Networked Creativity on the Censored Web 2.0
Networked Creativity on the Censored Web 2.0
-Chinese Users’ Twitter-based Activities on the Issue of Internet Censorship-
Journal of Contemporary Eastern Asia. 2015. May, 14(1): 23-43
Copyright @ 2015, Asia Triple Helix Society
  • Published : May 31, 2015
Export by style
Cited by
About the Authors
Weiai Wayne Xu
Weiai Wayne Xu (MA, University of Wisconsin-Milwaukee) is a doctoral student at the University at Buffalo(SUNY-Buffalo), USA.
Miao Feng
Miao Feng (MA, University of Wisconsin-Milwaukee) is a PhD candidate in the Department of Communicationand a research assistant at the Health Media Collaboratory at the University of Illinois at Chicago.

In most of the world, the current trend in information technology is for open data movement thatpromotes transparency and equal access. An opposite trend is observed in China, which has theworld’s largest Internet population. The country has implemented sophisticated cyberinfrastructureand practices under the name of The Golden Shield Project (commonly referred toas the Great Firewall) to limit access to popular international web services and to filter trafficcontaining ‘undesirable’ political content. Increasingly, tech-savvy Chinese bypass this firewalland use Twitter to share knowledge on censorship circumvention and encryption to collectivelytroubleshoot firewall evasion methods, and even mobilize actions that border on activism. Usinga mixed mythological approach, the current study addresses such networked knowledge sharingamong citizens in a restricted web ecosystem. On the theoretical front, this study useswebometric approaches to understand change agents and positive deviant in the diffusion of censorshipcircumvention technology. On policy-level, the study provides insights for Internet regulatorsand digital rights groups to help best utilize communication networks of positive deviantsto counter Internet control.
Social media for social change: slacktivism oractivism?
Many have viewed the Internet as a positive force for social changes (Papacharissi, 2002). Recently,from the Arab Spring to the Occupy Wall Street movement, social media have played anunequivocal role in galvanizing public action ( Skinner, 2011 ). However, the enthusiasm for theidea that revolutions will be tweeted has met with criticism over slacktivism—the notion thatonline actions do not translate into offline impact ( Christensen , 2011;Morozov, 2009). Amidstthe contrasting arguments, Van Laer and Van Aelst (2009) note some convergence of online andoffline public spheres and a blurring of lines between online influence and offline outcomes. TheInternet affords various forms of virtual participation in activism. Whereas some of these virtual actions, such as email ‘bombing’, virtual sit-ins and petitioning, have low thresholds for participation and entail little or no political and legal risks (Van Laer &Van Aelst, 2009), some actions, such as culture jamming (e.g. virtual graffiti, parody, ironic and artistic alteration to challenge the authority) and hacking, require technological and cultural sophistication and risk-taking (Powell, 2011; Van Laer &Van Aelst, 2009). These virtual actions utilize the Internet’s creative function to expand traditional tactics used in social movements (Powell, 2011; Van Laer &Van Aelst, 2009). Such creative use of internet’s interactive features o challenge institutions of power is also referred to as technical activist actions (TAA) by Powell (2011). TAA is widely used in digital activism around technology-related policy issues such as Stop Online Piracy Act (SOPA) (Powell, 2011). Powell (2011) notes TAA users are likely to be tech-savvy and young. To revisit the debate regarding activism vs. slacktivism, virtual actions may not produce offline activism, but are nevertheless important in that they reflect Internet users’ creativity in adapting to and altering Internet technology and culture for collective goals.
However, the current studies of virtual collective actions have yet taken into consideration two factors. The first factor is that Web 2.0 platforms enable ‘word-of-mouth’ diffusion based on weak-tie connections and such infrastructure builds an online community for virtual actions to propagate and progress (Katona, Zubcsek, & Sarvary, 2011). It is less known how social media users utilize s networked nature within the community for virtual actions. The second factor concerns a restrictive internet environment underlying a Web 2.0 community. That is, what happens if the Internet infrastructure is severely restricted and disrupted by a controlling authority such as a government? How would Internet users adjust their actions to adapt to a restricted web environment? The two factors have led us to consider the context of Chinese Twitter users’ collective actions in the face of Internet censorship by the Chinese government.
Internet censorship in China: grassroots adaptation and activism
Internet censorship is a common practice of information control in authoritarian regimes. According to OpenNet Initiative (2011), over 40 countries implement various degrees of Internet censorship. In Iran, for instance, the government implements "halal Internet" that shields the country from Western cultural influence ( Shirazi, 2014 ). In Russia, the government maintains a blacklist of banned websites, and requires that all Russian user data be stored within the country for monitoring (Tselikov, 2014). The most well-known case is China’s ‘Great Firewall’ (GFW). The GFW blocks access to websites deemed politically sensitive or having the potential for facil-itating social movements (Feng & Guo, 2012). Blocked websites include popular social networking services such as Facebook and Twitter, and search engines like Google (Reuters, 2014). The GFW also conducts keyword filtering to police Internet traffic (Freedom House, 2013). Although the GFW has been widely criticized as a threat to Internet freedom, China’s moves to regulate the Internet are accompanied by beefed-up rhetoric about cybersecurity and national defense (South China Morning Post, 2014). Furthermore, amidst Edward Snowden’s revelations about U.S. eavesdropping on global communications and monitoring of terrorist organizations’ use of social media for recruitment and propaganda, there have been calls for “Internet sovereignty”— the idea that national governments should regulate and manage domestic cyberspace (The Dip-lomat, 2014).
