast, the Radicals like these previous profiles, endorse participation in elections while also endorsing the participation in illegal protest as a form of future civic engagement. Thus, the chances are that members of the socially-engaged and duty-based groups are more likely to be members of the Rationals than the Radicals. Conversely, students from the comprehensive profile are more likely than students from the socially-engaged and duty-based profiles to endorse illegal forms of protest. This suggests that the Rationals have more in common with these groups than the Radicals, even though the Rationals also endorsed illegal protest as a form of future civic engagement. What the three groups have in common is a view of the future where they will be actively engaged, although not always in the same way.

Second, explaining the lack of significance related to the other civic norms may be related to the extended forms of engagement characterizing those norms compared to that of the groups in the current study. This highlights an important point concerning the nature and extent of civic engagement. We have shown in the current study that illegal protest will be considered by a minority of students and this represents one boundary. The civic norms show much more nuanced forms of civic engagement—sometimes quite passive and sometimes reflecting one form of engagement than another. All forms of civic engagement are important—from the radical to the passive. An important issue is seeing these different forms of civic engagement on a continuum and appreciating the choices young people have when it comes to their involvement in the future.

6 Conclusion

The results of this study are consistent with what has been observed recently in Asian contexts. More and more, both legal and illegal protest are being used to secure civic goals, whether it is the removal of a President as in Korea, advocating for universal suffrage as in Hong Kong, or protecting Chinese Taipei’s independence. This study has contributed to a better understanding of the use of illegal protest as a possible strategy for civic engagement. It is not always considered as the least preferable form of engagement. Indeed, it appears for some students to be a preferred form of engagement to be used perhaps when other forms will not achieve desired social or political objectives.

It seems important that more work is undertaken in this area and the current focus of much civic and citizenship education on conventional forms of citizenship engagement needs to be expanded. If young people are to make decisions about engaging in illegal protest, they need to be aware of the issues involved, the possible consequences and the likely outcomes. This is particularly so when one of the main influences on current civic engagement is social media, the effects of which remain largely unknown (Kennedy 2019). Young people immersed in social media, for example, need to learn how to recognize the “echo chamber” (Quattrociocchi 2017) and “filter bubble” (Curkovic 2019) effects so they can make decisions based on a broad range of information and available options for engagement.

Engaging in illegal protest is not an insignificant matter. Young people need to understand in detail what is involved, to make informed judgments about it and to be sure that this form of engagement will help them to secure their civic objectives. This will be an important role for civic and citizenship education in the future and hopefully the study reported here will benefit the ongoing decision making needed by policymakers and schools.