The influence of online forum and SNS use on online political ...

Telematics and Informatics 30 (2013) 359–369

Contents lists available at SciVerse ScienceDirect

Telematics and Informatics journal homepage: www.elsevier.com/locate/tele

The influence of online forum and SNS use on online political discussion in China: Assessing ‘‘Spirals of Trust’’ Yi Mou a,⇑, David Atkin b, Hanlong Fu c, Carolyn A. Lin b, T.Y. Lau d a

Faculty of Humanities and Arts of Macau University of Science and Technology, Macau Department of Communication Sciences at University of Connecticut, USA c Humanities Department of Colby-Sawyer College, USA d University of Southern California, USA b

a r t i c l e

i n f o

Article history: Received 23 November 2012 Received in revised form 10 April 2013 Accepted 15 April 2013 Available online 22 April 2013 Keywords: Social media Online forum SNS Political discussion China

a b s t r a c t Echoing the significance of mobile online networks in fueling the Arab Spring, the present study seeks to better understand social media influences in China by studying political activity among Chinese netizens. A survey of Chinese college students examines the influence of online social networks in the context of political attitudes and political participation. Study results reveal a moderate but positive impact of online forum and social networking site use on online political discussion. Implications for political change in the social networking era, particularly in regimes that practice Internet censorship like China’s, are discussed. Ó 2013 Elsevier Ltd. All rights reserved.

1. Introduction Dating back to Poole’s (1983) prescient treatise on Technologies of Freedom, the diffusion of new media has inspired hopes that they can reinvigorate democracy (e.g., Dahlberg, 2001; Lagerkvist, 2010; Papacharissi, 2006). As recent headlines taken from 2011s ‘‘Arab Spring’’ and U.K. riots bear out, the rise of mobile and social media are revolutionizing Habermas (1962, 1989) conception of a public sphere, or forum for political discussion (Jeffres et al., 2008b; Jennings and Zeitner, 2003; Kellner, 1998; Rheingold, 1993), even emerging as the tool of choice for dissidents in the middle and near East (Lagerkvist, 2010). New media have played crucial roles in organizing activity and spreading information in those movements, especially ‘‘in light of the absence of an open media and a civil society’’ (Khondker, 2011, p. 675). Social media are also surfacing as key players in the Far East (Sun, 2010), where they stimulated unprecedented online campaigning in Singapore (Khoo, 2011), helping propel the opposition to gain six parliamentary seats in this otherwise tightly controlled one-party state (Mydans, 2011).1 Exploring the reciprocal influences between new media diffusion, economic growth and political participation, scholars contend that GDP is a good indicator of the possibility of transition (e.g., Gilley, 2004; Lum, 2004). When a post-totalitarian

⇑ Corresponding author. Address: C202, Macau University of Science and Technology, Avenida Wai Long, Taipa, Macau. E-mail addresses: [email protected] (Y. Mou), [email protected] (D. Atkin), [email protected] (H. Fu), [email protected] (C.A. Lin), [email protected] (T.Y. Lau). 1 The Singaporean example is particularly compelling because, like so many other countries in the region, the island nation remains dominated by a single party that controls the national media system (e.g., Lau et al., 2008). As in these other statist regimes, unrestricted political speech remains practically illegal and thus demands a greater level of risk and commitment than in the West (e.g., Anokwa et al., 2003; Lagerkvist, 2010). However, as one island observer noted, ‘‘social media have lowered the barriers of entry into political discourse everywhere. . .the effects have been electric’’ (Cenite cited in Mydans (2011, p. 2)). 0736-5853/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.tele.2013.04.002

