2.1 Life-cycle theory and youth participation

The subject of youth participation in politics has been an ongoing concern since the early 1990s (Bennett, Reference Bennett1997; Park, Reference Park, Jowell, Curtice, Park, Brook, Thomson and Bryson2000; Henn et al., Reference Henn, Weinstein and Wring2002; Kimberlee, Reference Kimberlee2002). Studies dealing with youth political engagement often treat young people as a generation apart. Not only do young citizens turn out in lower numbers to vote than older generations (Russell et al., Reference Russell, Fieldhouse, Purdam and Kalra2002), but also they are not interested in, engaged with, or concerned about politics (O'Toole et al., Reference O'Toole, Lister, Marsh, Jones and McDonagh2003), have relatively weak commitments to political parties, and are less likely than older people to be members of political organizations (Parry et al., Reference Parry, Moyser and Day1992; Park, Reference Park, Jowell, Curtice, Park, Brook, Thomson and Bryson2000; Kimberlee, Reference Kimberlee2002). This decline in the political involvement of younger generations and the decreasing levels of youth participation not only endanger the democratic representativeness of today, but also jeopardize the democracy of tomorrow.

Previous research attributes these generational differences to age (Parry et al., Reference Parry, Moyser and Day1992; Russell et al., Reference Russell, Fieldhouse, Purdam and Kalra2002). Unlike older people who have been socialized for much longer and hence have accumulated more political resources and experiences through life, young people, as newcomers to politics, are chronically politically apathetic and lack interest in and connections to politics (Parry et al., Reference Parry, Moyser and Day1992; Park, Reference Park, Jowell, Curtice, Park, Brook, Thomson and Bryson2000; Kimberlee, Reference Kimberlee2002; Henn et al., Reference Henn, Weinstein and Forrest2005). As young people become older, socialized, and educated, their own cumulative political experiences will enable them to make more informed choices with regard to electoral and political processes (Quintelier, Reference Quintelier2007, 173).

2.2 The Internet and political participation

Witnessing the dramatic growth in communication technologies and the rapid rise of online social networks such as Facebook, Twitter, Plurk, YouTube, etc., scholars have pinned their hopes on the Internet for enhancing political participation, especially among young voters (Delli Carpini, Reference Delli Carpini2000; Xenos and Foot, Reference Xenos, Foot and Lance Bennett2007). The reasoning behind these expectations consists of four elements (Strandberg, Reference Strandberg2006). First, the Internet is seen as providing opportunities for spontaneous engagement in politics, as online voting in polls, debating, blogging, and so forth all allow end users to freely express their opinions (Strandberg, Reference Strandberg2006; Ward and Vedel, Reference Ward and Vedel2006). Second, with the popularity of smartphones and other mobile devices, the diffusion of ICTs should widen the penetration rate of the Internet and further lower the entry barrier for participation in online politics (Best and Krueger, Reference Best and Krueger2005). Third, the Internet provides up-to-date information essential for participating in civic life and public discussions and a level playing field for the political opposition (Delli Carpini and Keeter, Reference Delli Carpini and Keeter1996; Hale et al., Reference Hale, Musso, Weare, Hage and Loader1999; Tolbert and McNeal, Reference Tolbert and McNeal2003). Last but not least, social media permit users to exchange information, ideas, and opinions of all sorts (messages, images, videos, and files), and to openly debate with other users (Barber, Reference Barber2003; Shalom, Reference Shalom2005; Soep, Reference Soep2014). The interactive nature of social media has increased the efficiency of direct democracy and facilitated relations between citizens, political parties, and politicians (West, Reference West2004; Di Gennaro and Dutton, Reference Di Gennaro and Dutton2006; Steinberg, Reference Steinberg2015, Reference Steinberg and Çoban2016). In short, social media could solve collective action problems that have long impeded the participation in mainstream politics.

