Developing network indicators for ideological landscapes from the political blogosphere in South Korea


  • Han Woo PARK,

    1. YeungNam University
      Korea (South)
    Search for more papers by this author
    • Park, H. W., & Thelwall, M. (2008/2009 forthcoming). Developing network indicators for ideological landscapes from the political blogosphere in South Korea. Journal of Computer-Mediated Communication.

    • Han Woo PARK is an Assistant Professor in the Department of Communication & Information, YeungNam University, South Korea, where he teach courses and conduct research on the use of new communication technologies in extending social and political networks, and e-science/e-research topics. Over past few years, he has contributed important works in the area of Link Analysis (also called, Webometrics) from the perspective of Social Network Analysis. Currently, he is a coeditor of the Journal of Contemporary Eastern Asia.

      Email: /
      Website: http:/
      Address: Dept of Communication & Information, YeungNam University
      214-1, Dae-dong, Gyeongsan-si,Gyeongsangbuk-do,
      South Korea, Zip Code 712-749

  • Mike Thelwall

    1. Statistical Cybermetrics Research Group
      School of Computing and IT
      University of Wolverhampton
      Wolverhampton, UK
    Search for more papers by this author
    • Park, H. W., & Thelwall, M. (2008/2009 forthcoming). Developing network indicators for ideological landscapes from the political blogosphere in South Korea. Journal of Computer-Mediated Communication.

    • Mike Thelwall is Professor of Information Science and leader of the Statistical Cybermetrics Research Group at the University of Wolverhampton, UK. He is also visiting fellow of the Amsterdam Virtual Knowledge Studio, a Docent at Åbo Akademi University Department of Information Studies, and a research associate at the Oxford Internet Institute. He has developed tools for downloading and analyzing web sites, blogs and social networking sites, including the research web crawler SocSciBot and the LexiURL link analysis software. He is the author of the book Link Analysis: An Information Science Approach, and sits on nine editorial boards.

      Address: School of Computing and IT, University of Wolverhampton, Wulfruna Street
      Wolverhampton, WV1 1LY, UK


This paper investigates hyperlink patterns in the South Korean political blogosphere. Using sampling from the blog sidebar hyperlinks of elected politicians (National Assemblymen), the top 79 elite citizen blogs were selected. Two data sets were manually compiled during January, 2007: (a) links between politicians and citizens, and (b) links amongst citizens. A variety of social network analytical methods were then applied. The results show that more top blogs have reciprocal links with politicians than have unidirectional links. The structure of hyperlink interconnectivity suggests that the ruling Uri party affiliated blogs are key in the blog network. For example, the blogs tied with the Uri party have a higher centrality and are more densely connected. Network diagrams also suggest that the top blogs are polarized by party. However, some blogs are located at the center of the Uri and GNP clusters and are connected to both camps. In other words, there are a number of citizen blogs that link to both the Uri and GNP members, because their political identities are not completely shaped but also remain between 2 different ideologies. This suggests that binary opposition in online political discourse is slowly changing.


The Internet offers an opportunity for political actors (e.g., politicians, activists, government officers) to create their own political communication networks. The latest developments in digital technologies particularly affect communication by enabling new forms of political participation and citizen engagement. Since blogs have become one of the fastest growing media, large numbers of political actors have engaged with blogging. In the 2004 U.S. Presidential election, blogs played the role of attracting citizens’ attention to critical issues and enhancing candidate visibility (Williams et al., 2005). The interconnections between of these blogs were important to support the exchange of information, debate and to express support.

This paper employs the social network model to measure aspects of the political blogosphere in South Korea, represented using a hyperlink network. The goal of our research is to develop methods to identify underlying hyperlinking patterns of political significance between politicians’ blogs and citizens’ blogs, and amongst important citizens’ blogs. There are at least three significant reasons to focus on political blogs and hyperlinks in Korea.

First, official websites are largely decorative and symbolic channels of formal communication between political actors. For example, politicians’ websites are often managed by professionals who may be relatively detached from inner discussions within the organization. Thus, website content and services do not necessary represent the activities of those who are represented on the sites. However, blogs seem to be more personal than websites and political actors are probably more involved in blog-based activities, because blogs are relatively easy to maintain and update.

Second, web users in South Korea often express their personal opinions on topics related to national political interests, although this is typically expressed in an emotional manner rather than as part of a reasoned discourse on matters of public importance (Hague & Uhm, 2005). Online conversations and alliances between political bloggers may be visible online through hyperlinks between them.

Finally, there are few studies about Asian blogs in the international literature (e.g., see Bruns & Jacobs, 2006). Although the Internet and blogs have become embedded in the political and cultural activities of many people around the world, previous research has tended to be concentrated in the North American and Western European experiences. Furthermore, the study of the political dimensions of the Internet in Asia draws upon a conceptual and methodological approach based on a technological-deterministic point of view rather than exploring the implications of the new digital medium within a contextually anchored lens (Ho, Kluver, & Yang, 2003; Goggin & McLelland, 2008; Lim, 2007). This study contributes to gaining a non-Western understanding of, and a social-constructivist approach to, the current and emerging impact of blogs on Korean political practice.