360

Y. Mou et al. / Telematics and Informatics 30 (2013) 359–369

country’s per capita reaches around $3200, it has entered ‘‘the danger zone.’’ When it reaches above $4500, political pluralism becomes highly likely. The probability of democratization doubles every year as income per capita grows from $1700 to $8400. Although this does not imply causality, the correlation between economic prosperity and democratic transition remains robust (Lipset, 1959; Lum, 2000; Gilley, 2004). Economic development can predict political participation such that people in the more economically developed regions are more likely to participate in politics because they are more knowledgeable and politically active. This begs the question of how much further China’s robust economy can grow in the absence of corollary media liberalization, particularly as new media continue to diffuse throughout the country. Social media are thus playing an important role in spreading the news and challenging the Chinese government (Sun, 2010). Reporting on a July 2011 bullet train crash—which resulted in 39 dead and more than 200 injured—USA Today (MacLeon, 2011) observed that: An increasingly vocal public, aided by Twitter-like social media, has criticized the government for mishandling the aftermath, and challenged the limited explanation given to the date of the crash. . .Chinese bloggers and journalists have aggressively questioned the government’s handling of the crisis, from the death toll to the lack of reliable information (p. 1). In light of this transformative potential, the present study addresses the influence of the Internet and emerging social media (e.g., social networks) in China by studying their effects on the political activity of Chinese users. Study findings should help us gain a better understanding of the political influence of social media in similar political and media contexts. 2. Literature review With roughly double the number of Internet users as her American counterparts, China also stands poised to be transformed at the hands of emerging online and social media. Given that mass media are strictly regulated in China (e.g., Anokwa et al., 2003; Lagerkvist, 2010; Wu et al., 2011), the Internet presents an ideal ‘‘public sphere’’ for less filtered political discussion, particularly on politically sensitive topics. This raises the question of whether enhanced Internet connectivity promotes more political participation in China. Recent events suggest that Chinese leaders see the Internet as a force to reckon with, as President Obama’s 2010 visit to China was broadcast live online. At that time, Chinese leader Wen Jiabao used the Internet to answer a set of screened questions, although a portion of Obama’s sensitive speech was not broadcast. Chinese leaders are thus searching for a balance between Internet access and political participation, as the country’s burgeoning user base is a pacesetter among emerging economic powers. By July of 2011, China boasted 485,000,000 Internet users—double that of the U.S.—with penetration recently hitting a ‘critical mass’ exceeding a third (36.2%) of the population (CNNIC, 2011). The overall number of social networking users in 2010 was 235 million—or over 50% of China’s Internet base that year. Online forums or bulletin board system (BBS) retain their significance, accounting for 148 million users, or 32.4% of the total online population (CNNIC, 2011). The Internet thus provides an accessible, if unevenly filtered forum for Chinese users—including dissidents—to share information and even politically sensitive opinions (Sun, 2010). This stands in marked contrast to the state-controlled media which, true to the authoritarian model (Anokwa et al., 2003), remain heavily censored (Lagerkvist, 2010). Hence the Internet’s unique status as an open channel helps explain its relative popularity in China, particularly among younger users, and even belies the portrait of political apathy painted by several observers (e.g., Elegant, 2007; Schell and Shambaugh, 1999). In fact, youthful Internauts have assumed the vanguard in movements arising around such issues as the emerging transnational Chinese cultural sphere (Yang, 2003a), Chinese cyber-nationalism (Wu, 2007), the setting of public agendas that can affect government decision making (Zheng and Wu, 2005), and the formation of collective action and social norms (Arsene, 2008). In-depth social scientific investigations of the relationships between media use, political affairs and democratic aspirations are, in light of political sensitivities, rarer and more difficult to conduct in state-controlled media systems such as China’s (Mou et al., 2011). Here social networks sites (SNSs) such as Renren.com (a Facebook imitation) and Weibo (a Twitter imitation) are closely monitored. In light of these limitations, it is useful to consider parallel work addressing social media influences in other national contexts. 2.1. Understanding social media use The different impacts of various media forms on political participation have long been recognized by scholars (e.g., McLeod et al., 1999; Xenos and Moy, 2007). Distinct from old forms of media, social media can include both web-based and mobile technologies. These might encompass Internet forums, email, social networking sites, blogs, microblogs, vlogs, wikis, and voice-over IP, all of which emphasize extending traditional social networks online, while sharing information. As a consequence, the dynamism that social media bring to political participation may exceed that of traditional media, particularly as catalysts for social capital (e.g., Ellison et al., 2007). Scholars (e.g., Shah et al., 2001b) uncover strong ties between informational uses of media and social capital (Coleman, 1988; Putnam, 2000), which can be defined in several different ways: their own conception incorporated volunteer work, attending organizational meetings and working on projects. Early claims that the Internet could build social and political capital (Dahlberg, 2001) were not borne out by the data, however, at least not during its initial diffusion phase (Castells, 2001). In fact, some commentators cautioned that the proliferation of fragmented cable and online modalities may well