Some scholars recognize the significant potential of the Internet to enhance communication both horizontally among members of the public and vertically between the public and government officials, but suspect that the Internet merely reinforces existing patterns of political engagement and thus magnifies the gap between activists and apathists (Norris, Reference Norris2001, 228; Gibson et al., Reference Gibson, Lusoli and Ward2005; Di Gennaro and Dutton, Reference Di Gennaro and Dutton2006, 300). They examined the Internet usage of young adults and found that the ‘disengaged’ were largely of low socioeconomic status and were less likely to interact with political sites and to visit civic webpages and further concluded that the Internet appears to exacerbate the socioeconomic biases already exhibited in civic and political participation (Vromen, Reference Vromen2003, Reference Vromen2007; Weber et al., Reference Weber, Loumakis and Bergman2003; Livingstone et al., Reference Livingstone, Bober and Helsper2004). Other scholars acknowledge the utility of social media for facilitating online political engagement and the aggregation of individuals around common causes, but question its effectiveness in creating a collective identity (Fenton and Barassi, Reference Fenton and Barassi2011; Juris, Reference Juris2012) and enhancing offline participation (Brants et al., Reference Brants, Huizenga and van Meerten1996; Streck, Reference Streck1997). Their studies show that although young people use the Internet to acquire political information more frequently, they are no more likely to vote in national elections (Bimber, Reference Bimber1998, Reference Bimber2001; Couldry et al., Reference Couldry, Livingstone and Markham2006).

Recent studies further remind us that in spite of the Internet's undeniable advantages, it is like a double-edged sword. On the one hand, social media can serve as a ‘liberation technology’ which allows people to communicate with peers about current events, gather like-minded comrades, organize protests and movements, and support political candidates and parties (Diamond, Reference Diamond, Diamond and Plattner2012). On the other hand, it might stimulate online social extremism and political polarization and thus jeopardize the foundation of democracy. Political psychologists argue that to avoid cognitive dissonance and attitudinal ambivalence, online users tend to be motivated by directional goals and selectively expose to information that simply confirms their prior views (Taber and Lodge, Reference Taber and Lodge2006, 757; Slothuus and de Vreese, Reference Slothuus and de Vreese2010; Bakshy et al., Reference Bakshy, Messing and Adamic2015) and deliberately interact with like-minded. Such segregated information environments not only facilitate rumors, fake news, extreme attitudes, and misperception of facts (Flynn et al., Reference Flynn, Nyhan and Reifler2017; Lazer et al., Reference Lazer2018), but also monger hatred toward others and foster online cleavages (Bessi, Reference Bessi2015; Del Vicario et al., Reference Del Vicario, Vivaldo, Bessi, Zollo, Scala, Caldarelli and Quattrociocchi2016; Lazer et al., Reference Lazer2018; Stewart et al., Reference Stewart, Arif and Starbird2018).

2.3 The Sunflower Movement and its research merits

Marking the first time in which the legislature had been occupied by citizens in Taiwan, the Sunflower Student Movement showed that young Taiwanese are not uninterested in or apathetic to politics. On 17 March 2014, the KMT attempted to void the agreement with the opposition Democratic Progressive Party (DPP) and to force the CSSTA to vote for ratification in the KMT-controlled legislature. This unilateral move triggered the student occupation of the parliament the following day. From 18 March to 10 April 2014, hundreds of university students and activists occupied the Legislative Yuan to protest against the passing of the controversial CSSTA without a clause-by-clause review. Because Taiwan's media environment had been highly polarized between the incumbent KMT and the opposition DPP, at the outset of the Sunflower Movement, traditional media merely simplified the occupation as a conflict between the two camps. Therefore, to overcome the limitations of traditional media and to keep the world up to date about the constant goings-on of the movement, student occupiers waged their battles on social media platforms. Among the industrial countries listed in Figure 1, Taiwan has the highest social media penetration rate. About 80% of the population has actively used social networking services to connect with friends, share information and interests, play games and more. Figure 2 further demonstrates that Facebook, YouTube, Instagram, Line, and Google are the most popular social media platforms among Taiwanese netizens. Given the interactivity, connectability, and easy accessibility of social media, the students decided to make use of Facebook, Twitter, and other popular bulletin boards and online forums to disseminate their appeals, mobilize supporters, and challenge the mainstream media (Ho, Reference Ho2015; Lo, Reference Lo2015).