Literature review: Related studies

This section begins with background information about online political campaigning in Korea. Related studies on political blogging in general are then discussed as well as the role and use of blogging technologies in online political communication. Finally, some relevant statistics and trends related to the Internet and blogging in Korea are introduced.

Online campaigning in South Korea

Online campaigning in South Korea started in the 2000 National Assembly election when candidates were allowed to use the Internet (e.g., websites). A second wave of web campaigning was conducted during the 2002 presidential election. The current president Moo-Hyun Roh used campaign tactics and strategies with a specific focus on the mobilization of the reform-minded younger population via cyberspace. The impact of Internet-mediated communication practices on political networking was unprecedented within South Korea and apparently successful for the Uri party (Lee, 2004). More recently, the Internet has also served as the primary medium to help the ruling Uri party win in the 17th National Assembly election in April 2004 (Hague & Uhm, 2005; Hara & Jo, 2007; Kim & Park, 2007).

The Korean Social Science Data Center’s (KSDC, 2004) national survey after the 2004 National Assembly election found that emergent technologies had become embedded into Koreans’ lives in different ways (N = 1,500). According to the KSDC’s results, 44.0% of the ruling Uri Party’s supporters used the Internet every day compared to 16.4% for their GNP opponents. While only 26.6% of those who were favorable to the Uri party had never been online, nearly half of the GNP supporters (49.2%) had never used the Internet.

Since the 17th Assembly members are relevant to the current paper, the election results are briefly explained. Two major parties and two minor parties succeeded in winning seats in the 2004 election: the ruling Uri Party (Uri, Yeo-lin-uri in Korean), the major opposition Grand National Party (GNP, Han-na-ra in Korean), and two minor opposition parties: the Democratic Labor Party (DLP, Min-joo-no-dong in Korean) and the Democratic Party (DP, Min-joo in Korean) which had earlier changed its name from the Millennium Democratic Party (MDP). The Uri and the GNP parties roughly correspond to the Democratic and Republican parties in the U.S. system, respectively. These two parties occupied more than 270 out of the 299 seats in the Assembly. The DP is the former ruling liberal party, while the DLP has progressive (left-wing) policies. A few members are affiliated with small parties or are independent.

Since the 2004 national election, new Internet technologies have been rapidly adopted and web-based personal media have brought challenges for the regulators. The Singapore government banned its citizens from using RSS (Really Simple Syndication) feeds, a major blog technology, for political purposes in the last election (Skoric et al., 2007). Similarly, the National Election Committee (NEC) of South Korea launched a full-scale crackdown on possible irregularities in textual information and audio and visual clips involving civil service elections on 25 April 2007 ahead of the 19 December presidential election. In particular, the NEC set guidelines for unethical practices with homemade Internet video, widely known as UCC (user-created content) in Korea.

Hyperlinks have been used to reflect or cement alliances between political actors and to amplify their voices in cyberspace (Park, Thelwall, & Kluver, 2005). The well-known Loose Change video about the September 11 conspiracy is a good example because more than 10 million people watched this movie in 2006. This rapid underground diffusion would not have been possible without hyperlinks and other forms of information sharing. Links in blogs, as Reynolds (2006) emphasizes, let people experience a new communication world in a way that traditional media can’t. Of course, the link isn’t a guarantee of collaborative social ties between the linker and linked – the relationship may be hostile. In networked politics, successful online campaigning can be helped by creating and maintaining ties with ‘smart crowds’. Political actors, including politicians, are trying to maximize the power of digital media by distributing video clips, photos, and news favorable to them on the blogosphere.

Political blogosphere as networked communication space

Blogs, personal websites that serve as publicly accessible online journals, have become a new communication channel for politics. Politicians, previously reliant upon the mass media (e.g., broadcasting), can now speak and listen to their constituents through blogs (Coleman, 2005). Citizen reporters and professional journalists are increasingly setting up their own blogs in order to report political information and news commentaries. Ordinary citizens, who wish to connect to other people with similar interests, also use blogs to communicate and build interpersonal networks (Reynolds, 2006).

In parallel with the rise of political blogging, there has been research into the political characteristics of the contemporary blogosphere, for example, gender and age distributions of blog authors for some types of blogs (e.g., filter and knowledge blogs) which has shown that the majority of blogs are personal diaries, even though the most publicized bloggers engage with the news in some way (Herring et al., 2005). In a review of political blogging research, Bahnisch (2006) notes that scholarly interest in the way that blogs influence politics and the political process is another significant trend. For instance, some political blogs such as the “Daily Kos” helped Democratic candidates to receive half a million dollars in donations during the 2004 U.S. presidential campaign. Johnson (2006) examined the role of the major political blogs, the conservative “Instapundit” and the liberal “Eschaton,” during the U.S. 2004 presidential debates. The blogs not only fact-checked coverage of the New York Times on the debates between candidates but also presented their own analysis. This reveals that political blogs can attempt to fulfill the key journalistic roles of gatekeeper and government watchdog.

One particularly important issue in the literature on political blogging concerns the hyperlink-mediated networking power of blogs. A blogosphere, created by a collection of blogs, can serve as an online social space. It is known that the whole blogosphere has a globally sparse but locally dense linking structure (Marlow, 2006). This blogosphere property works well for the epidemic-like diffusion of information. A-list blogs, the best-known (most read, most linked to) blogs, could play a key role in triggering or mediating information diffusion within this.