361

erode social capital (e.g., Putnam, 2000), with the ‘‘long-tail’’ of the new media leading to ever-narrowing taste-publics expressing a declining interest in public affairs content (Bucy et al., 2007; Delli-Carpini, 2000). Mindich (2004), for instance, notes that younger Americans are increasingly ‘‘tuned out’’ of public affairs outlets like newspapers and only 11% list news as a reason for going online; this malaise contributes to lower levels of political knowledge and participation (e.g., Diddi and LaRose, 2006). Others suggest that online participation is dampened by a belief that the Internet is a tool for government control (Cherry, 2005; Lessig, 1999). Technology determinists, on the other hand, see the Internet as a facilitator of online spheres and the rise of civic society generally (Yang, 2003b,c), ultimately aiding in the process of democratization in China (e.g., Chase and Mulvenon, 2002; Lagerkvist, 2010). Within 2 years of Facebook’s debut on American college campuses in 2005, social networking services quickly established themselves as key contributors to social capital which is conceived as the number of interpersonal ties (Boyd and Ellison, 2008; Ellison et al., 2007). Even after controlling for demographic factors, Internet use, and psychological well-being measures, level of Facebook use was a powerful predictor of social capital, explaining more than 40% of the variance (Ellison et al., 2007). Greater proficiency in online use is thus linked with social capital. Social capital, in turn, is linked to civic engagement and political participation in studies of general Internet use (e.g., Delli-Carpini, 2000; Jeffres et al., 2008a; Jennings and Zeitner, 2003). For instance, work in the knowledge gap tradition finds that newspaper readership predicts political knowledge holding and participation (e.g., Tichenor et al., 1980; Jeffres et al., 2008b). This framework could be grouped alongside other media S–R approaches into a social influence model, one predicting that emerging media enhance democracy because online news is a critical source of political information, which can easily be shared (e.g., Mou et al., 2011; Salwen et al., 2005). Online social media could thus facilitate public affairs consumption and political participation (e.g., Donsbach and Traugott, 2007; Miller and Krosnick, 2004; Salwen et al., 2005; Shah et al., 2001a). This invigoration of the public sphere thus extends to more mature democratic contexts like the U.S., where PEW data (Madden and Zickuhr, 2011) suggest that Facebook users are much more politically engaged than most people.2 Consequently, even in relatively mature democratic contexts like the U.S., military leaders express concern that social networking presents a threat to social stability (Pentagon to monitor social networking sites, 2011). This, combined with the fact that SNS’s have been linked to a range of pro-democracy movements across the globe, suggests that heavier users are likely to have a greater appetite for public affairs knowledge. In the Chinese context, we assume that information seekers would seek independent online sources, particularly if they believe that accounts from state media are censored or controlled by the government.3 Given the raft of recent work suggesting that SNS’s are building social and political capital—and past research showing that the use of media channels is interrelated (e.g., Jeffres et al., 2008a; McLeod et al., 1999)—we draw from the social influence model’s more optimistic assumptions about media orientations; namely, that heavier SNS users will make greater use of various media channels to satisfy their needs for public affairs information. More formally: H1a. Time spent online is positively related to online forum use. H1b. Internet proficiency is positively related to online forum use. H2. Time spent online is positively related to Internet proficiency. In addition to the government filtering mechanisms mentioned above, the Internet’s potential as a democratic tool in China is dampened by citizens’ general unwillingness to engage in political expression. Political nonparticipation is in effect a form of self-censorship (Hayes et al., 2006), which is prompted by concerns that engaging in such discussion could impose a social cost (e.g., alienation among friends, family and coworkers) (e.g., Chase and Mulvenon, 2002; Rosenberg, 1954). Such concerns help explain why even Westerners show little interest in political discussion (e.g., Eveland and Lively, 2009; Miller and Krosnick, 2004), particularly when it leaves them vulnerable to criticism (Hayes et al., 2006). Such concerns could prompt a ‘‘spiral of silence’’ in public settings (Noelle-Neumann, 1984), one that may mitigate any liberalizing presented by emerging online discussion forums. One’s political network thus affects one’s willingness to engage in political discussion (Jeffres et al., 2002; Kwak et al., 2005; Price et al., 2002; Scheufele et al., 2003), which Scheufele (2002) regards as the ‘‘soul of democracy.’’ Beck (1991) found that Americans were more likely to discuss politics with like-minded friends and family, although they encountered more dissonant messages from coworkers. Individuals who discuss politics more frequently in volunteer groups, in turn, engage in more political activities (Scheufele et al., 2003). Based on this dynamic, we hypothesize that: H3. Online forum and SNS use is positively related to frequency of political discussion. 2 PEW data (Madden and Zickuhr, 2011) also suggest that close to half of all U.S. adults now use social networking sites (SNS)—double the number of users in 2008—and the average user is getting older. Of particular interest: (1) Internet users were over twice as likely to attend a political meeting, 78% more likely to try influencing someone’s vote, and 53% more likely to have voted or intended to vote, and (2) compared with other internet users, and users of other SNS platforms, a Facebook user who uses the site multiple times per day was an additional two and half times more likely to attend a political rally or meeting, 57% more likely to persuade someone on their vote, and an additional 43% were more likely to have said they would vote. 3 It should be noted that there are several kinds of participation, the open forms of which are more difficult to facilitate in China owing to restrictions on speech. But privately, many Chinese citizens are frank and critical, especially among friends. SNS’s of course provide a rare medium through which such views can be shared.