Despite draconian Internet control, the online public sphere in China is teeming with opinions, including opinions critical of the government ( Yang, 2013 ). On several occasions, collective action and public sentiment have resulted in legislative changes and the downfall of high-ranking government officials (Yang, 2013). The empowering role of the Internet in the country’s societal development is made possible by Internet users’ creative virtual actions that allow them to bypass censorship (Yang, 2014). Their virtual actions fall under two domains: cultural tactics and technological tactics. As to cultural tactics, to get around keyword filtering, Internet users create code words and images to replace sensitive words used in politically sensitive discussions (Yang, 2014). For example, ‘Grass Mud Horse’, is an invented Chinese word with a shape similar to an alpaca, and its pronunciation is similar to a profane Chinese phrase. It was created to refer to China’s Internet censorship and has become an Internet meme in China. This word is used as a symbol of defiance and protest over the government’s Internet censorship ( Meng, 2011 ). As to technological tactics, users adopt and alter various technologies to circumvent censorship. Business applications such as Virtual Private Networks (VPNs), the Google App Engine and cloud computing services provided by Amazon and Microsoft have been used as censorship circumvention tools (Mou, Wu, &Atkin 2014; OpenITP, 2013). Using both the cultural and technological tactics, a sizable number of Chinese users are able to bypass the GFW. Reportedly, 35.5 million domestic Chinese users were active on Twitter in 2013 (Forbes, 2013), and domestic Chinese users of Facebook were estimated at 63.5 million in 2012 (Bloomberg, 2012). These numbers are intriguing considering China’s ban on both platforms. Even on domestic websites, some discussions of sensitive topics, including the issue of Internet censorship, slip past censors (Carter, 2013).
Social media empower users’ creative adaptation to Internet censorship. Web 2.0 platforms provide a connected online community for tech-savvy users to share insights and mobilize actions. On Twitter, for instance, hashtags such as #f**kgfw and #gfw are used to disseminate information about circumvention and encryption and issues related to Internet freedom (China Digital Times, 2009). Petition websites are used for launching activism targeted at the GFW. In 2013, an online petition was circulating on the U/S. White House’s petition website “We the People” (Global Times, 2013). The petition asked the U.S. government to deny visas to the developers of China’s Golden Shield Project.
Chinese Internet users’ creative approach to resist Internet censorship has increasingly received scholarly attention. Some researchers summarize the counter-censorship technologies adopted by Chinese users (see Burnett & Feamster, 2013; Leberknight, Chiang, Poor, &Wong, 2012), and more recent studies examine how Chinese users bypass the GFW for seeking information seeking, socializing and entertainment (Yang & Liu, 2014). Chinese Internet users’ adoption of anti-censorship tools is predicated on an array of factors, including attitude towards censorship, willingness to self-censor, political and media trust, and digital literacy (Mou, Wu, &Atkin, 2014). Some scholars offer thick description of Chinese users’creative cultural adaptation. Meng (2011) considers online spoofs as alternative political discourse but argues that online spoofs are not rational debates and do not lead to concrete political consequences. Nevertheless, online spoofs do convey political criticism and provide emotional bonding for users. Yang (2014) suggests that the use of code words and images to convey sensitive political information is a cultural response to censorship and “resistance” against institutional power. The current study adds to this body of work by providing an integrated framework to examine three elements in users’ creative collective actions targeting censorship. The three elements are users, relationships and messages , and all of these constitute the web ecosystem that supports collective actions.
The emphasis on these three elements is informed by webometrics, which is the study of quantitative aspects of web communication to explore the characteristics of online information diffusion and social interactions (Thelwall, 2009). Webometric approaches are commonly used in studying online activism (Introna &Gibbons, 2009) and public discussions (Woo-young & Park, 2012; Shapiro &Park, 2014). Along with these three elements, webometrics examines the characteristics of Internet users based on profile information self-disclosed on social media, in particular, information concerning political affiliation (Park et. al., 2011; Xu, Sang, Blasiola, & Park, 2014). Concerning relationship, webometrics uses social network analysis (SNA) to study the structure of social networks based on studying social ties from forged from friendships, conversations, and endorsements, affiliations and joint activities (Park and Leydesdorff, 2013). Network structure and positions in the network reveal paths of influence and coordination in collective actions and social movements (Huang & Sun, 2014; Xu, Sang, Blasiola, & Park, 2014) and in the formation of collective identity (Ackland & O’Neil, 2011). In information diffusion, for instance, the size of a network, and the density and clustering of relationships can influence the speed and scope of message propagation (Haythornthwaite, 1996). Occupying different positions in a network’s social hierarchy, users play different roles in fomenting collective actions. with users in central network positions tending to have considerable reach and influence over others (Xu, Sang, Blasiola, & Park, 2014). Additionally, users who connect two otherwise disconnected groups of people act as influence brokers across different social spheres (Gruzd, Wellman, & Takhteyev, 2011; Theocharis, 2013). Concerning the message, webometrics identifies important themes and linguistic features in public conversations. For example, based on social movement action repertoires (Van Laer and Van Aelst’s, 2010), Choi and Park (2013) find that online political groups use massive “retweeting” and “culture jamming” to enhance solidarity.
In the context of the current Internet censorship in China, these three elements may be applied to describe Chinese users who bypass the GFW to engage in collective actions against Internet censorship as positive deviants – innovators who challenge the establishment and norms, who evangelize for and embrace alternative approaches for resisting censorship (Pascale, Sternin, & Sternin, 2010). Although there is a lack of credible sources for estimating the number of censorship-evading users in China, the snowball sample used in a recent OpenITP (2013) report shows that 90% of surveyed respondents were male users – young and highly-educated with 80% having university degrees, and two thirds of them working or studying in information technology fields. The demographic profile of these respondents is consistent with an earlier survey by Jason Ng, an entrepreneur and activist in Beijing (2010). The censorship-evading users clearly deviate from the general demographic of Chinese Internet users reported by CNNIC (2014). Thus, the first research question of this study is on individual user characteristics:
  • RQ1: What are the demographic and behavioral characteristics of the users involved in Twitter-based discussion and mobilization on the issue of Internet censorship in China?
Bonded by a shared interest and concern, these users use the interactive features of social media to foster online communities. For example, on Twitter, they use the ‘retweet’ function to pass along information they deem valuable. Massive retweeting, as discussed earlier, is a form of collective action for creating solidarity (Choi & Park, 2014). It is also a form of crowd sourced gatekeeping used to raise valuable voices above the ‘noise’ inherent in user-generated content (Meraz & Papacharissi, 2013). Users also leverage Twitter’s reply feature for directed conversation. Twitter reply is interpersonal, thus likely leading to interpersonal relationships (Xu &Feng, 2014). More importantly, these communication features interconnect users through the flow of information, mutual association and joint activities (Nam, Lee & Park, 2013). Study of users’ social networks reveals salient participants and their influence. It has been noted that influentialusers can be agents of change who sway national political discourse and media reporting on censorship(Qiang, 2012). Overall, the patterns of interactions and social connections can indicatewhether an online ecosystem provides enough momentum for collective actions. Thus, the secondquestion of this study is about the characteristics of the social networks of users involved incollective actions against censorship.
  • RQ2; What are the characteristics of peer interactions among the users involved in Twitter-baseddiscussion and mobilization on the issue of Internet censorship?
Lastly, following previous studies that identify unique actions and tactics used in online activism(Choi & Park, 2014; Powell, 2011), the current study aims to identify a typology of virtual actionsin the context of Internet censorship.
  • RQ3: What virtual collective actions are used by the users involved in Twitter-based discussionsand mobilization on the issue of Internet censorship?
Selected hashtags
Hashtag name Hashtag definition Type
#GFW Discussions related to GFW (the Great Firewall of China) Issue-based
#F***GFW Discussions related to GFW (the Great Firewall of China) Issue-based
#翻墙 Discussions related to “climbing the wall,” a Chinese phrase referring tocensorship evasion Issue-based
#真理部 Discussions related to the Ministry of Truth, a term originated from GeorgeOrwell’s novel 1984, used to mock China’s censors Issue-based
#方滨兴 Discussion related to Fang Binxing, the father of GFW Issue-based
#goagent Discussions related to GoAgent, an open-source circumvention tool. technical
#shadowsocks Discussions related to the circumvention tool called shadowsocks technical
#vpn Discussions related to VPN technologies used for evading censorship technical
Data collection
Tweets that include hashtags related to China’s Internet censorship were download from theTwitter API. Eight hashtags were selected, among them five are general issue-based, and threehashtags named for specific censorship circumvention technologies (seeTable 1) . The selectionrepresents a broad array of topics and technologies related to the issue. Tweets sent during themonths of June, July and August of 2014 were included. Several censorship-related events occurredduring this time period, including the blocking of the Internet services company Google,the instant messaging (IM) tool Line, and the online storage service OneDrive, and the implementationof a stricter law targeting domestic IM users (Storify, 2014). The raw dataset includes160,605 tweets sent between June 11 and August 17, 2014. Several iterations of datacleaning3wereperformed to delete tweets not related to the issue. Given our focus on ChineseTwitter users, Non-Chinese tweets were excluded, resulting in a sample of 60,232 tweets.
PPT Slide
Lager Image
A breakdown of tweet volume by hashtag
With the current study addressing the networked aspect of public actions on censorship, we onlyinclude tweets that seek to form social ties between users. Therefore, tweets not directed at specificusers are excluded. Directed tweets embody three Twitter features—retweet, mention, andreply (Boyd, Golder, &Lotan, 2010). All of these are indicated by the symbol @, but each hasdifferent underlying communication motivations and goals (Xu & Feng, 2014). For this study,we distinguish between mention/reply and retweet because retweet is an act of forwarding another’s message, meant to agree and acknowledge (Xu, Sang, Blasiola, & Park, 2014), whereasmention/reply is meant to engage targeted users in conversations (Xu & Feng, 2014). There are866 directed tweets in the final sample. 94% (n=816) are retweets, and 6% (n=50) are mention/reply. The directed tweets were sent by 577 unique users, targeting 162 unique users. MostTwitter activities were driven by issue-based hashtags (see Figure 1 ).
- Profile analysis
For RQ1, user characteristics were inferred from Twitter user profiles. Three types of profileswere included in profile coding: Twitter bio, location, and profile images. Profile information isorganized into three dimensions: the degree of self-disclosure, the nature of self-disclosure , and the tactics of self-disclosure . First, the degree of self-disclosure is gauged by the variety of revealedpersonal information. An inductive analysis shows that users typically disclose informationin the following areas: name (N), occupation (O), location (L), political opinion (PO),personality and lifestyle (P) and contact information (C). The nature of self-disclosure noteswhether users reveal political ideology and viewpoints in profiles, and whether they express interestor involvement in Internet technology.
Profile coding scheme
Categories Definition/Note
The degree of selfdisclosure Name (N) not including pseudonym
Location (L) Specifying different types of location information:(a) China-base users’ locationdetailed to province or city-level; (b) generallocation information without identifiableprovince or city name, or latitude information;(c) locations outside of MainlandChina, including Chinese territories of HongKong, and Taiwan where Chinese censorshiplaw does not apply
Occupation (O)
Political opinion/identity (PO)
Personality and lifestyle (P) A user
Contact information Email address, phone number, IM accounts,profiles on other social networking sites
The nature of selfdisclosure Expressed political identity, interest andinvolvement
Expressed interest and involvement in technology
Tactics of selfdisclosure Creative expressions of identities and opinions The use of sarcasm, parody, spoof, codelanguage/image
The rationale for focusing on the political and technology-related disclosure is because censorshipis a politicized technology issue. The tactics of self-disclosure is an open coding categoryused to identify creative ways of expressing personal identity. Table 2 presents a detailed codingof categories with definitions. Two authors coded all available profiles along the three disclosuredimensions and the agreement percentage is 91.6% for intercoder reliability.
Content coding scheme
Dimension Category Definition
Technological Reporting Describing technological problems and processes oftroubleshooting the problems
Sentiment Expression of frustration and anger
Commentary Expressing opinions about a technology issue
Solution Providing tips and how-to guide to solve a technologicalproblem involving censorship evasion
Mobilization Engaging others to join an event or collective actionson a technologic problem
Political Sharing news Forwarding news articles
Commentary Expressing opinions about a political issue
Mobilization Engaging others to join a political cause
Creative expression The use of code words to for discussions of sensitiveevents and topics
Social network analysis
For RQ2, three types of network are considered: a global Twitter network including both the retweetand mention/reply Twitter features, a retweet network based on flows of acknowledgement,and a reply network based on conversation flows. In the networks, participating users arethe nodes, connected with other users by ties emanating from retweet and mention/reply. If UserA retweets/mentions/replies to User B, we set up a tie from A to B. When a sender targets multipleusers in a tweet, it is treated as multiple ties originating from the sender. Social networkanalysis (SNA) is used to examine the structure and users’ positions in the networks (Haythornthwaite,1996). Using the SNA programs NodeXL and UCINET, nodal-level information isextracted to identify salient and influential users. Specifically, in-degree centrality is used to sortout users frequently retweeted by other users or targeted in conversations for feedback. High indegreecentrality indicates opinion leadership and importance recognized by other users (Xu,Chiu, Chen, & Mukherjee, 2014). Out-degree centrality identifies users who are active in forwarding others’ messages and in starting or replying conversations. On the network level, cen-tralization is used to show the distribution of power/engagement in conversations and infor-mation diffusion (Freeman, 1979). A highly centralized network means that a small group of people dominate the discussion and diffusion. Density measure is used to examine whether users are well connected with one another to form a cohesive community (Borgatti, Mehra, Brass, & Labianc, 2009). Additionally, K-core was used to identify a cluster of closely connected users (Seidman, 1983).
Content analysis
To answer RQ3, content analysis is conducted to draw the typology of actions used in the Twitter discussions. A pilot coding reveals various themes, leading to the development of a final coding scheme that centers on three areas of actions: technological actions, political actions, and creative actions. Under each, specific coding categories are developed to reflect actions with various levels of participation threshold and potential consequence. Table 3 presents the content coding scheme.
PPT Slide
Lager Image
Self-disclosure on Twitter profiles
To address RQ1 which deals with user characteristics (see Table 3 ), among 576 senders in the discussions, 68% (n=389) disclosed at least one element of personal information. On average, a user discloses one item (mean = 1.13; s.d.= 1.02). Among the disclosers, about 70% disclosed location information, of which 230 users disclosed location information as detailed as a Chinese province or city or even a geographic coordinate. The location information reveals that 40 of the users are identified as living outside of Mainland China (that is People’s Republic of China, excluding Taiwan, Hong Kong and Macau--three regions with a political system different from therest of China). The second most frequently disclosed item, political identity and opinion, is 32.13%. More private personal information such as name and contact information are revealed less often. Only five users reveal name and 29 users disclose contact information. In addition, creative expressions of identity are found in 60 profiles.
PPT Slide
Lager Image
Visualization of the global network of retweet and mention/reply
PPT Slide
Lager Image
To address RQ2, the network based on Twitter retweet and mention/reply consists of 677 users, including 576 unique senders and 164 unique receivers. These users created 1,263 ties. Figure 3is a visualization of the Twitter network. Table 4a and 4b present a list of top users by in-degreeand out-degree centrality. This global network is decentralized, judged by the 0.5 threshold forcentralized/decentralized network (in-degree & out-degree centralization = 0.17) and sparse(density = 0.02).
Identity attributes of engaged users
Users identified by 5K-core Identity attributes
H***a Political dissident
L***i Auser who has expressed nostalgia towards Republic of China
S***n Writer
x***s Average user
M***L A Chinese living in the U.S. with interest in politics and civil rights
S***g A journalist working for a Western media
X***y Political dissident
d***e The Chinese page of a German news media
Top users by in-degree centrality in the global network of retweet/mention/reply
User screenname In-degree centrality Note on user profile
H***a 102 Political dissident
C***w 100 A research project about Chinese Internet censorship
A***0 67 A user with expressed interest in information security
O***x 53 A user with expressed interest in Internet censorship issue
l***5 50 Not disclosed
X***t 40 A user with expressed interest in Internet censorship issue
D***g 37 A user with expressed interest in Internet censorship issue
A***4 31 A user with expressed interest in technology
S***n 20 A user with expressed interest in Internet censorship issue
S***g 18 A reporter with expressed interest in politics
X***t 18 Auser with expressed interest in Internet censorship issue
P***u 18 A user with expressed interest in technology
Yet, the network graph shows a set of core users being common recipients in Twitter retweetingand replying. This set of users is identified by K-core. Figure 4 of 5K-core shows a cluster of usersconnected to at least five other users (see Figure 4 ). Table 4 shows that this small set of engagedusers are involved in politics and technology.
Top users by out-degree centrality in the global network of retweet/mention/reply
Screen name Out-degree centrality Note on user profile
H***a 21 Political dissident
D***g 17 A user with expressed interest in Internet censorship issue
L***i 15 A user who has expressed nostalgia towards Republic of China
O***x 10 A user with expressed interest in Internet censorship issue
F**6 10 Not disclosed
X***n 9 A user who expressed determination for justice in the society
G***9 9 Political dissent
M***L 8 A user expressed nationalist sentiment
C***a 7 A user with expressed interest in Internet censorship issue
D***n 7 Political dissent
M***i 7 A user with expressed interest in Internet censorship issue
Top users by in-degree centrality in the retweet network
Screen name In-degree centrality Note on user profile
H***a 102 Political dissident
C***w 100 A research project about Chinese Internet censorship
A***0 67 A user with expressed interest in information security
O***x 51 A user with expressed interest in Internet censorship issue
L***5 47 Not disclosed
X***t 39 A user with expressed interest in Internet censorship issue
D***g 36 A user with expressed interest in Internet censorship issue
A***4 30 A user with expressed interest in technology
S***n 19 A user with expressed interest in Internet censorship issue
S***g 18 A reporter with expressed interest in politics
Figure 5 visualizes the retweet network, consisting of 643 users with 1,194 ties. Tables 5a and 5bpresent top users by in-degree and out-degree centrality. In-degree centralization for the retweetnetwork is .16 and out-degree centralization is .02. Density of the network is .002.
Top users by out-degree centrality in the retweet network
Screen name 17 Note on user profile
H***a 15 Political dissident
L***i 10 Auser who has expressed nostalgia towards Republic of China
F***6 9 Not disclosed
O***x 9 A user with expressed interest in Internet censorship issue
G***9 8 Political activist
D***g 8 A user with expressed interest in Internet censorship issue
X***n 8 A user who expressed determination for justice in the society
M***l 8 A user expressed nationalist sentiment
M***i 8 A user with expressed interest in Internet censorship issue
L***n 7 A user expressed nationalist sentiment
Top users by in-degree centrality in the mention/reply network
Screen name In-degree centrality Note on user profile
S***x 4 Programmer
L***5 3 Not disclosed
O***x 2 A user with expressed interest in Internet censorship issue
A***1 2 Not disclosed
C***q 2 A user who expressed political ideologies
C***4 2 A user with expressed interest in Internet censorship issue
S***a 2 A user with expressed interest in technology
D***g 2 A user with expressed interest in Internet censorship issue
PPT Slide
Lager Image
Visualization of retweet network
Top users by out-degree centrality in the mention/reply network
Screen name In-degree centrality Note on user profile
D***g 9 A user with expressed interest in Internet censorship issue
H***a 4 Political dissident
F***n 4 A user with expressed interest in politics
C***a 3 A user with expressed interest in Internet censorship issue
N***y 3 Not disclosed
O***x 2 A user with expressed interest in Internet censorship issue
X***s 2 Not disclosed
D***n 2 Political dissident
J***h 2 Not disclosed
F***0 2 Not disclosed
Figure 6 visualizes the mention/reply network. The network includes 81 users and 71 ties. Bothin-degree and out-degree centralization are .10 and the network has a density of .02. Table 6aand 6b presents top users by in-degree and out-degree in the network.
Distribution of actions and themes
Dimension Category No. of Tweets No. of Retweets
Technological Reporting 172 3702
Sentiment 54 146
Commentary 87 188
Solution 273 7019
Mobilization 20 212
Political Sharing news 130 684
Commentary 121 741
Mobilization ? 69
To address RQ3, the majority of tweets deal with technical aspects of censorship (70%), despite the fact that most Twitter activities are driven by issue-based hashtags rather than technologybased hashtags. Among the technical contents, providing solutions is the most salient theme (n=273), followed by describing or reporting problems (n=172). For both technical and political content, only a small proportion deals with mobilization. Table 7 presents the distribution of actions and themes.
PPT Slide
Lager Image
Visualization of mention/reply network
Discussion and Conclusion
Internet censorship and resistance to Internet censorship have increasingly provoked debate both inside and outside of authoritarian regimes. Over the past few years, it has been demonstrated that resistance can be enhanced by collective actions via social media platforms. China presents a compelling case for study on this issue due to the strong role the Chinese government plays in Internet censorship and because of the creativity exhibited by those who oppose this censorship. In exploring Chinese Twitter users’ collective response to Internet censorship, this study shows how their collective response is supported by the social media networks and a variety of messaging strategies. We have documented that Chinese Internet users agilely adapt to the restricted web environment, not only by developing technological solutions to bypass blockage of certain services, but also by experimenting with creative means of self-expression and opinion sharing. Our analysis reveals the collective nature of their response by using an integrated webometric approach that examines three elements in a web ecosystem: users, relationship and usergenerated content. Employing a mixed methodological approach allowed this study to focus onthe quantitative aspects of Twitter-based social networks and analyze the themes found in usergenerated content. This approach further allowed in-depth qualitative interpretation of the symbols used in censorship-related discussion and to mobilize activism.
Regarding the characteristics of users, our findings reveal that collective response to Internet censorship in China has been prompted and promoted among a group of engaged and experienced individuals who identify themselves as enthusiasts of technology and politics. Such user charicteristics have led to the question of whether the impact of such collective response could reach the general public, who lack the knowledge necessary to evade the ‘Great Firewall’. The engaged, tech savvy few represent positive deviants who could become change agents bringing counter-censorship technologies to achieve the ideal of an open Internet to the general public. However, the demographic composition of this positive deviant network alludes to a potential digital gap among Chinese Internet users, between the haves of unfettered Internet access and the have-nots of an Internet in government fetters.
Regarding the structure of user relationships, the findings from network analysis shows a relatively decentralized and sparse network. There are few dominant parties, and the involvement and enthusiasm for collective actions are most visible in a small set of users, consisting of political dissidents and technical activists. This group of users are active in starting and following up on conversations. Their messages are also widely retweeted, indicating their opinion leader status.
Regarding user-generated content, our analysis suggests that the majority of tweets focus on technical information sharing such as discussions on services, tools, problems, and countercensorship solutions, among which the majority center on providing tips on bypassing censorship. Technical information is also frequently retweeted by the entire Twitterverse. In relation to Van Laer and Van Aelst’s (2010) digital action repertoire along the low/high threshold dimen-sion, it can be argued that providing technological solutions generally requires a higher degree of technological sophistication than reporting and commenting on technological problems. Collective technical information sharing among the users is done in a conversational and timely fashion, featuring continuous Q&A and troubleshooting. This kind of information sharing likely reinforces technological solutions to censorship and further diffuses counter-censorship technologies. In the political domain, it is found that most tweets focus on sharing political news and opinions. This shows that users not only view Internet censorship as a technological issue, but an issue intertwined with Chinese politics. This potentially paves the way for the collective action on censorship to converge with political activism on other political issues.
While it may be hoped that the connectivity of social media enables users to engage in activism to address issues head-on, we find that tweets that touch on the theme of mobilization are minimal, particularly in the political domain. A very small portion of tweets, less than 1% in the current study, are devoted to discussions of mobilization. Engaging others in collectively addressing an issue requires certain level of influence, and likely a greater degree of risk, especially in China. This finding, interestingly, is similar to a previous study (King et al., 2014) showing that Chinese censors allows criticism of the government to some extent, but adopt more strict control over messages that call for collective actions in China. Among the content with mobilization themes, some offer direct paths toward action. For example, a tweet asks users to submit a form requesting government public information from Chinese regulators in probing the censorship of Google.
Despite the paucity of content on mobilization, the content revealed by users on their Twitter profiles indicates their creativity in resistance. There are many uses of sarcasm, parody,spoofs, and code words in their profiles. Unexpectedly, location information on Twitter is politicized by the users. Some identify their location as the “old capital of the Republic of China,” as a way to show nostalgia for the pre-Communist era, and a way to voice their discontent toward the current government. Some use “West Korea” (west of North Korea) to mock China’s resemblance to North Korea in terms of speech control, and some use “enemy-occupied area” to express their hostility towards the current government. In Twitter bios, similarly, some use the line, “Across the Great Firewall we can reach every corner in the world.” A line originally used by China’s state media at the launch of China’s first email service in 1987 (Xinhua, 2012). It can be argued that these creative expressions of opinion do not necessarily represent a tactic to evade censorship since the users’ speech on Twitter does not fall under the jurisdiction of Chinese cen-sors, but are a form of domestic Internet culture. These terms are developed and adopted by Chinese users, and used in environments with and without keyword censorship. In contrast to Chinese domestic web services which are heavily policed by keyword filtering, Twitter provides a platform for unfettered speech. There are bold statements expressing political ideals: in some profiles, users express willingness to sacrifice for liberal political ideologies such as ’freedom’, ’democracy’, ’human rights’, and ’justice.’ It can be argued that the creative and bold profile expressions indicating political ideals may represent a new way to respond to China’s Internet control. The politicization of the issue may foster the strong sense of citizenship necessary for social movement in the future. For the current research, the depth of self-disclosure varies among the users participating in the discussions on the issues examined in this study. In spite of the creative and discursive strategies they use to describe themselves, a sizable number of users have disclosed very detailed identity information, which may motivate others and foster a shared identity.
To relate the findings to the general debate on whether social media forges new forms of civil participation, we argue that social media provide infrastructure for networked activism, yet this potential is limited by the political power exercised over a country’s Internet openness. Despite this limitation, users develop new ways (e.g. using location information to express dissatisfaction) to voice dissent and call for collective action. To study collective action in an authoritarian society, one needs to consider the higher threshold required for and risks involved in participation in online discussion, debate, and protest. Content simply passing along political information or providing technological solution should be given equal weight with content that directly calls for action.
Ackland R. , O’Neil M. 2011 Online collective identity: The case of the environmental movement Social Networks 33 (3) 177 - 190
Bloomberg 2012 Accessed on 01/13/2015 at:
boyd D. , Golder S. , Lotan G. IEEE 2010 Tweet, tweet, retweet: Conversational aspects of retweeting on twitter. In System Sciences (HICSS) 2010 43rd Hawaii International Conference on 1 - 10
Borgatti S. P. , Mehra A. , Brass D. J. , Labianca G 2009 Network analysis in the social sciences science 323 (5916) 892 - 895
Burnett S. , Feamster N 2013 Making sense of Internet censorship: a new frontier for In-ternet measurement ACM SIGCOMM Computer Communication Review 43 (3) 84 - 89
Carter L. China Digital Times Inc 2013 The Grass-Mud Horse Lexicon:Classic Netizen Language
China Digital Times 2009 Accessed on 01/13/2015 at:
Choi S. , Park H. W. 2013 An exploratory approach to a Twitter-based community centered on a political goal in South Korea: Who organized it, what they shared, and how they act-ed New Media & Society, 1461444813487956
Christensen H. S. 2011 Political activities on the Internet: Slacktivism or political participa-tion by other means? First Monday 16 (2)
CNNIC 2014 Accessed on 01/13/2015 at:
The Diplomat 2014 Accessed on 01/13/2015 at
Freedom House 2013 Accessed on 01/13/2015 at
Freeman L. C. 1979 Centrality in social networks conceptual clarification Social networks 1 (3) 215 - 239
Forbes 2013 Accessed on 01/13/2015 at:
Global Times 2013
Gruzd A. , Wellman B. , Takhteyev Y. 2011 Imagining Twitter as an imagined community American Behavioral Scientist 55 (10) 1294 - 1318
Haythornthwaite C. 1996 Social network analysis: An approach and technique for the study of information exchange Library & information science research 18 (4) 323 - 342
Huang R. , Sun X. 2014 Weibo network, information diffusion and implications for collec-tive action in China Information, Communication & Society 17 (1) 86 - 104
Introna L. D. , Gibbons A. 2009 Networks and Resistance: Investigating online advocacy networks as a modality for resisting state surveillance Surveillance & Society 6 (3) 233 - 258
Katona Z. , Zubcsek P. P. , Sarvary M. 2011 Network effects and personal influences: The diffusion of an online social network Journal of Marketing Research 48 (3) 425 - 443
Kenengba 2010 Accessed on 01/13/2015 at:
King G. , Pan J. , Roberts M. E. 2013 How censorship in China allows government criti-cism but silences collective expression American Political Science Review 107 (02) 326 - 343
Leberknight C. S. , Chiang M. , Poor H. V. , Wong F. 2012 A taxonomy of Internet censor-ship and anti-censorship
Meng B. 2011 From steamed bun to grass mud horse: E Gao as alternative political discourse on the Chinese Internet Global Media and Communication 7 (1) 33 - 51
Meraz S. , Papacharissi Z. Networked gatekeeping and networked framing on# Egypt The international journal of press/politics, 1940161212474472
Morozov E. 2009 Iran: Downside to the" twitter revolution" Dissent 56 (4) 10 - 14
Mou Y. , Wu K. , Atkin D. 2014 Understanding the use of circumvention tools to bypass online censorship New Media & Society, 1461444814548994
Nam Y. , Lee Y. O. , Park H. W. 2013 Can web ecology provide a clearer understanding of people’s information behavior during election campaigns? Social Science Information 52 (1) 91 - 109
OpenITP 2013 Accessed on 01/13/2015 at:
OpenNet Initiative 2011 Accessed on 01/13/2015 at:
Papacharissi Z. 2002 The virtual sphere: The Internet as a public sphere New media & socie-ty 4 (1) 9 - 27
Park H. W. , Leydesdorff L. 2013 Decomposing social and semantic networks in emerging “big data” research Journal of Informetrics 7 (3) 756 - 765
Park S. J. , Lim Y. S. , Sams S. , Nam S. M. , Park H. W. 2011 Networked politics on Cy-world: The text and sentiment of Korean political profiles Social Science Computer Re-view 29 (3) 288 - 299
Pascale R. T. , Sternin J. , Sternin M. Harvard Business Press 2010 The power of positive deviance: How unlikely innovators solve the world's toughest problems (Vol. 1)
Powell A. 2012 Assessing the Influence of Online Activism on Internet Policy-Making: The Case of SOPA/PIPA and ACTA    DOI :
Reuters 2014 Accessed on 01/13/2015 at:
Seidman S. B. 1983 Network structure and minimum degree Social networks 5 (3) 269 - 287
Shapiro M. A. , Park H. W. 2014 More than entertainment: YouTube and public responses to the science of global warming and climate change Social Science Information, 0539018414554730
Shirazi F> 2014 Interrogating Iran’s restricted public cloud: An Actor Network Theory per-spective Telematics and Informatics 31 (2) 228 - 236
Skinner J. 2011 Social media and revolution: The arab spring and the occupy movement as seen through three information studies paradigms
Storify 2014 Accessed on 01/13/2015 at:
South China Morning Post 2014 Accessed on 01/13/2015 at
Thelwall M. 2009 Introduction to webometrics: Quantitative web research for the social sci-ences Synthesis lectures on information concepts, retrieval, and services 1 (1) 1 - 116
Theocharis Y. 2013 The wealth of (Occupation) networks? Communication patterns and in-formation distribution in a Twitter protest network Journal of Information Technology & Politics 10 (1) 35 - 56
Tselikov A. 2014 The Tightening Web of Russian Internet Regulation Berkman Center Re-search Publication, (2014-15)
Van Laer J. , Van Aelst P. 2009 Cyber-protest and civil society: the Internet and action rep-ertoires in social movements Handbook on Internet crime 230 - 254
Van Laer J. , Van Aelst P. 2010 Internet and social movement action repertoires: Opportu-nities and limitations Information, Communication & Society 13 (8) 1146 - 1171
Woo‐young C. , Park H. W. 2012 The network structure of the Korean blogosphere Jour-nal of Computer‐Mediated Communication 17 (2) 216 - 230
Xinhua News 2012 Accessed on 01/13/2015 at:
Xu W. W. , Chiu I. H. , Chen Y. , Mukherjee T. 2014 Twitter hashtags for health: applying network and content analyses to understand the health knowledge sharing in a Twitter-based community of practice Quality & Quantity 1 (20)
Xu W, W, , Feng M. 2014 Talking to the Broadcasters on Twitter: Networked Gatekeeping in Twitter Conversations with Journalists Journal of Broadcasting & Electronic Media 48 (3) 420 - 437
Xu W. W. , Sang Y. , Blasiola S. , Park H. W. 2014 Predicting Opinion Leaders in Twitter Activism Networks The Case of the Wisconsin Recall Election American Behavioral Scientist, 0002764214527091
Yang G. 2013 The power of the Internet in China: Citizen activism online Columbia Univer-sity Press
Yang F. 2014 Rethinking China’s Internet censorship: The practice of recoding and the poli-tics of visibility New Media & Society, 1461444814555951
Yang Q. , Liu Y. 2014 What’s on the other side of the great firewall? Chinese Web users’ motivations for bypassing the Internet censorship.Computers in Human Behavior 37 249 - 257