Figure 1. Social media penetration rate in industrial countries and China (Source: Statista; https://www.statista.com/).

Figure 2. Leading social networking services in Taiwan (Source: Digital in 2018 in Eastern Asia Essential Insights).

The intensive use of social media had significantly aroused public concern and anxieties over the political consequences of closer economic relations with China, with some citizens worrying that the increasing economic dependency on China might erode Taiwan's political autonomy and may result in the loss of their political liberties. On 30 March, 500,000 people rallied in the streets demanding that the CSSTA first be withdrawn from the legislature, the bill on Cross-Strait Agreement Supervision (CSAS) be enacted, the CSAS bill be legislated before the legal review of the CSSTA could take place, and a citizens' constitutional conference be convened. The Sunflower Movement thus became the last straw for President Ma and his KMT government. Ma's disapproval rating rose from 64 to 74.5% and his approval rating was down to 13.6%. On 29 November 2014, in the Taiwanese local mayors and magistrates elections held in Ma's final term, the KMT-controlled areas dropped from four municipalities and eleven counties to only one municipality and five counties. Thus began the KMT's secession of power to the DPP, culminating in the DPP candidate Tsai Ing-wen being elected and the KMT's seats in the legislature dropping from 64 to 35 in the 2016 presidential and legislative elections.

The Sunflower Student Movement stands in stark contrast to the life-cycle and motivated reasoning theories and serves as an inspiration to young people in overcoming their political apathy and lack of political efficacy by bringing them back to politics via the Internet. First of all, the life-cycle theory oversimplifies the differences in turnout rates by attributing them to generational effects. Recent studies (Quintelier, Reference Quintelier2007, 176) and the Sunflower Movement both illustrate that younger generations are no less interested in politics than their older counterparts. Second, the life-cycle theory critiques young people's lack of participation according to a very narrow conception of ‘formal' politics – i.e. politics that is concerned with the formal institutions of government, the main political parties and traditional forms of political behavior such as voting in elections (Henn et al., Reference Henn, Weinstein and Wring2002). Despite its importance in the regulation of democratic systems, voting is not the only form of political participation. Numerous recent studies and the Sunflower Movement both underscore this idea by demonstrating that new forms of participation have diverted younger generations from traditional forms of political engagement practiced by older generations (cf. Norris, Reference Norris2001; Quintelier, Reference Quintelier2007, 165).

Third, the intensive use of social media during the Sunflower Movement shows that the Internet not only has the potential to expand the base of younger citizens' political participation, but also to have inspired their concerns about and interest in public affairs and enhanced their sense of political efficacy. According to post-Sunflower Movement surveys, the percentage of young adults interested in politics increased from only 25% in 2010 to 50% in 2014 and that of young people who thought they had influence over what the government does also expanded from 34 to 50% (Data Source: TEDS 2010, ABS Wave IV). Last but not least, instead of creating an echo chamber which reinforces and amplifies the belief of its supporters by rumors, fake news, and false or dubious information, the Sunflower Movement disseminated their appeals, recruited supporters, and challenged the mainstream media by providing the most real time and transparent information from the occupied legislative floor (Lo, Reference Lo2015).

The above discussion highlights a series of important questions that emerge when studying Taiwan's political participation in the digital era. Given that we know that student activists who participated in and supported the Sunflower Movement used the Internet extensively to both coordinate among themselves and to disseminate their appeals to the public, to what extent did the Movement and the associated Internet use have a broader impact on democratic political participation? If online political information, like its offline counterpart, can encourage voter turnout (Matsusaka, Reference Matsusaka1995; Feddersen and Pesendorfer, Reference Feddersen and Pesendorfer1996, Reference Feddersen and Pesendorfer1999; Stromberg, Reference Stromberg2004), we should observe a consistent and positive association between the use of the Internet and political participation. However, if political participation was temporarily inspired by the Sunflower Movement via its intensive use of the Internet, the positive relationship between the Internet usage and political participation should only exist in the post-Sunflower Movement period, but not in other elections. The second question looks at the different impacts of the Internet on different generations. Did use of the Internet have a consistent effect on the political participation of different age groups or did it affect different generations in different ways? If online political information can inspire citizens' political participation, young adults who are more skilled at zooming from site to site, sorting, sifting and assessing online information than older citizens should be more efficient in turning their online political engagement into political participation in the real world.

Given the Internet's accessibility, interactivity, and connectability, I argue that it is hard to deny its potential in promoting political participation, especially that of young adults. However, because online behavior is often considered negligent and exaggerated, prejudice surrounding the Internet thus hinders its interactivity from the bottom up and delimits the development of online democracy. Although the Internet has widely been thought to facilitate government efficiency, considering Internet security, the representativeness of the end user, and the reliability of online information, the government rarely allows voters to officially exert their influence on policy making or hold their elected representatives accountable via ICTs. Consequently, social media often serves as a weapon of opposing activists, allowing them to disseminate their critiques on government or politicians. As shown in the Sunflower Movement, online cynicism can even bring down a government when it arouses enough sympathy. Nevertheless, it also fades away quickly and has no long-term impact when the discontent was covered by the immerse information on the Internet.

3.1 The Internet and electoral participation in Taiwan

Figure 3 depicts the generational differences in voting turnout in recent elections in Taiwan. In line with the findings of previous studies on youth participation in politics, young Taiwanese citizens are less engaged in voting. Their average turnout rates are 10–20% lower than that of older adults. I then classify respondents into four different types along with their frequency of searching for information about elections, politics, and government on the Internet: ‘No Use,' ‘Occasional Use,' ‘Regular Use,’ and ‘Frequent Use.’ To dynamically investigate whether use of the Internet is differently associated with older and younger generations' turnout, I estimated a logit model to predict voting turnout as a function of respondents' generation, the frequency of their searching for political information online, and the interaction of the two variables. If the Internet, as suggested by previous studies, exerts greater influence on younger people than on older adult cohorts, we could expect that online political information has a greater marginal association with younger people's voter turnout than on that of older citizens across the three elections. To derive a more robust estimation, I further controlled for respondents' frequency of discussing politics (from [1] never to [4] often), gender ([1] male, [2] female), education level (from [1] illiterate to [13] graduate school), and partisanship ([0] no party affiliation and [2] with party affiliation), which have been thoroughly discussed in studies of voting behavior (see the Appendix for details).

Figure 3. Generational difference in electoral participation (Source: Chu (Reference Chu2012), Huang (Reference Huang2014, Reference Huang2016)).

Table 1 presents the results from a multivariate logit analysis of respondents' turnout in the most recent three elections in Taiwan. Since the TEDS 2014 survey did not draw on a national sample but was only conducted in the three municipalities of Taipei, Taichung, and Kaohsiung, in order to derive a more reliable comparison I also estimate samples from the municipalities in the TEDS 2012 and 2016 datasets. As shown in Table 1, the two different samples yield the similar results. The statistics in the first row of Table 1 show that the generational difference in voting engagement between younger and older adults is consistent with the findings of most youth studies and Figure 3. It reveals that, other things being equal, the younger age group was less likely to vote than older adults. The coefficients in the second and third rows respectively indicate the average marginal effect of Internet use on the older group's voting participation and the average difference in the marginal effect of Internet use on the younger and older groups' voting participation. To facilitate our understanding of the statistics, Table 2 calculates the estimated marginal effect of the Internet use on different generations' turnout rates, showing that the 2014 election stands out from the other two as the only election in which the Internet had a significantly positive association with the younger group's voting participation. Figure 4 further illustrates the predicted marginal effect of the Internet usage on different generations' turnout rates in the 2014 election and their corresponding 95% confidence intervals derived from Table 1. As depicted in the figure, as the Internet usage increases from regular use to frequent use for the younger group, the predicted turnout rate of that election increases from 85.1 to 88.7%. However, the same increase in the internet usage does not yield significant effect on older adults' voting turnout.