362


2.2. An emerging public forum Habermas’s (1962, 1989) notion of a public sphere initially encompassed such forums as cafes, alongside traditional (e.g., print) media outlets. According to Habermas’s theory of rational communication, this public sphere provides citizens outside of the governing class with a vehicle for governance. Rheingold (1993) is among those early technology determinists arguing that the Internet represents a powerful extension of the public sphere. And true to the theory’s prognostications, the public sphere represents a ‘‘budding form of democracy’’ (Poor, 2006, ‘‘The public sphere’’, para. 1). SNSs thus provide an unprecedented avenue for vox populi to circumvent mass media filters and even facilitate political organizing in real time; that is, they represent an ideal online public sphere (Blumler and Gurevitch, 2001; Jeffres et al., 2008b; Rheingold, 1993). This is particularly significant in an authoritarian context such as China’s, where the Internet is characterized as the primary vehicle for citizens ‘‘to criticize government policies and to participate in politics’’ (Zheng and Wu, 2005, p. 525). In addition, given that users can be blind to each other’s identity in cyberspace, the abundance of online information enables users to interact with each other equally, regardless of ethnic background (Gray, 2005), political views (Dahlberg, 1998, 2001), or even gender (Hert, 1997; Lea and Spears, 1992), etc. That said, the Internet’s potential as an online forum is compromised by the general civic malaise accompanying the rise of fragmented new media environments (e.g., Althaus and Tewksbury, 2000; Bucy et al., 2007), and the commodification of cyberspace by corporate interests (e.g., Dahlberg, 1998; Papacharissi, 2002). In order to gain a better understanding of how social media might influence political participation in an authoritarian context like China, it is useful to provide fuller explication of political participation dynamics in that context. Owing, in part, to an imperial tradition that dates back thousands of years, Chinese netizens have not enjoyed the democratic traditions of their Western counterparts (e.g., Lagerkvist, 2010). Consistent with the authoritarian model of press functioning (e.g., Anokwa et al., 2003), traditional and emerging media in China have been strictly controlled by the Communist government. Of course, as a distributed network, the Internet presents a formidable challenge to government bureaucrats hoping to monitor the online discourse of nearly a half-billion netizens. The Chinese government has been aggressive, for instance, in cracking down on websites that they deem to be offensive or subversive (e.g., Zittrain, 2008).4 Technology firms hoping to conduct business in China, such as Cisco, have also been compelled to provide online addresses of users deemed to have offered offending postings; other companies refusing to comply with such requests (e.g., Google) have been denied access to the Chinese market. Still, in an apparent concession to online users, the Chinese government recently suspended plans to implement Green Dam filtering software, a software patch that could have allowed unprecedented government monitoring of online movements (e.g., Wu et al., 2011). This strong online government presence may have a chilling effect on online political expression in China. Wu et al. (2011) suggest that the PRC government’s ‘‘raised eyebrow’’ thus erodes user confidence in the privacy of their online communication, prompting many to engage in selfcensorship. In fact, a survey by China Youth Daily found that 15.1% of interviewees felt they were under digital surveillance (Survey reveals, 2002). This may help explain why, for instance, several longitudinal nation-wide surveys found that only 10% post their opinions frequently, while roughly 30% do so some times; most users do not post any opinions at all (Shen et al., 2009). This unwillingness to engage in political participatory behavior stems from such factors as a person’s level of political efficacy and trust in the political system which, have in turn been linked with use of information channels (McLeod et al., 1999). It is worth noting that the concept of political efficacy is multi-dimensional, including external and internal efficacy. Of particular interest here, internal efficacy encompasses ‘‘beliefs about one’s own competence to understand and participate effectively in politics’’ (Craig et al., 1990, p. 290). Previous studies have established the positive relationship between media use and political internal efficacy in democratic (e.g., Newhagen, 1994; Lee, 2006; Kenski and Stroud, 2006) and authoritarian countries (e.g., Chan et al., 2012). It is reasonable to consider that citizens gain political knowledge from mass media and become confident that they can understand the political process and public affairs, and eventually get involved. However, political external efficacy concerns the beliefs governing the responsiveness and effectiveness of government (Craig et al., 1990). The relationship between media use and external political efficacy remains unclear. The correlation between internal and external efficacy was found to be about .26 (Morrell, 2003). Yet, when analyzing national panel surveys from 1952 to 1980 in the United States, Abramson (1983) note steady declines in external efficacy, which corresponds with the decline of voter turnout and political trust, while the internal efficacy remains stable. A recent study on American college students finds a negative effect of Internet use on their external political efficacy (Lee, 2006). Gamson (1968) predicts that a combination of high internal efficacy and low external efficacy is optimal for political mobilization. Research suggests that political internal efficacy positively predicts political activities (e.g., McCluskey et al., 2004). In a traditional media context, for instance, the impact of local television news viewing on civic participation is mediated through political efficacy (Hoffman and Thomson, 2009). Based on the assumption that these dynamics operate online, we posit that:

4 Zittrain (2008, p. 1) notes, in particular, that: The Chinese government has used the Chinese version of Skype instant messaging software to monitor text conversations and block undesirable words and phrases. It and other authoritarian regimes routinely monitor all Internet traffic—which, except for e-commerce and banking transactions, is rarely encrypted against prying eyes. Over 500 Chinese cities have established Internet police units and a 30,000 member police unit to check Internet traffic and monitor websites, blogs, etc. for sensitive materials.


363

H4a. Political internal efficacy is positively related to online forum and SNS use. H4b. Political internal efficacy is positively related to frequency of political discussion. RQ1. What is the relationship between online forum and SNS use and political external efficacy? Defining social capital as trust, Whitely (2000) found that social capital is a strong predictor of variations in economic growth, regardless of differences in socio-political systems. As a result, the levels of economic development influence the levels of social capital in political discourse, which in turn could influence interaction between different social groups, social dynamics, and political participation. Pioneering work by Ellison et al. (2007), for instance, finds that social media use is a primary contributor to political activity and other forms of social capital. Lower levels of trust have been linked with greater levels of cynicism and lower internal efficacy (Kim, 2005). In addition, a person’s dispositional anxiety or discomfort may impact on his/her tendency to engage in political participation via media and other channels (Hayes et al., 2006; Jeffres et al., 2009). We assume that Chinese users are aware of online surveillance mechanisms that may be used by the government to monitor them and that this could reduce the necessary trust levels to engage in political participation via social media. Put another way, willingness to use SNSs will be a function of one’s trust in the communication paradigm that they are entering. Himbelboim et al. (2012) found that generalized trust (trust toward people with weak social ties) is a major predictor of using social media to interact with others about politics. Similarly, Cohen et al. (2001) found that personal-psychological variables such as efficacy and involvement are antecedents to political participation. Fu (2012) found that trust toward people with close ties is important to engaging in informal political participation, while trust toward people with weaker ties matters more to engaging in formal political participation. Another dimension of trust encompasses one’s faith in the credibility of information presented in the media, as distrust of conventional outlets could prompt citizens to seek out online sources of information and expression. In the online realm, this could be conceptualized as Internet trust. Trust should thus work in concert with internal self efficacy to encourage political activity (Mou et al., 2011; McLeod et al., 1999); more formally: H5a. Internet trust is positively related to online forum use. H5b. Internet trust is positively related to online political discussion. H6. Internet proficiency is positively related to Internet trust. More generally, to gain a panoramic picture of all above proposed relationships, a path model is structured to link online political discussion with antecedents (see Fig. 1). Identification of the process model of online political discussion could help achieve a fuller understanding of related phenomena. Toward that end, a path model is desirable as it could help identify the

Fig. 1. A proposed path model to predict online political discussion. Note: ⁄⁄⁄p < .001; ⁄⁄p < .01; ⁄p < .05; n.s., non-significant. Model fit: v2(6) = 9.94, p = .13; CFI = .93; RMSEA = .06.

364


process model by testing the relationships among variables under an integrated framework. In setting up the path model, Internet proficiency and time spent online were hypothesized to be related, both of which contribute to the frequency of online forum use. Online forum use, in turn, was hypothesized to influence the likelihood of online political discussion. In the meantime, political attitudes—specifically Internet trust and political internal efficacy—cast independent impacts on online political discussion.