The social role of hyperlinking between blogs has been discussed by Drezner and Farell (2004) in one of the early studies of political blogging. They examined the structural characteristics of mass discussions and information exchange in the U.S. political blogosphere. A few hub blogs had considerable power to promote small pieces of information and to channel public attention to minor blogs. According to Herring et al. (2005) there are a few central blogs with many incoming links from others. Thus, hub blogs contribute to shortening the ‘distance’ between blogs in a massively connected blogosphere. To give a concrete example, in Bruns’ (2007) case study of Australian–born Guantanamo detainee David Hicks, Australian domestic blogs were more strongly engaged in debating news and politics on the case than those outside Australia. A strong and sustained engagement with this case by Australian top political blogs played an important role in diffusing blog discussions in spite of the lack of serious coverage of the case in the mainstream media.

Another feature of A-list blogs is their connections with other blogs sharing similar political stances and ideologies. Previous research suggests that political blog discourse tends to be clustered around a handful of sites run by A-list bloggers with shared political orientations. Studies conducted in the U.S. (Adamic & Glance, 2005; Hargittai et al., 2005), Canada (Elmer et al., 2007), and South Korea (Park & Kluver, 2008) confirm that political subdivisions tend to emerge in blog networks. Participation rarely takes place across ideological borderlines and political issues. More importantly, the ego-network of A-list blogs seems to contribute to polarization among ideological lines.

The networking power of A-list blogs through trackbacks, blogroll links, and citations can also increase the visibility of marginal political actors when they get noticed by these influential bloggers. Nevertheless, the role of hyperlinks between blogs has been relatively little studied and so their real impact remains unknown. For example nothing is known about how common it is for blog reader to follow the various types of links in political blogs.

Blogging and hyperlinking

Hyperlinks permit users to select and reference anything, and so the interconnectivity pattern of hyperlinks between websites can reveal aspects of the sociocommunication landscape of the web. Although the hyperlink analysis of websites has been established in the information and social sciences (Park & Thelwall, 2003; Rogers, 2004; Thelwall, 2004), this is inadequate for hyperlinks in the blogosphere because there are several types that are unique to blogs.

One early study about hyperlinking practices in blogs provides interesting results. During the 2004 U.S. campaign, Williams et al. (2005) compared hyperlink destinations of official websites and the blogs of George W. Bush and John Kerry. Official sites were much more likely to hyperlink to promotional materials (e.g., campaign-related products), political advertisements, and donation requests than blogs. In contrast, when linking to external sources, the most common targets of blog links were blog authors’ official websites and mass media sites. Blogs were judged to be supplemental to other modes of communication.

Another example comes from an interview analysis by Nardi et al. (2004), who found that, compared to websites, interviewees perceived their blogging activities (e.g., writing, hyperlinking, archiving) as more dynamic and somewhat more personal. Blogs often served as a place to post and share thoughts about news and magazine stories. In this regard, Halavais (2006) claimed that blogs are particularly useful for drawing attention to what Merton (1987) calls “specified ignorance”; that is, a new awareness of what is not yet understood but is worth knowing, as tentative ideas can receive immediate and intensive feedback. Blogging in practice often includes hyperlinks to other blogs. Halavais (2006) believes that the hyperlink network among blog authors can represent a snapshot of conversations between bloggers.

In a recent article, Schmidt (2007) argues that “communities of blogging practices” are framed by the three structural dimensions of rule, relation, and code. In terms of hyperlinks, bloggers have a networking rule that guides them to express a social tie to another person through different technical means, for example hyperlinks within a posting or a blogroll, or adding a comment to another blog. He also claims that establishing relations between blogs occurs via hyperlinks. In particular, blogroll links can convey important aspects of social relations, such as being a sign of personal acquaintance or friendship, professional affiliation, or expressing consent (or dissent) with the linked blog.

As discussed above, the underlying architecture of the blogosphere is partly structured by the use of hyperlinks. The analysis of hyperlink connections between blogs can thus be a starting point for the understanding of blogging practices.

Use of the Internet and blogs in Korea

According to an International Telecommunication Union (2006) report, the adoption of digital technologies in South Korea is widespread. Korea had the world’s second highest broadband access rate (25.20%) and about 80% used a mobile phone in 2005. By December 2006, the National Internet Development Agency of Korea (NIDA, 2007) announced that there were 34.12M Internet users (74.8%), a 1.11M increase from the previous year. Moreover, as of December 2006, an incredible 39.6% of Korean Internet users had published a blog (including a Korean style “minihompy” blog) and the most active blog user age group was the twentysomethings (68.2%) (NIDA, 2007). A total of 47.4% of Korean netizens claimed to have visited other people’s blogs. A minihompy (inline image) is in fact a location within the social network site Cyworld, which is similar to MySpace (boyd & Ellison, 2007). This location includes blog functions in addition to other features, such as friending (inline image) and a photo gallery. Its distinctive attribute is that an owner’s avatar ‘lives’ in a virtual room in the site.