Figure 4. Marginal influence of Internet use on voting engagement.

Table 1 The generational difference in the marginal effects of the Internet on voting turnout

***p < 0.01, **p < 0.05, *p < 0.1.

Date Source: Chu (Reference Chu2012), Huang (Reference Huang2014, Reference Huang2016).

Table 2 Estimated marginal effect of the Internet usage on voting turnout for different generation

***p < 0.01, **p < 0.05, *p < 0.1.

3.1.1 Non-electoral participation in Taiwan

Next is an examination of the association between Internet usage and non-electoral participation using the ABS Wave IV data, which was collected between June and September 2014, only a few short months after the Sunflower Movement. The survey asked interviewees whether they had contacted politicians, elected officials, or representatives. It also queried their experiences of participating in self-help activities, including raising issues, signing petitions, or attending demonstrations, protests, or marches for political ends (see the Appendix for details). I applied the two batteries above to generate two corresponding indicator variables, lobbying and self-help. If a respondent had practiced any of the activities above at least once in the past year, I coded the corresponding variable as 1; otherwise it was coded 0 (see the Appendix for details). Then the same interaction model was applied to estimate the association between Internet use and respondents' non-electoral participation. If the interactivity, communication feedback, and faster conveyance of information of the ICTs, as suggested by sociologists, could be used in organizing and maintaining social activities and in accomplishing e-participation (Biernacka-Ligieza, Reference Biernacka-Ligieza and Çoban2016, 112; Çoban, Reference Çoban and Çoban2016, vii), we should be able to observe that access to online political information also has a positive relationship with young citizens' participation in lobbying and self-help activities.

Table 3 shows that while control for citizens' frequency of discussing politics, gender, education levels, and partisanship, the estimated coefficient of the internet usage turns out to be insignificant. In other words, access to online political information has no association with older adults' participation in lobbying and self-help activities. In contrast to use of the Internet for political causes, citizens' offline social–political engagement and their partisanship are more significantly associated with their non-electoral participation.

Table 3 The generational difference in the marginal effects of the Internet on non-electoral participation

***p < 0.01, **p < 0.05, *p < 0.1.

Data Source: ABS Wave IV.

Moreover, Table 4 demonstrates that Internet usage had different associations with younger age groups' participation in different non-electoral activities. As demonstrated in Figure 5, all other things being equal, a one-unit increase in the frequency of Internet usage for political causes is significantly associated with a 3.9% increase in young adults' participation rate in self-help activities (standard error: 0.009). However, the same unit increase in Internet usage is insignificantly associated with a 0.2% increase in young citizens' participation in lobbying activities (standard error: 0.005). In other words, while the use of the Internet for political purposes has a positive association with young citizens' engagement in petitions and protests, it neither encouraged older citizens' non-electoral participation nor inspired young citizens to contact influential people.

Table 4 Estimated marginal effect of the internet usage on self help activities and lobbying practice for different generation

***p < 0.01, **p < 0.05, *p < 0.1.

Figure 5. The generational difference in the marginal effects of the Internet use on non-electoral participation.

Why did Internet usage, as demonstrated by the findings above, boost young adults' participation in electoral and self-help activities, but not inspire them to carry out lobbying activities in 2014? The Sunflower Movement was ignited by the ruling party's controversial maneuvering over the CSSTA, and the occupation of the Legislative Yuan was a sign of students' disenchantment with representative politics. Mobilized by the Internet and the Sunflower Movement, younger citizens decided to take action themselves rather than relying on conventional media, political parties, politicians, or influential celebrities.