3. Method 3.1. Sample profile Because any inquiries assessing political attitudes can be sensitive in a one-party state, a confidential paper-and-pencil survey was deemed the most secure research modality. The instrument was administered in two large universities located in two Eastern Chinese cities in June of 2010. The sample was both convenient and purposive, having been chosen because college students represent the first truly online generation in China (CNNIC, 2011).5 College students also represent a more accessible and relatively safer population for political communication researchers to survey. Given the politically sensitive nature of this study, respondents were assured of confidentiality and anonymity. Within the total sample of 181 respondents, 102 (56%) reported their gender as female, while 79 (44%) were male; some 177 (97.8%) indicated they were from the Han ethnic group, and four (2.2%) were minorities. Respondents age ranged from age 17 to 25, with an average of 20.31 (SD = 1.24). 3.2. Measures In the survey, participant demographic information was collected using commonly accepted measures, which encompassed gender, ethnicity, age and annual household income. This was assessed alongside media use and attitudinal measures, which are described in turn. Upon I.R.B. approval, the original English version of the questionnaire was translated into a Chinese version, and then back translated by two independent bilingual translators to check the semantic validity of the translated measurement. Items included the following. 3.2.1. Internet use and proficiency To capture the overall picture of Internet use, the respondents were asked to report how long they have been using the Internet. Overall Internet use time was measured by the estimation of weekly Internet use time (open ended, in hours and minutes). Self-rated Internet use proficiency was measured on a 7-point Likert-type scale, ranging from (1) not at all proficient to (7) very proficient (M = 4.51; SD = 1.36). 3.2.2. Online political participation Online political participation was measured in two steps. First, respondents were asked ‘‘Do you participate in online civic/political discussions?’’ They were required to select one of three answer options: (1) Yes; (2) No discussions, viewing/ reading others’ discussions only; and (3) No online forums/bulletin board use at all. If the answer was ‘‘yes’’, respondents were further asked to indicate ‘‘how frequently do you participate in online civic/political discussions’’, on a scale from (2) occasionally to (7) very frequently. Answers to these two questions were combined as online political participation frequency: for those answers of ‘‘No discussions, viewing/reading others’ discussions only’’ and ‘‘No online forums/bulletin board use at all,’’ the frequency was (1) never. Overall, online political participation frequency was measured on a 7-Likert scale from (1) never to (7) very frequently (M = 1.47, SD = 1.44). 3.2.3. Political efficacy Respondents’ internal political efficacy was measured by a revised version of Craig et al.’s (1990) scale. Six Likert scale items that best reflect the Chinese political system were used in the study (M = 4.20; SD = .83). The alpha reliability coefficient was .76. External political efficacy was gauged by eight items adopted from Craig et al.’s (1990) scale. The alpha reliability coefficient was .84 (M = 3.38, SD = 1.12). 3.2.4. Time spent online Time spent online was measured by the estimation of weekly time spent on the Internet (M = 21.46 h, SD = 18.06 h). To grasp the full picture of Internet use, respondents were asked to specify their time spent on online activities. Out of a variety of online activities, Instant Messaging (M = 6.47 h, SD = 7.99 h), information seeking (M = 3.47 h, SD = 3.53 h), and Forum use (M = 2.31 h, SD = 3.69 h) emerged as the three activities on which the respondents spend most time. 5 Chinese students represent the largest group of Internet users in China, accounting for 31.7% of the total online population, according to the 24th statistical report on Internet development in China (CNNIC, 2011); moreover, users aged from 10 to 29 account for 62.8% of total Internet users.


365