However, not all of those who use the Internet engage in putting news and personal commentaries together. According to Edelman (2007) about 43% of Koreans read blogs more than once a week (France: 22%, UK: 23%, USA: 27%). A total of 63% of Korean blog readers identified themselves as social influencers and read blogs 3.06 times per week. Social influencers are those who have taken part in at least three activities (e.g., email or call politicians, attend political events or public hearings, serve as a public committee member, petition, make a public speech or media contribution). This is very high compared to other countries (France: 37% with 0.97 blog readings, UK: 35% with 1.11 blog readings, USA: 34% with 1.20 blog readings). The survey was conducted across 1,000 people during November 2006.

Some blogs have been very active in political discussions. Park and Jankowski (2008) found that popular political blogs run by lay persons in Korea explicitly express a political stance both as producers and as readers of content. For example, their blog titles are self-expressive and strongly imply ideological affiliations in Korean politics. As Reynolds (2006, pp.117–118) states, a personal voice is important for good blogging and the selection of hyperlinks helps to represent this.

Research question

Given the rising influence of online networking in Korean politics, this paper focuses on influential citizen blogs in South Korean politics, addressing the following research question. What hyperlink patterns exist between influential citizen bloggers and politicians in South Korea?

Methods: Network data and measurement

Methods for detecting A-list blogs

There are several types of hyperlink in blogs. A blogroll link is a blog sidebar hyperlink to other blogs that author reads or otherwise recommends. A content link is a hyperlink embedded in a post or commentary. A navigation link directs visitors to internal blog features. A paid link is a hyperlink to advertisers (e.g., Google ad-sense). Data on social networks may be gathered for all linking elements of a blog, but this research examines only blogroll links. Amongst these blog hyperlinks, a list of blogroll links to other blogs is of particular interest to blog researchers since they can provide clues about the blogger’s political agenda and affiliations (Schmidt, 2007).

Most blog link studies have dealt with hyperlinks in blogrolls because these links are set up and maintained by bloggers themselves perhaps for regular navigation to linked blogs. Blog hyperlinks may be tokens of the blog producer’s communication networks, informal associations with the other blogs. But, should blog hyperlinks, as social ties, be regarded as strong or weak ties? (Granovetter, 1973). People with strong ties tend to have mutual obligations, and give support and affection when others are in need. On the other hand, social networks in cyberspace generally tend to be weaker ties by nature (Haythornthwaite, 2002). Since online ties often do not reflect a long-term acquaintanceship, mutual duties, social commitment, and visual information, and even rely at times on false names, they are typically easier to break (Adamic & Adar, 2005). However, the classification of online relationships in cyberspace into strong or weak ties can be constrained by a number of contextual factors. People with strong ties are prone to employ more kinds of communication media to meet their informational and emotional needs, particularly when they are temporarily unavailable and geographically distant (Kim et al., 2007). In this case, online technologies such as blogs often are used as the functional equivalent of other communication methods, because of the asynchronous, cheap nature. Thus, the strength of online ties can vary according to the specific context under investigation. Furthermore, the characteristics of offline ties can be reflected in online behavior. For example, since politicians sometimes link to one another in their websites, the quality of e-relationships conveyed via blogroll links can be partially attributed to traditional bonds from primary group affiliations such as gender, locality, and party.

We first identified Assembly members that use the blog service: is the most popular web portal in South Korea. According to Park and Jankowski (2008), 124 out of 299 National Assembly members maintained personal blogs during July 2005. The majority, 115, used the Naver blog service. Thus, we decided to focus on the Naver in this study. More specifically, we visited the official websites of Assembly members and collected their blog URLs. Search engines were also used to trace blogs. An identification process starting from the list of already identified blogs was also employed. We excluded from our analysis a few inactive blogs with no posts after the initial posting.

To find influential citizen bloggers, we used quasi-snowball sampling from the blogroll links of the 124 Korean National Assembly members, finding a total of 1,904 citizen blogs. Of these, 79 blogs were linked more than three times, forming our top linked ‘elite citizen bloggers’ set. Two data sets: (a) blogroll links between politicians and citizens, and (b) blogroll links amongst elite citizens were manually compiled during January 2007. Tables 1, 2, and 3 give a detailed overview of the sampling procedure and data. Note that not all of the politicians linked to one or more citizens. The average number of hyperlink connections to citizens was 2.85 with a standard deviation of 5.96 and a range of 0-40 links. Some politicians linked to other politicians, parties, or NGOs on their blogrolls but any such links were ignored.

Table 1.  The sampling procedure for citizen blogs
StageActionMethodNo. of blogs
1Systematic identificationBlog-using politicians traced via their official sites and multiple search engines124 politicians
2Snowball samplingLayperson’s blogs in the blogrolls of the politicians1904 citizens
3Purposive samplingCitizens linked to more than 3 times (a set of politically active actors)79 citizens
4Making the full-networkInformation about each citizen’s links to all others collected79 citizens
5Dissecting the full-network79 citizens divided into groups based on link reciprocity between citizens and politicians47 citizens
Table 2.  Distribution of citizen blogs according to linking members’ party affiliation
PartyNo. of politiciansNo. of citizens linked by politiciansNo. of elite citizens linked by politicians
Table 3.  Distribution of inlinks to citizen blogs according to linking members’ party affiliation
PartyPoliticiansLinks to all 1,904 citizen blogsLinks to the 79 elite citizen blogs

Social network analysis design

Social network analysis is a set of methods to identify interconnectivity patterns amongst individuals (also called “nodes”) based on shared attributes (Hanneman & Riddle, 2005). Several techniques are able to reveal relational features amongst components within a social system under investigation, and between groups based upon the shared attributes of group members. For this study, social network analysis was performed using standard routines from UciNet for Windows (Borgatti, Everett, & Freeman, 2002).

In this paper, we concentrate on two primary analytical techniques for measuring social relationships represented by (a) blogroll links between politicians and citizens, and (b) blogroll links amongst citizens. First, for the data set (a), “positional” properties of particular network actors were identified through a network analysis of citizens’ affiliations with politicians. Positional approaches were employed in order to determine citizens’ tie-strengths with politicians and political characteristics of citizens.

The tie-strengths of citizen blogs were classified according to their interaction types. Citizens having mutual links with a National Assembly member were categorized as “strongly bonded”. Those which did not link back to Assemblymen were called “weakly bonded”. Amongst the 79 elite blogs, 47 are strongly bonded but only 32 are weakly bonded. Next, the partisanship of the elite blogs was examined with an event-based approach depending on involvement in selected activities. The primary criterion for the identification of partisanship was the existence of a hyperlink between a citizen and a politician. In support of this, Soon and Kluver (2007), in their analysis of Singaporean political organizations, found that hyperlinks served as new type of political marker because link network structure, to some degree, matched ideologies. Sunstein (2007), in his recent book, 2.0, emphasizes that links embedded in the website need to be analyzed as a form of political action in the digital age. Table 4 summarizes the distribution of strongly and weakly bonded blogs by partisanship.

Table 4.  Party connections of the 79 elite blogs
ItemsStrongly bondedWeakly bondedTotal
Uri + GNP111325
Uri + DLP303
GNP + DLP101
Uri +GNP + DLP101

Second, for data set (b), relational properties of civic actors based on network structure are examined through a whole network analysis. For the relational analysis, a “who-links-to-whom” matrix amongst the 79 citizen blogs was made. Relational network analysis is particularly useful when employed with a positional approach. The 79 * 79 matrix illustrates social connectedness within a single set of network actors. While positional (affiliation) network analysis is restricted to outgoing links of network actors, relational (whole) network analysis yields insights into internal connectedness represented in all links between actors. The frequently used indicators and techniques in this study are as follows.

Degree centrality measures the number of relationships involving individual actors; in this case, citizen blogs (Freeman, 1979). Degree centrality can be separated according to the direction of relationships between actors. Indegree refers to the number of direct connections from others in the network. Outdegree means the number of others to which a blog links. Centrality usually comes with its sister indicator “centralization” that shows the extent to which relationships in network are associated with a central actor. A high centralization suggests that the central actor is influential in shaping the overall network. While centrality seeks to report individual attributes, “density” measures the internal relations of the whole network. Density is the number of existing connections amongst actors divided by their possible connections (Wasserman & Faust, 1994).

Finally, citizen blogs are clustered naturally by the positional approach described earlier and it is useful to investigate the interrelations between those subclusters within the larger network. In this regard, the frequency of link “exchange” between subclusters was measured, and their link “traffic” was visualized with a simple diagram.

Analysis and Results

Positional affiliation network analysis

The affiliation network of the 47 strongly bonded elite blogs is shown in Figure 1. There are apparently substantial differences in civic involvement between the ruling and opposition parties. The percentage of citizens affiliated with the ruling Uri party is large in comparison to the opposition GNP, confirming previous findings with different data sets (Park & Jankowski, 2008). Using a similar approach, Park and Jankowski (2008) found that the progressive Uri camp in South Korea had been consistently better organized than the opposition in terms of the number of blog hyperlink connections after two major elections.

Figure 1.

The affiliation network of the 47 strongly bonded elite blogs.

As described in the method section, the assignment of individual citizen blogs to partisanship categories was made according to the existence of reciprocal hyperlinks between politicians and citizens. In order to check this assumption, a content analysis was made of several citizen blogs’ posts and titles, and text accompanying their blogroll links. One of the common features used by bloggers to differentiate themselves from others is their blog titles. In regard to posts, Korean bloggers using the Naver blogs under investigation are able to block strangers from reading their blog posts, although their post titles are generally open to everybody.

One GNP-affiliated citizen, “lionrise,” with a fairly high centrality gave his/her blog a striking label, “devotion to my country and people, Dear Geun-Hye Park”. Miss Park is the former president of the opposition GNP party and is a daughter of the military dictator Jeong-Hee Park who got assassinated during his term. The “lionrise” blog was also included in the category of strict and legitimate conservatives on the side bar in another GNP blogger “jskim8875” whose ID represents another Korean conservative politician Jong-Phil Kim. Compared to the GNP, the ruling Uri-tied blogs frequently attempted to address public concerns about the reforming policies of incumbent president Roh. For example, one Uri blogger “tank1801” asserted the importance of listing and removing Korean elites with positions above medium ranked-governmental/military officer and who were in a close relationship with Japan during colonization period. According to the writings of some Uri bloggers, this government program was necessary to restore national self-esteem. Furthermore, as this specific issue had become popular, blog conversations were shaped along similar ideological lines because the ruling party tried to include deceased president Park in this pro-Japanese tier. Another Uri blogger (asraee) identified him/herself as an Uri supporter by setting up a noticeable link to Mr. Roh’s presidential office. This blogger also packed in their blogroll a set of blogs that (s)he describes as “the best Roh people”. The DLP is well known for its social and democratic policies. Individuals (e.g., since4299) associated with the DLP strongly provided their own party preference using self-referential framing of the anti-American fighter’s issue stances with respect to the expansion of the U.S. army base in Korea.

Interestingly, the current results partly contrast with a study (Park & Kluver, 2008) of online networks amongst the blogs of South Korean politicians alone. The study showed that conservative GNP members were highly visible and noticeable in cyberspace in terms of technology adoption rates. After the GNP lost the 2002 presidential and 2004 national elections, its politicians increasingly created their own websites or became concerned with online political activities and services, either to complement offline weakness or to strengthen their competitive positions online. Park and Kluver (2008) found that there was a concerted use of blogs and other associated technologies by GNP members. Hyperlinking by GNP members was also more extensive than by Uri members. The findings from the current study show, however, that this is not true in the context of affiliation between politicians and citizens.

Another significant finding from the affiliation network analysis is that several blogs are located at the center of the Uri and GNP clusters and are connected to both camps. There are a number of citizen blogs with a dual “partisanship” as they link to both Uri and GNP members. A qualitative investigation of these blogs reveals that three of the citizen bloggers are professional journalists (jay1124, rogos0119, scaletqueen1) with a relatively long experience of blogging. There are also prospective politicians (cys2007, wkh55), a progressive poet (imim0123), and some laypersons. We were not able to identify the remaining bloggers due to a lack of information on their blogs. Judging from their blog titles and posts, the majority of blogs with dual partisanship are politically engaged citizens that provide quick comments on ongoing political events. However, it was not possible to determine if they have formal offline party affiliations. Presumably their “dual partisanship” had the purpose of portraying impartiality or the possession of influential friends.

Table 5 shows that blogs affiliated with the ruling Uri party are very active online in terms of informing and networking. Activity was measured using four items: posts, citations, hyperlinks, and visitors. Since these four items do not have a Gaussian normal distribution, the mean is not a good measure of average and so the median is used instead. Each item is defined as follows:

Table 5.  Indicators of online social capital for the 47 strongly bonded elite blogs
PartisanshipStatisticsPostsCitationsBlogroll linksVisitors
Uri (N = 18)Min9644277,256
GNP (N = 11)Min26251110,219
DLP (N = 2)Min34712917616,586
Uri + GNP (N = 11)Min8710166,054
Uri + DLP (N = 3)Min6010262,852
GNP + DLP (N = 1)Min4216397,047
Uri +GNP + DLP (N = 1)MinNA42285162,196
  • • Posts: total number of posts in the blog
  • • Citations: total number of citations from other bloggers as measured by inlink frequencies
  • • Blogroll links: total number of outgoing blogroll links
  • • Visitors: total number of visitors as recorded by visitor counts

What is important about the findings of this additional analysis? It seems that the political blogosphere of Korea is constructed in such a way that the Uri party is attempting to gain a competitive advantage through hyperlink-mediated associations with its supporters. The Uri-related bloggers form an independent group separate from the remaining party groups, as the former was the most active information content providers among all political actors and played as dissemination axes of blog commentary. The Uri-affiliated citizen blogs functioned as platform, whereas the Uri politicians’ blogs were placed firmly on the agenda of the blogosphere and reached a number of civic bloggers in an immediate way. In past elections, Uri social relationships established through online communication channels were used as interactive and participatory tools. Similarly to U.S. Democratic candidates in 2004, the extension of the online network may have reduced election costs for Roh. In summary, the analyses of blog activity and engagements, when taken with link network indicators, reveals that the Uri-tied blogs, as opinion leaders and agenda settlers, benefited from this technology in the sense of involving bloggers in the ruling party’s domain.

Relational whole network analysis

A whole network analysis of the 79 citizen blogs was conducted to examine their interconnectivity.

Tables 6 and 7 list the top 10 blogs in terms of indegree and outdegree values. Both incoming and outgoing blogroll links are connected to progressive and central blogs affiliated with either the Uri or the Uri and GNP parties. There are no conservative GNP-only blogs in these first-tier groups.

Table 6.  Top 10 citizen blogs in terms of outdegree centralities
1ssk3169Uri + GNP18
2ek3906Uri + DLP15
3josjosjosUri + GNP14
4cys2007Uri + GNP13
6since1934Uri + DLP12
Table 7.  Top 10 citizen blogs in terms of indegree centralities
2kwclUri + GNP11
3josjosjosUri + GNP9
4doolykingUri + GNP9
6rogos0119Uri + GNP8
7lucky31Uri + GNP7

Figure 2 illustrates the hyperlink relationships between the 79 elite blogs. The most linked blog, myhong0910, is located at the center of the whole network map. Also, this diagram shows the centrality of blogs favorable to the ruling Uri party.

Figure 2.

Blogroll interlinking of the 79 elite blogs.

• Uri – Yellow, GNP – Blue, DLP – Light green, DP – Red, Uri + GNP – Sky blue, Uri + DLP – Purple, GNP+DLP –Pink, Uri+GNP+DLP–Grey. Node size is proportionally bigger according to node indegree. Circle nodes are strongly bonded blogs and square nodes are weakly bonded blogs.

Additionally, a system analysis of the network yields useful information for understanding the structural characteristics of the network. The low density value 0.0409 indicates that the network of the elite blogs is sparse. Only 4% of possible ties exist in the blogrolls. Inbound and outbound network centralizations are 15.36% and 19.23% respectively. These relatively moderate measures reveal that this network forms a distributed chain shape rather than a hub-and-spoke topology. This suggests the dominance of horizontal networking. Ten blogging citizens were found to be isolated, with no hyperlinks to and from other members in the network.

Combined network analysis

Next, we used a combined network analysis to seek hidden relations amongst the 47 strongly bonded elite blogs. In Figure 3, the links between the 47 blogs (expressed as thin and light gray lines) are inserted within the affiliation network diagram. Given that blogroll links are used as a surrogate measure of ideological affiliation, note that the blogs on the left and central area in Figure 4 have a number of transverse links from one to another.

Figure 3.

Blogroll interlinking of the 47 elite strongly bonded blogs.

Figure 4.

The strength of linking amongst the party groupings.

• See Figure 2 for the colors.

On other hand, the opposition GNP-affiliated citizen blogs do not have a high degree of hyperlinking amongst themselves. This suggests that blog technology helps the progressive Uri party-affiliated citizens extend their information potential and deepen their alliances. Reform-minded citizens are perhaps exploring the ways social networking technologies, including blogs, can help build relationships. During this process, hyperlinks have perhaps become a channel for signifying political affinity and connecting to related information.

For the additional analysis, the 47 strongly bonded elite citizen blogs were organized according their party affiliations. The results show a discrepancy between the ruling Uri party-affiliated blogs and opposition blogs. Interestingly, there are 34 and 2 hyperlinks within the Uri and GNP parties respectively. Nearly all blogs with dual partisanship (Uri & GNP) send hyperlinks to Uri citizen blogs. Uri and GNP-tied bloggers (11 citizens) had 24 out of 50 hyperlinks to Uri blogs. This is more than the blogroll links within the group. Examining reverse linking, Uri blogs (18 citizens) send 14 links out of 59 to Uri and GNP blogs. This suggests that the overall level of network integration within the Uri party is extremely high. These results are shown in Figure 4. Arrowheads represent the direction of blogroll link traffic and arrow widths are proportional to link counts. The more self-links, the bigger the node.


This research suggests that blogs have become a popular channel for political communication whilst also being interwoven with many political actors. As a whole, online interactions via blogs fill communication gaps between the ruling Uri party members and their core supporters. Although the Internet sometimes enhances only weak online ties, the progressive Uri camp seems to be obtaining political and organizational capital from the Internet in an affordable and convenient way. With a relatively homogeneous population sharing ethnicity and language and with a high proportion having broadband access, the political discourse provided by A-list blogs can potentially enjoy widespread penetration. At the time of data collection, the Uri party and President Roh were becoming unpopular due to Roh’s frequent verbal slip-ups concerning important national issues. Despite this, the ruling Uri-tied citizen bloggers who are influential in the blogosphere seem to provide significant support.

However, some finding suggests that influential A-list blog authors are individuals who have become disenchanted with the Uri party and several bloggers have been drawn either to a politically central position or to the conservative camp. Since the National Assembly election in 2004, the importance of ideological issues has been declining among the young and progressive citizens who were the primary supporters of the ruling Uri party (Kim, Choi, and Cho, 2008). Their apparently misguided reform policies and rapidly increasing unemployment have clearly caused problems. This might have lead to a situation in which the importance of blogosphere activists in rallying support is even greater- particularly for the more centrally placed bloggers; conversely, a disinterested population may no longer want to read blogs about politics.

Nevertheless, bloggers are mainly polarized by political stance in South Korea. New media technologies may thus reinforce certain political and organizational advantages of the central conservatives, just as this position was exploited by progressive liberals in the past. As increasingly many individuals are now able to participate in political discussions online because of the ease of joining groups of similar others, our findings suggest that blogging reflects, or even contributes to, the fragmentation to Korean politics.

Similarly, political blogs of American citizens seem to be grouped along ideological lines, for example, with Andrew Sullivan’s blog on the right or the Daily Kos on the left. In the 2004 U.S. presidential elections, American politicians appeared to be active bloggers. They tended to use blogging to inform voters about the campaign, particularly during the election period (Williams et al., 2005). In contrast, British politicians are known to have a motivation to blog whilst in office as an effort to communicate with their constituents about ongoing political events (Trammell, 2006). More strikingly, Cameron, a leader of the UK conservative party, has run a video blog ( since October of 2006 (Woodward, 2006). Korean politicians also attempt to quickly adopt new technologies to reach their constituents at various levels. As seen in this study, the blog is being recognized by Korean political players as a cyber-campaign tool that helps to build a network. To be more specific, given that the distinct national political culture (e.g., collectivistic political discussions), online chain-referral via blogroll links is often perceived as fairly powerful and credible and is likely to generate a strong impact on the political process and electoral results (Park & Kluver, 2008). Thus, there are a number of ordinary blogs run by those who are elected as well as prospective politicians.

When it comes to citizen blogs and their possible “whispering effect” on political information and communication, the Korean model seems to be both similar to and different from that of the US. Social activists and (ex)journalists are prominent political bloggers in the US and their professional blogging often motivates the public to question political campaigns. Furthermore, American A-list elite bloggers became a financial channel with links to donation sites. In contrast, it seems that Korean political bloggers range from college students to salaried white collar workers. The Korean blogosphere seems to be repeatedly filled with poor quality comment and debate, and with hate speech. Academics seem to seldom run political blogs. This perhaps reflects the emotional attachment of Korean people to political events and the self-censorship of Korean elites.


This research provides some interesting interpretations with respect to South Korean national politics. First, the percentage of elite blogging citizens affiliated with the ruling Uri party is large in comparison to the opposition GNP, confirming previous research. Furthermore, blogs affiliated with the ruling Uri party are the most active online. Many GNP members run blogs but they are weakly coupled with citizen groups. Thus, the information penetration of the opposition conservative party is narrow in the popular blogosphere. On the other hand, few Uri Assemblymen are active bloggers but they have relatively many strong ties with ordinary people in the blogosphere. This suggests that the Uri party sustains its online presence through supporters’ blogs and its messages are probably efficiently communicated through blog networks. This perhaps reflects the power that “an army of Davids” (Reynolds, 2006) can have on politics.

The blog network topology was relatively sparse because only 4% of possible ties among the 79 elite citizen blogs existed in the blogrolls. Nevertheless, the analysis points to the Uri party attracting the DLP away from the GNP in terms of links. Moreover, several blogs are at the center of the Uri and GNP link clusters and are connected to both camps. These are potentially the most important blogs, and only a minority of them belongs to journalists. The rise of this “central citizen” group is potentially a signal for the development of the political landscape in the near future. These two factors suggest that the landscape of Korean political discourse may be changing. At the time of data gathering for this research, the Korea presidential election was a year away.

In the Korean 2007 presidential elections, the candidate of the main opposition party (GNP) Lee Myung-bak won the presidency by a significant margin. The importance of “economic development” rather than “democratic participation” had rapidly grown and as a result, the influence of online networking power probably declined during the 2007 elections. The most noticeable difference from the previous elections, was that the Internet is no longer the exclusive preserve of young liberals who are more likely to be affiliated with the ruling party (Park & Lee, 2008). A growing number of conservatives have utilized political websites since defeat in the 2002 presidential and 2004 parliamentary elections. Widening access to political websites across the generational and political spectrums is one of the contributing factors to the conservatization of Korean cyberspace. Our research clearly suggested that the networking capacity of the ruling Uri party, measured in terms of the number of citizen bloggers hyperlinked both to and from politician sites, would gradually wane.

As the first analysis of its kind, this study has wider implications for the value of blogroll links as an analytical tool to help understand the relationship between politicians and influential citizen bloggers in Korea, the US, or elsewhere. Although we have not validated our results in the sense of taking steps to assess their reasonableness, this study has clearly given a way to make sense of what might otherwise be an undifferentiated mass blogosphere. Future research may reveal whether this is the best approach and whether the information gained has genuine practical use. Nevertheless, the outlook is promising in the sense of the development of a practical set of methods that have yielded (so far) simple and reasonable results.


This research was supported by a Korea Research Foundation Grant (KRF-2004-042-H00004). This paper is a part of the “Web Sphere Analysis for Political Websites” Project.

About the Author

  1. Han Woo PARK is an Assistant Professor in the Department of Communication & Information, YeungNam University, South Korea, where he teach courses and conduct research on the use of new communication technologies in extending social and political networks, and e-science/e-research topics. Over past few years, he has contributed important works in the area of Link Analysis (also called, Webometrics) from the perspective of Social Network Analysis. Currently, he is a coeditor of the Journal of Contemporary Eastern Asia.Email: /
    Website: http:/
    Address: Dept of Communication & Information, YeungNam University
    214-1, Dae-dong, Gyeongsan-si,Gyeongsangbuk-do,
    South Korea, Zip Code 712-749

  2. Mike Thelwall is Professor of Information Science and leader of the Statistical Cybermetrics Research Group at the University of Wolverhampton, UK. He is also visiting fellow of the Amsterdam Virtual Knowledge Studio, a Docent at Åbo Akademi University Department of Information Studies, and a research associate at the Oxford Internet Institute. He has developed tools for downloading and analyzing web sites, blogs and social networking sites, including the research web crawler SocSciBot and the LexiURL link analysis software. He is the author of the book Link Analysis: An Information Science Approach, and sits on nine editorial boards.Email:
    Address: School of Computing and IT, University of Wolverhampton, Wulfruna Street
    Wolverhampton, WV1 1LY, UK