3.2.5. Internet trust Trust in the Internet was gauged by the respondents’ trust level toward the information on the Internet. Three items were measured on a 7-point Likert-type scale ranging from (1) strongly disagree to (7) strongly agree (M = 3.95; SD = 1.11). They included (1) ‘‘I believe the news from the Internet in general,’’ (2) ‘‘I trust the information from the Internet,’’ and (3) ‘‘I believe the news from the Internet is skewed in general.’’ The reliability coefficient Cronbach’s alpha was .81. 3.2.5.1. Online forum and SNS use. Online forum and SNS use was measured by the amount of online forums and SNSs that the respondents visited last week among a provided list of eight online forums and SNSs in China (see Appendix). The list includes the two major SNSs: Kaixin and Renren; five major online forums where college students most likely to engage in political discussion (e.g., Tianya) are also included. Considering the proximal impacts, each college’s own campus bulletin board system is included as well. Although online forum and SNS are two different types of social media services, the variable of online forum and SNS use was operationalized as an additive index by counting the number of forums users have visited during the past week. Two reasons justify the operationalization. Practically, the scarcity of popular SNSs in China—the only two major SNSs—makes it difficult to detach SNS use as an independent variable. And theoretically, users tend to use functionally similar communication technologies (e.g., LaRose and Atkin, 1992; Lin, 2006), such as online forum and SNS in this case. The concept of technology clusters (Rogers, 2003) reflects this idea. 3.3. Preliminary analysis Prior to data analysis, independent-sample t tests were conducted to rule out any potential differences attributable to the different data collection sites. No significant differences involving the major variables of interest were detected between the two samples. This result was also been confirmed by a further multiple regression analyses (with the dummy coded location as the independent variable and the key variables as dependent variables). Analyses conducted on the pooled samples are thus acceptable. Pearson correlations were first computed to identify any possible multicollinearity among the predictor variables. The correlation results and V.I.F. indicators suggest that multicollinearity was not a problem, as the highest correlation coefficient is .28. The hypotheses were then assessed by hierarchical multiple regression analyses. Later, as a supplemental step, the path model was finally analyzed to further check the hypotheses testing. 4. Results 4.1. Descriptive results Some 12.7% of respondents report participating in online political discussion. The majority (76.8%) of respondents lurks only, without discussion, and 10.5% refrained from any kind of lurking or discussion at all. Among the 23 people who discuss political issues online, 56.5% do it frequently (either every day or several times a week) and the rest (43.5%) do so on occasion. Each of the eight online forums and SNSs provided in the list was visited by at least one of the participants (M = 2.98, SD = 1.54). In one of those forums, 58 (33.2%) respondents showed a specific interest in politics. This percentage is higher than that for online political discussion, but lower than the percentage lurking in political discussions. The data thus indicate that, although some of users are particularly interested in politics, online the majority still refrain from actively participating themselves. 4.2. Hypotheses testing Hierarchical multiple regression analysis was conducted to test the hypotheses. Hypothesis 1a predicted that time spent online would be positively associated with online forum use, which we found to be the case here (b = .22, p < .01). Hypothesis 1a was therefore supported. Hypothesis 1b predicted that Internet proficiency would be positively related to online forum use. Results likewise indicated support for that prediction (b = .17; p < .05), providing support for Hypothesis 1 generally. Hypothesis 2 predicted that time spent online would be positively related to Internet proficiency. Results found the prediction supported: b = .36, p < .001. Hypothesis 3 predicted that online forum use would be positively related to online political discussion. Results indicated that online forum use was, in fact, related to online political discussion (b = .19, p = .01). Hypothesis 3 was therefore supported. Hypothesis 4a predicted that political internal efficacy would be positively associated with online forum use. A regression analysis found that their relationship is not significant (b = .03, p = .73). Hypothesis 4a was thus not supported. Hypothesis 4b predicted that political internal efficacy would be positively related to online political discussion. Results indicated that political internal efficacy was indeed significantly related to online political discussion (b = .22, p < .01). Hypothesis 4b was therefore supported. RQ1 queries the relationship between political external efficacy and online forum and SNS use. A correlation analysis found they are not related: r = .06, p = .41.

366


Hypotheses 5a and 5b predicted that Internet trust would be positively related to both online use and online political discussion. Results indicated that Internet trust is non-significantly related to online forum use: b = .03, p = .69. Therefore, Hypothesis 5a was not supported. A non-significant relationship was detected between Internet trust and online political discussion as well: b = .07, p = .34. Therefore, Hypothesis 5b was not supported. Hypothesis 6 predicted that Internet proficiency would be positively related to Internet trust. Results indicated otherwise (b = .09, p = .25), leaving the hypothesis without support. The hypothesized path model (Fig. 1) fits the data reasonably well: v2(6) = 9.94, p = .13; the comparative fit index (CFI) was .93; and the root mean square error of approximation (RMSEA) was .06 (.00–.13 for 95% C.I.). Nine percent of the variance in online discussion was explained by the proposed model. The path model supports a pattern of relationship in which online frequency and proficiency increase the likelihood of online forum use, and online forum use provides a catalyst for online political discussion. Additionally, political internal efficacy can positively influence online political discussion. Trust in the Internet, however, seems orthogonal to online political discussion. Taken together, the model echoes the relationships borne out in the regression analysis, providing a validity check of our results.

5. Discussion The present study set out to assess the influence of social media in political affairs amongst Chinese college students. On balance, study findings uncover positive relationships between use of online social networks, political internal efficacy, and level of online political discussion. Although a significant minority of respondents does engage in online political discussion, the norm of silent noninvolvement for Chinese students suggests that social media influences—in the online context, at least—remain more cautiously evolutionary than steadfastly revolutionary. While this suppressed political appetite might be partially explained by the relative satiation that Chinese students feel— as a relatively elite segment of society—a similar pattern of results was found in a series of large-scale surveys (World Internet Project cited in Shen et al., 2009). Here the typical suppressors of political participation found around the globe—apathy, lack of access to social media, and the like—are joined by the chilling (perceived) influence of government cyber-surveillance (e.g., Wu et al., 2011). The government’s presence online can erode social trust, which is logically a predictor of time spent in the online forum. Put another way, users are not likely to engage in extensive online discussion if they cannot trust their peers online (e.g., Jeffres et al., 2008b). Thus, the present study was keen on testing a spiral of trust dynamic here, one that is not inconsistent with the notion of opinion spirals (Noelle-Neumann, 1984); that is, we predicted that users are not likely to engage in political discussion until they feel requisite levels of political trust in the fact that their expression will be secure, accepted and not used against them at a later time. Of course, in a state controlled system like China’s, sanctions for speaking out may be more severe than in liberal democratic contexts where criticism of the government is more widely tolerated. There may thus be a number of unique factors that will need to be explored if we wish to better explain the Chinese populations’ online political forum use (see Sun, 2010). These may include a host of social and cultural explanations that are deep-rooted in the collectivist Chinese tradition, where expressing one’s opinions strongly and openly in public is neither encouraged nor admired. The saying that ‘‘silence is golden’’ still rings largely true as a Chinese social norm, where children are socialized by the convention that instructs them ‘‘to be seen, not heard’’ (see Lagerkvist, 2010). In addition, the Chinese media system is rather different from the Western system with which most communication scholars are familiar. Our pilot study, nonetheless, was able to generate a preliminary understanding of the Chinese population’s adoption of online political discussion forums and some of the reasons that could explain this phenomenon. The proportion of variance explained by our multivariate model is commensurate with past work, as the present study points to a new set of online involvement measures that can complement conventional measures of political interest and activity. The present study is limited in several respects. As mentioned earlier, the class-based paper-and-pencil survey approach is not as externally valid as some telephone and Internet modalities. But given the presence of several sensitive items, this approach was deemed the most likely guarantor of respondent privacy (and hence cooperation or trust). Such considerations help explain the relative paucity of political research on China to date, nearly a quarter century after the Tiananmen Square student demonstrations. Later work might profitably engage these issues with a larger, more diverse probability sample. In particular, this agenda could address the impact of government initiatives to regulate online speech. Future research might also consider how social media continue to shape China’s public sphere (e.g., the content and influence of political messages posted online).6 As Lum (2000) suggests, when capitalist development is successful in reaching economic growth from which a sizeable proportion of the population benefits, pressures to democratize will continue to grow. Based on Down’s (1957) conception of a rational actor, political actors’ willingness to social change will be determined by their evaluation of potential gains and losses. Easing of censorship on Internet communication represents perhaps the most pivotal change confronting Chinese government authorities. The relatively low levels of student involvement online uncovered here suggests that the Internet 6 This is likely to remain a moving target. The Southern Weekly newspaper recently reported, for instance, on a government proposal to establish a special restriction-free zone in Chonqing City—where Chinese netizens can access formerly inaccessible foreign websites for e-commerce purpose (as cited in Apple Daily (2011)).


367

is not yet a revolutionary force in China, hence both the ideological and logistical barriers for continuing such liberalization at the grassroots level remain stubbornly strong. Still, as Ai Wei Wei famously noted, ‘‘expressing oneself online is like a drug, I’m addicted to it’’ (Ai, 2011). He and others have noted that the source of China’s territorial unification has long depended upon the central authority’s ability to maintain geographical cohesion by preventing its people from congregating to form public opinions, politically. The Internet represents the greatest challenge to this long-established system of central government control over public opinion formation and expression. Nonetheless, the Chinese government is not anxious to take any chances, having recently ramped-up its censorship of media (which, in turn, triggered protest in the blogosphere; see Pierson, 2013). A key question going forward, then, is: can Chinese leaders control the message being sent in the Internet age, where so many Internet users use Weibo and other Internet channels to solicit information? One of two possible scenarios is likely to emerge; that is, the PRC government is likely to either (1) continue to closely filter and monitor online political discourse, or (2) embrace social media, perhaps as something of a ‘‘safety valve’’ for the expression of dissent as China’s growing middle class demands political freedoms commensurate with their economic attainment. The government’s ultimate course of action will depend on the degree to which they regard social media as a stable public forum, or a catalyst for insurrection. Parallel experience involving another authoritarian capitalist state in the region, Singapore, point to possible accommodations for social media. In a recent media report, one Island official observed that ‘‘the Government would have to be more adept at engaging Singaporeans online,’’ while another observer suggested that an active citizenry was ‘‘a good way to encourage Singaporeans to come together as one’’ (Ng, 2011, p. 1). In closing, our study supports a moderate but positive impact of online social networks on online civic participation, lending qualified credence to the positive role of social media in facilitating recent events such as the Arab Spring and other grassroots protests. Our study suggests that online social networks (measured by online forum use) and online political discussion are positively related. Frequent users of online social networks are more likely to engage in political discussion online. Online political discussion is also influenced by political efficacy such that more efficacious citizenry is related to more robust online civic activity. However, online political discussion is a multivariate phenomenon that could be attributed to a variety of factors. Our study has uncovered a small but valuable set of predictors. It is up to future research to ferret out a more comprehensive model that can be applied in other contexts. Appendix A. A list of online forums and SNSs (1) (2) (3) (4) (5) (6) (7) (8)

Campus BBS Tianya forum Qiangguo forum Xici MOP Wu You Zhi Xiang Renren Kaixin

References Abramson, P., 1983. Political Attitudes in America. W.H. Freeman and Company, San Francisco. Ai, W.W., 2011. In China, Lack of Truth is Suffocating. National Public Radio. Available from: