Is There an Elite Hold? Traditional Media to Social Media Agenda Setting Influence in Blog Networks


According to several 2008 reports, blogging continues to attract writers and readers (Comscore Media Matrix, 2008; eMarketer, 2008; Sifry, 2008; Universal McCann, 2008). This form of Web content creation has matured beyond public personal journaling to support citizen journalism or journalism produced by independent bloggers unaffiliated with professional newsrooms (Lenhart & Fox, 2006). The popularity of blogs is in part fueled by its interactive format: The blog tool is popularly believed to be a vehicle of democracy because it fosters decentralized citizen control as opposed to hierarchical, elite control (Crumlish, 2004; Levine, Locke, Searls, & Weinberger, 2001; Rosen, 2006; Scoble & Israel, 2006; Suroweicki, 2005; Weinberger, 2003, 2008). This inversion of elite control is the social outcome of a more interactive format. Blogs are popularly viewed as a form of social media, or media that is architected by design to readily support participation, peer-to-peer conversation, collaboration, and community (O’Reilly, 2004). Social media tools such as blogs enable Web content creators to circumvent the high transaction costs that once characterized usage of earlier media technologies (Gillmor, 2004; Benkler, 2006; Bowman & Willis, 2003; Shirky, 2008).

Independent political bloggers that comment on day-to-day news command a readership rivaling that of traditional media entities (Armstrong & Moulitsas Zuniga, 2006). The initial public derision heaped by traditional media entities on these independent bloggers unaffiliated with traditional, professional newsrooms (Rosen, 2005) continues to wane as these bloggers gain respect among Web readers (Johnson & Kaye, 2004). Top independent political bloggers have played an influential role in holding public officials accountable from Trent Lott to Dan Rather (Meraz, 2008). The blog form has matured to resemble traditional journalism in form and practice: Top, independent bloggers now hire editors, blog full-time, and engage in investigative journalism acts (Stoller, 2007; Strupp, 2008). The growth in the independent political blogger's credibility has taken place against the backdrop of traditional media's loss of a media monopoly on news creation and dissemination (Gillmor, 2004; Rosen, 2006). In an effort to survive, traditional newsrooms have embraced newsroom blogs as an alternative vehicle for news delivery (Meraz, 2008). This trend in traditional media blogging first grew to public visibility in the United States during the 2004 Democratic and Republican national conventions when traditional media bloggers showed up to “live-blog” these events with top independent political bloggers who were given press passes (Meraz, 2005).

This study questions the agenda setting and social influence of elite traditional media outlets both among top independent political bloggers and among elite newsroom blogs. Using the hyperlink as an external marker of source influence, this study assesses hyperlink usage across three diverse issues in 2007 within the 11 newsroom political blogs of the New York Times and the Washington Post and within 18 top U.S. independent political blogs across the political spectrum.

The timeliness of this study is underscored by two central concerns. Firstly, blogs have been widely heralded as a “democratizing innovation” (Von Hippel, 2005, Bruns, 2005, 2008); yet, this study questions whether citizen media outlets are really inverting the power structure between citizens and elite traditional media entities (Armstrong & Moulitsas Zuniga, 2006; Reynolds, 2006) through analyzing source influence. Secondly, this study examines how traditional media is utilizing the blog tool. The question can be posed: Do elite media entities utilize their newsroom political blogs as tools for forging networked conversations with the best of independent political blogging, or has it continued to “normalize the blog form” (Singer, 2005) with insular and parochial linking practices to other traditional mass media newsrooms? Such practices would indicate that this “more of the same journalism” may not be an effective gatekeeping practice for traditional newsrooms who seek to reconnect with engaged Web publics.

Literature Review

Agenda Setting and Gatekeeping

Agenda setting has systematically sought to document the effects of mass media on the audience's cognitions. By virtue of creating a shared, national pseudo-environment, mass media fulfill the important function of building a public consensus on the important issues of the day (Lippmann, 1949; McCombs, 1997; McCombs, 2004). As one of the more enduring and most researched theories in mass communication and political communication (Bryant & Miron, 2004; McCombs, 2004; Weaver, 2007) , agenda setting has matured as a theory to include a second-level agenda setting component (attribute agenda setting), a psychological component to explain individual-level agenda setting effects (need for orientation), an emphasis on how the media's agenda is shaped, and an explanation for the shared news agenda among different media (intermedia agenda setting). The theory, formulated when traditional media entities owned the tools of content creation and content distribution, remains largely untested in the new, interactive media age (McCombs, 2005).

This study is centrally concerned with the intermedia agenda setting dynamics between elite traditional newsroom blogs and top independent political blogs, as well as the intermedia agenda setting influence within these said social networks. Prior research has found strong intermedia agenda setting effects between elite-to-less-elite traditional mass media entities (Reese & Danielian, 1989; Lim, 2006), newspaper coverage to television news broadcasts (Lopez-Escobar et al., 1998; Roberts & McCombs, 1994) and newspaper coverage to wire services (Lim, 2006). In the new media environment, evidence has also been found for convergent agendas between portal news outlets and traditional, elite media entities in television and newspaper industries (Yu & Aikat, 2006).

Few studies have been conducted on traditional mass media's agenda setting ability in citizen-controlled Internet spaces. Lee et al. (2005) found higher correlations in support of newspaper influence on online bulletin board conversations in South Korea. Roberts, Wanta, and Dzwo (2002) found that U.S. online media were successfully able to set the agenda for U.S. electronic bulletin board users for three issues with a time-lag influence varying between 1 to 7 days. In one of the few studies to test mutual influence between traditional media and the independent political blogosphere, Cornfield et al. (2005) found a correlation of .78 for traditional media-to-blog influence as opposed to a correlation of .68 for blog-to-traditional media influence.

Issue characteristics have been shown to play a strong role in mediating the traditional media's issue agenda setting effect. Prior research has shown that media's agenda setting impact is stronger with unobtrusive issues or issues that the audience has less direct, personalized experience with in real-world contexts (Zucker, 1978; Winter, Eyal, & Rogers, 1982; Smith, 1987; Weaver et al., 1981). The media's ability to influence the audience's agenda with issues is also related to the nature of the issue: Abstract issues are more difficult to transfer to the audience than concrete issues (Yagade & Dozier, 1990). The media are also better able to set the agenda in relation to more dramatic and conflict-laden events (MacKuen &Coombs, 1981; Wanta & Hu, 1993) and with shorter time-frame issues (Downs, 1972; Zucker, 1978). When assessing the agenda setting influence between traditional mass media and the independent political blogosphere, issue transfer is more feasible with issues that can support buzz and partisan framing within the highly partisan U.S. political blogosphere (Meraz, 2008; Cornfield et al., 2005).

Gatekeeping is another relevant theory that has been subsumed under agenda setting. Many scholars question whether gatekeeping can be a tenable theory in the decentralized, new media environment where media abundance negates the role of a central news gatekeeper (Anderson, 2005; Bennett, 2004; Katz, 1992; Kovach & Rosensteil, 2007; Lasica, 1996; Williams & Deli Carpini, 2000, 2004). Bruns (2005, 2008) reconceptualizes gatekeeping as gatewatching to account for the increased power of the decentralized ‘produser’ netizen who can republish, refilter, and produce media content for a networked Web 2.0 environment.

Scholars who study gatekeeping in online media have used the hyperlink as a symbolic representation of a single gatekeeping act (Dimitrova et al., 2003; Trench & Quinn, 2003). Though traditional media newsrooms' usage of hyperlinks has grown as a function of time (Tremayne, 2004), traditional media continue practice internal, insular linking practices within their traditional newspapers (Deuze, 2003; Dimitrova et al., 2003). Examining blog influence on media reports, scholars have found evidence of traditional mass media's dependence on top, political bloggers (Farrell & Drezner, 2008; Cornfield et al., 2005; Herring, Kouper, Scheidt & Wright, 2004; Meraz, 2008). In one of the few studies on traditional newsroom blogs, Singer (2005) found that traditional media outlets were more prone to legitimize other traditional media outlets. In the realm of traditional-to-citizen media influence, Reese et al. (2007) found greater reliance on traditional media than citizen media within top, citizen media outlets.

This study provides a more robust test to these prior studies through tracing influence across three issues within blogs of elite traditional media entities and within top, independent political blogs of diverse ideological perspectives. Tapping into a diverse array of issues and an elite blog sample, this study questions the nature of citizen media to traditional media interdependence through the lens of source influence in their news reports.

Social Network Analysis

Social network analysis is an interdisciplinary theory of structural relations that focuses on the relationships and linkages between individuals. Wasserman and Faust (1997) define social network analysis as the study of “relationships among social entities, and on the patterns and implications of these relationships.” Emphasis is placed on the interdependence of actors and their actions, the relational ties between actors, the network structure of ties between and among individuals, and the conceptualization of network structure along social, political, and economic dimensions. This study takes the approach that social network theory can be used to explain the boundaries of potential source influence and the potential power of agenda setters within specified social networks.

In World Wide Web networks like the independent blogosphere, relationships among actors are expressed through the hyperlink, and a series of studies examine the ‘elite’ bias, or the ‘power-law’ principle of these growing, open networks where a few sites command the majority of attention (Farrell & Drezner, 2008; Kottke, 2003; Shirky, 2003; Thompson, 2006). Often called scale-free networks because there are no barriers to entry, World Wide networks such as the blogosphere eventually acquire a hidden order through preferential attachment, a situation where a few actors gain disproportionate influence and attention (Barabasi & Albert, 1999; Kleinberg & Lawrence, 2001). This scenario, modeled through the power law (see Figure 1) highlights an inequality in influence. Within the blogosphere, this inequality among popular, “A-list” or “short-head” blogs is driven by the egalitarian metrics of link traffic and link totals: Popular blogs are propelled to celebrity status through proven credibility and reputation, which leads to even more traffic. Though it has been suggested that this growth in blogger aristocracy threatens the democratic potential of blogs (Bowers, 2007; Thompson, 2006), in-coming entrants to the blogosphere are not barred from attaining similar celebrity status (Shirky, 2003). As hubs, popular political blogs also drive traffic to lesser-known blogs by connecting and filtering the disparate parts of the blogosphere through hyperlink usage in their news stories (Kleinberg, 1997). Anderson (2005) argues that in many of these networks, the traffic amassed from long-tail options can outweigh or exceed that of the elite actors, thus neutralizing the singular influence of the powerful few.

Figure 1.

Power Law Inequality In Independent Political Blog Networks and Traditional Media Blogs.

Of particular importance to this study is the subfacet of social network theory called homophily theory or “birds of a feather flock together” theory. The powerful tendency that individuals have to network along homogeneous lines is driven by shared beliefs, interests, and social status (Lazersfeld & Merton, 1954; McPhearson et al., 2001). Within the independent U.S. political blogosphere, several studies have found segmentation along partisan lines in diverse environments (Adamic & Glance, 2005; Meraz, 2005; Tremayne et al., 2006). Sunstein (2000, 2001, 2002) warns that this partisan, group polarization limits deliberative, democratic conversation by encouraging “enclave extremism.”

This study seeks to examine homophily through the specific lens of status homophily (Lazersfeld & Merton, 1954). This study contemplates the following: To what extent do elite traditional media newsrooms extend their elite bias by choosing to link to other professional newsroom organizations as opposed to amateur, citizen media? Alternatively, it could be asked, to what extent does the independent political blogger's outside status from mainstream media result in greater dependence on other citizen bloggers as sources in news reports? The tendency for traditional media to erect walled gardens (Singer, 2005) and the popular perception that citizen media platforms provide the average citizen greater opportunity to be a news source (Gillmor, 2004; Bowman & Willis, 2003) will both be tested in this study.

Research Questions

This study sought to examine whether traditional mass media entities remain dominant as source influencers in the independent political blogosphere It is argued that blogs can restore the balance of power between the average citizen and traditional media entities (Armstrong & Moulitsas Zuniga, 2006; Reynolds, 2006). This study sought palpable evidence for such power shifts through assessing source influence in the construction of news reports. This study advanced the following research question.

  • RQ1:To what extent do independent bloggers of diverse political ideologies depend on citizen versus traditional media?Prior agenda setting studies have highlighted that issue characteristics play a strong role in determining the extent of traditional media's agenda setting influence. To date, few to no studies have tested this finding in the blogosphere. This study took the first step by questioning whether the different issues selected in this study would result in variable dependence on traditional to citizen media within independent political blog networks. This study advanced the following research question:
  • RQ2:Is there evidence that the balance of mass media to citizen media influence differs among independent bloggers of diverse political ideologies based on different types of issues?Prior studies have found that elite traditional media entities influence less elite media outlets (Reese & Danielian, 1989). Given the fact that there is unlimited media availability on the Web (Anderson, 2005), this study questioned the agenda setting influence of the New York Times and the Washington Post within independent political blog environments. This study advanced the following research question:
  • RQ3:To what extent do the New York Times and the Washington Post maintain their elite status as source influencers among independent bloggers of diverse political ideologies?Is the traditional newsroom utilizing its blog to forge connections to the citizen media world? Studies on the early adoption of the blog form within traditional media newsrooms found a tendency for insular linking practices among traditional media entities (Singer, 2005). With little research conducted on traditional media blogs since this initial study, this study advanced the following research question:
  • RQ4:To what extent do the New York Times and the Washington Post depend on citizen media versus traditional media in their political blogs?These four research questions were tested against a blog sample of 18 independent political blogs (six left-leaning, six moderate-leaning, and six right-leaning) and 11 political blogs culled from elite traditional media entities the New York Times' and the Washington Post.


To extract the 18 most popular U.S. political blogs from across the ideological spectrum, this study used correlation techniques to cull six popular blogs from the left-leaning blogosphere and six popular blogs from the right-leaning blogosphere from noted blog aggregators Technorati, the Truth Laid Bear, and BlogPulse. All of these aggregators maintain top or popular blogs lists based primarily on link and traffic rankings. Though the relative rank of individual blogs were different, correlation techniques yielded broad agreement on the 12 most popular blogs from the left-leaning and the right-leaning blogosphere. These blogs comprised the following: The Daily Kos (left-leaning), Crooks and Liars (left-leaning), Think Progress (left-leaning), Talking Points Memo (left-leaning), The Huffington Post (left-leaning), FireDogLake (left-leaning), Instapundit (right-leaning), Michelle Malkin (right-leaning), Hot Air (right-leaning), Little Green Footballs (right-leaning), Powerline (right-leaning) and Captain's Quarters (right-leaning).

Moderate blogs rarely make it to these popular lists. In order to select the top moderate blogs, the following techniques were adopted: a. presence in the three aforementioned aggregators; b. popularity in the ‘Moderate Blog Advertising Network,’ a network listing the most popular moderate blogs and the dollar amount to advertise on each of these blogs; c. nomination in the Weblog Awards, a popular blog contest running since 2003 that determines popularity through net nominations; and d. frequency of appearance in blogroll listings of the aforementioned blogs. The following sample of moderate blogs was identified for study: Donklephant, The Moderate Voice, The Daily Dish, The Gun Toting Liberal, Central Sanity, and The Van Der Galien Gazette.

This study also examined the political blogs of the New York Times and the Washington Post. Included in this study were the political blogs The Caucus, The Lede, and the Opinionator from the New York Times, and the political blogs White House Watch, Bench Conference, Achenblog, Early Warning, Think Tank Town, The Fix, On Balance, and OFF/Beat from the Washington Post.

The period July 20, 2007 to September 30, 2007 was selected as a good time period for relevant issue identification due to the preponderance of significant political events occurring in this time period. Initially, a catalog of all central issues was generated. The list of all potential issues was eventually whittled to reveal core issues based on two main factors: 1. Each issue needed to be bound by a definite start and end point, a factor important in measuring the nature of agenda setting effects, and 2. Issues also needed to generate notable discussion across all blogs and media. Using this methodology, only three issues provided enough traction on citizen media and mass media blogs. These issues were the Alberto Gonzales hearings on the NSA wiretapping, the Larry Craig sex scandal, and Iraq, the latter focusing on the Petraeus report and the MoveOn ‘Betray us' ad.

To locate the “bursty period” for each issue (beginning and ending time point of issue investigation), the frequency count of blog posts and articles dedicated to each of the three aforementioned issues was created on a day-by-day basis for the entire time period under study. Applying this rubric to each of the issues, the three start- and end-time periods for the Gonzales NSA wiretapping testimony, the Larry Craig scandal, and the Iraq issue were July 24 to August 2, August 27 to September 7, and August 15 to September 24 respectively.

The diversity of these selected issues also provided a good test of agenda setting effects. Both the Alberto Gonzales and the Larry Craig issue were dramatic and conflict-laden, capable of supporting partisan dialogue. In the case of the Craig issue, the incident appeared sudden and unexpected, suggesting that traditional media had a strong opportunity to set the blogosphere's agenda. Unlike the Craig scandal, the Gonzales hearing was an expected event, and was thus discussed by both media and the blogosphere before the actual hearing. In the case of the Iraq issue, the delivery of the Petraeus report was connected to the larger issue of the War in Iraq. The public's long period of deliberation about the report (evidenced by the longevity of this issue period in comparison to the other two issues), and the coincidental creation of a partisan ad by the democratic partisan organization MoveOn both suggested that this issue would be more resistant to traditional media agenda setting influence.

To decipher source influence, URLs were coded in situ on the blog posts of both the 18 independent political blogs and the 11 traditional media blogs. Each URL could only be placed in one content category, and content categories for this study's URLs included the following: liberal blog, conservative blog, moderate/neutral blog, democratic partisan organization, conservative partisan organization, nonpartisan organization, traditional media (newspapers, television, radio, newsmagazines), traditional media blog, periodicals other (nonhard news), niche/independent media, other citizen media (including video-sharing sites), government site, education site, media conglomerate (eg. links to News Corporation, Time Warner, Comcast, or Walt Disney), portal news site (eg. Yahoo, Google, Topix), AP/wire syndicated content, social news aggregators (e.g. Digg, Reddit, Buzz, Technorati), and web-only news media (e.g. Salon).

To clarify this study's URL rubric, textual elaborations with comprehensive site-based examples were utilized during intercoder reliability sessions. At the data analysis stage, categories were eventually collapsed to provide macroanalysis of citizen-to-traditional media links. This collapsing avoided the error of double counting since URLs could only be placed in one category. At this data analysis stage, the following main URL categories were aggregated to represent links to traditional media entities: traditional media (newspapers, television, radio, and newsmagazines), traditional media blog, and AP/wire syndicated content. For citizen media entities, the URL categories liberal blog, conservative blog, moderate blog, other citizen media (including video-sharing sites) and social news aggregators (e.g. Digg, Reddit, Buzz, Technorati) were aggregated. URL links to organizational entities (partisan or nonpartisan), media conglomerates, governmental sites, educational sites, niche media entities, and nontraditional media entities (portal news sites and web-only news media entities) remained as standalone categories, operating outside this study's main comparative focus on traditional media and citizen media source citations.

In sum total, 3721 URLs were analyzed for their social network connections, with 3172 culled from the 18 blogs and 549 culled from the 11 traditional media blogs. Using Krippendorf's alpha, intercoder reliability tests on a random sample of 317 URLs produced estimates ranging from .90 to .98 through the different blog networks. URLs were then coded through a Web form and captured in a database. SQL queries were written against the database to provide data on frequency of URL linkages. These queries also provided numeric counts of linkage patterns against which descriptive and inferential statistics were run.

Results: General Statistics

Table 1 provides a summary of link frequency through the three different issue periods across all four media networks (left-leaning blogosphere, right-leaning blogosphere, moderate blogosphere, and traditional media blog networks). As the table highlights, link volume was more prolific on the Iraq issue primarily due to the longevity of the issue time period when compared to that of the Larry Craig issue and the Gonzales issue.

Table 1.  Link Summary Across Media Networks and Issue Periods
Media NetworkLinks on Issues
Traditional Media10653390

It is important to note that the left-leaning blogosphere was significantly more likely to link when compared to all other media networks. As Table 2 highlights, a one-way ANOVA revealed that link frequency varied significantly as a function of partisan orientation for the blogs in this study (f(2,15) = 15.47, p < .05) when equal variances are assumed. Tukey posthoc tests revealed that left-leaning blogs in this study were significantly more likely to link across all issues (M = 341, SD = 113.96) when compared to right-leaning blogs (M = 106.5, SD = 81.39) and moderate-leaning blogs (M = 81.33, SD = 64.97) in this study.

Table 2.  Total Hyperlinks Across the Different Blog Networks
 SSDfMean SquareFSig
Between Groups245771.4442122885.72215.469.000
Within Groups119157.667157943.844  

This aggregate finding was borne through the different issues. A one-way ANOVA revealed that link frequency varied significantly as a function across the Craig issue (f(2,15) = 6.10, p < .05), the Gonzales issue (f(2,15) = 9.617, p < .05), and the Iraq issue among blogs in this study (f(2,15) = 10.919, p < .05), with equal variances assumed in all scenarios. Left-leaning blogs in this study were significantly more likely to link about Craig (M = 62.5, SD = 21.44), Gonzales (M = 44.5, SD = 25.02), and the Iraq issue (M = 233.83, SD = 100.23) when compared to right-leaning blogs on Craig (M = 22, SD = 17.24), Gonzales (M = 6.33, SD = 8.41), and Iraq (M = 78.17, SD = 67.70), and moderate-leaning blogs on Craig (M = 26.33. SD = 26.89), Gonzales (M = 9.83, SD = 11.67), and Iraq (M = 45.17, SD = 45.91).

Figure 1 provides evidence of the power law structure for all URL domains in this study. The 29 blogs examined in this study yielded 3721 unique links and 646 unique domains ( is a domain in contrast to, which counts as a link). The top 20% of the unique 646 URL domains in this study command 2890 of the 3721 links, or 78% of attention across all networks. This network's close adherence to the 80/20 Pareto power law suggests that a few elite actors are in control of the majority of source influence throughout the entire network of traditional media and citizen media blog links.

Citizen to Traditional Media Links in Independent Blog Networks

Research question one probed whether there were differences in the traditional-to-citizen media links through the entire blogosphere network and through each separate network within the citizen media blogosphere. Table 3 presents the top 30 domains in the blogosphere and traditional media within the independent blog networks. Over 50% of the top domains point to traditional media entities while approximately 33% point to citizen media. The popularity of the Washington Post and the New York Times is evidenced by their appearance as the two most-linked-to sites.

Table 3.  Top Hyperlink Domains Across the Entire Network
URL DomainDomain Frequency
  1. *bolded sites represent citizen media


Is there a preference among all citizen blog networks for either traditional or citizen media? Table 4 provides the means and standard deviations for the three networks in their links to citizen media versus traditional media across all three issues. Table 5 reveals the results of an independent samples t-test, which revealed no significant differences (t(34) = −1.49, p > .05) in links to citizen media (M = 58.61, SD = 52.72) or traditional media (M = 92.5, SD = 80.8) across all three issues in the three ideological blog networks.

Table 4.  Means and Standard Deviations of Links to Citizen vs. Traditional Media for All Blog Networks Across All Issue Periods
 NMeanStandard DeviationStandard Error From Mean
Citizen Media1858.652.7312.42
Traditional Media1892.580.819.04
Table 5.  Independent-Samples T-Test of Differences in Links to Citizen Media vs. Traditional Media Across All Blog Networks and Issue Periods
Equal Variances Assumed4.415.04−1.4934p > .05

Interestingly, this finding was not a robust one on an issue-by-issue basis through the three blog networks, providing an answer to research question two. On the Iraq issue, blog networks were equally as likely to link (t(34) = −1.442, p > .05) to citizen media (M = 38.11, SD = 35.26) as to traditional media (M = 62.9, SD = 63.74). Similarly, on the Gonzales issue, citizen media (M = 9, SD = 12.4) was equally as likely to be linked to (t(34) = −.177, p > .05) as traditional media (M = 9.44, SD = 10.27) from all blog networks. However, through all three networks, bloggers were significantly more likely to link to traditional media (M = 20.17, SD = 14.7) as compared to citizen media (M = 11.5, SD = 9.5) for the Larry Craig issue (t(34) = −2.044, p < .05).

Was the above finding supported through the individual blog networks? Left-leaning bloggers showed no significant differences in their links to citizen media and traditional media through two of the three issues. For the Iraq issue, both citizen media (M = 73.6, SD = 35.4) and traditional media (M = 118.5, SD = 68.3) were equally as likely (t(10) = −1.44, p > .05) to be linked to from the left-leaning blogosphere. Similarly, on the Gonzales issue, left-leaning bloggers were equally as likely (t(10) = .110, p > .05) to link to citizen media (M = 19.5, SD = 11.2) as traditional media (M = 20.3, SD = 14.8). However, on the Craig issue, left-leaning blogs were significantly more likely (t(10) = −2.27, p < .05) to link to traditional media (M = 31.5, SD = 9.64) versus citizen media (t(10) = 18.33, SD = 9.68).

For the right-leaning and the moderate network, no significant differences were found in links to citizen media and traditional media across the three issues under consideration in this study. The conservative blogosphere were equally as likely (t(10) = −1.12, p > .05) to link to citizen media (M = 31.83, SD = 21.6) as to traditional media (M = 61.2, SD = 62.5) on the Craig issue. There were no significant differences (t(10) = −1.1, p > .05) in links to citizen media (M = 7.2, SD = 6.49) and traditional media (M = 13.3, SD = 12.4) on the Gonzales issue. On the Iraq issue, both citizen media (M = 23, SD = 17.4) and traditional media (M = 45.67, SD = 52.5) stood an equal chance (t(10) = −1.04, p > .05) of being linked from the conservative blogosphere.

Like the conservative blogosphere, no significant differences were found in links to citizen media and traditional media in the moderate blogosphere across all three issues. In relation to the Craig issue, the moderate blogosphere were equally as likely (t(10) = −.963, p > .05) to link to citizen media (M = 8.5, SD = 8.9) as they were to traditional media (M = 15.67, SD = 15.89). On the Gonzales issue, there were no significant differences (t(10) = .093, p > .05) between the moderate blogosphere's links to citizen media (M = 5, SD = 7.4) and traditional media (M = 4.67, SD = 4.67). No significant differences (t(10) = −.483, p > .05) were also found for links between citizen media (M = 18.2, SD = 21.7) and traditional media (M = 24.5, SD = 23.7) on the Iraq issue.

Probing Traditional Media Bias

Research question three questioned the extent to which the New York Times and the Washington Post maintained their elite status as agenda setters within the independent, political blog networks. Tables 6 through 8 present the top 20 links across the different ideological spectrums on an issue-by-issue basis. Notable about these lists is the continued strength of traditional media, a finding that is well supported through

Table 6.  Top 20 Links for the Left-Leaning Blogosphere On An Issue-by-Issue Basis
RankCraigGonzalesIraq (28) (35) (141) (20) (35) (94) (13) (25) (51) (13) (11) (40) (13) (9) (36) (9) (8) (31) (9) (7) (27) (9) (7) (26) (8) (7) (25) (8) (6) (24) (7) (6) (22) (7) (5) (21) (7) (5) (21) (7) (4) (19) (6) (4) (18) (6) (4) (18) (6) (4) (18) (6) (4) (17) (5) (3) (16) (4) (3) (16)
Table 7.  Top 20 Links for the Moderate Blogosphere On An Issue-by-Issue Basis
RankCraigGonzalesIraq (28) (35) (141) (20) (35) (94) (13) (25) (51) (13) (11) (40) (13) (9) (36) (9) (8) (31) (9) (7) (27) (9) (7) (26) (8) (7) (25) (8) (6) (24) (7) (6) (22) (7) (5) (21) (7) (5) (21) (7) (4) (19) (6) (4) (18) (6) (4) (18) (6) (4) (18) (6) (4) (17) (5) (3) (16) (4) (3) (16)
Table 8.  Top 20 Links for the Right-Leaning Blogosphere on an Issue-by-Issue Basis
RankCraigGonzalesIraq (8) (8) (30) (7) (3) (25) (7) (3) (24) (6) (2) (15) (5) (2) (14) (5) (2) (14) (5) (2) (12) (4) (1) (10) (4) (1) (9) (4) (1) (9) (4) (1) (8) (4) (1) (8) (4) (1) (7) (3) (1) (6) (3) (1) (6) (3) (1) (6) (3) (1) (6) (2) (1) (5) (2) (1) (5) (2) (1) (5)

each of the three issues. For two of the three issues, the Washington Post and the New York Times are present in the top three media choices for all three networks. Notable about the right-leaning blogosphere is the presence of the National Review, a media selection which was their top media choice for the Iraq issue in this study.

Do any of the independent blog networks privilege either the status of the Washington Post or the New York Times? Table 9 provides means and standard deviations of all three networks in their links to the New York Times and the Washington Post across all three issues. Table 10 reveals the results of an independent samples t-test, which showed no significant differences in links to the New York Times (M = 14.56, SD = 15.73) or the Washington Post (M = 10.5, SD = 10.82) across all issues (t(34) = .901, p > .05) in the three blog networks. On an issue-by-issue basis, this finding was also strongly supported. There were no significant differences in links to the Washington Post (M = 1.5, SD = 2.09) or the New York Times (M = 1.39, SD = 1.5) on the Craig issue (t(34) = .183, p > .05) in the three blog networks. Bloggers were equally as likely to link to the Washington Post (M = 1.5, SD = 2.09) or the New York Times (M = 1.05, SD = 1) on the Gonzales issue (t(34) = .813, p > .05) in the three blog networks. Similarly, links were equally as likely to be given to the Washington Post (M = 2.89, SD = 3.66) or the New York Times (M = 1.61, SD = 2.40) for the Iraq issue (t(34) = 1.238, p > .05) amongst the three blog networks.

Table 9.  Means and Standard Deviations of Links to the New York Times and the Washington Post Across All Blog Networks and Issue Periods
 NMeanStandard DeviationStandard Error From Mean
Washington Post1814.5615.733.71
New York Times1810.510.822.55
Table 10.  Independent-Samples T-Test Of Differences in Links to the New York Times and the Washington Post Across All Blog Networks and Issue Periods
Equal Variances Assumed1.7.201.90134p > .05

How does this finding hold up through each blog network? Across all issues, there were no significant differences (t(10) = 1.255, p > .05) in the left-leaning's blogosphere's links to the New York Times (M = 31.5, SD = 16.56) or the Washington Post (M = 21, SD = 12.1). No significant differences were found in these left-leaning bloggers' links to the Washington Post (M = 2.17, SD = 1.72) or the New York Times (M = 1.167, SD = .75) on the Craig issue (t(10) = 1.303, p > .05). Both the Washington Post (M = 5.83, SD = 4.96) and the New York Times (M = 4.17, SD = 2.63) were equally as likely to get links from left-leaning bloggers on the Gonzales issue (t(10) = .727, p > .05); similarly the Washington Post (M = 23.5, SD = 11.13) and the New York Times (M = 15.67, SD = 10.76) were similarly linked to on the Iraq issue (t(10) = 1.249, p > .05).

The conservative blogosphere produced similar results: There were no significant differences between the Washington Post (M = 5.83, SD = 4.75) or the New York Times (M = 5.17, SD = 5.63) in link totals across all issues (t(10) = .222, p > .05). No significant differences were found in right-leaning bloggers' links to the Washington Post (M = .33, SD = .52) or the New York Times (M = .66, SD = .82) on the Craig issue (t(10) = −.845, p > .05). Both the Washington Post (M = 1.33, SD = 1.75) and the New York Times (M = .5, SD = .84) were equally as likely to be linked to on the Gonzales issue (t(10) = 1.052, p > .05), and no significant differences were found between link totals for the Washington Post (M = 4.12, SD = 4.87) or the New York Times (M = 4.0, SD = 5.59) on the Iraq issue (t(10) = .055, p > .05).

Like the other two networks, the moderate blogosphere bore similar findings. The moderate blogosphere had no significant differences (t(10) = .349, p > .05) in its media links to the Washington Post (M = 6.3, SD = 5.3) or the New York Times (M = 5.3, SD = 4.68) across all issues. Both the Washington Post (M = 2, SD = 3.03) and the New York Times (M = 1.3, SD = 1.37) were equally as likely to be linked to on the Craig issue (t(10) = .491, p > .05). On the Gonzales issue, both the Washington Post (M = 1.5, SD = 1.5) and the New York Times (M = .17, SD = .41) held similar prestige (t(10) = 2.08, p > .05). Similarly, on the Iraq issue, both the Washington Post and the New York Times were equally as likely to be linked to from the moderate blogosphere.

Citizen Media Impact in Traditional Media Blog Networks

Research question 4 probed whether elite traditional media outlets were open to the intermedia agenda setting influence of citizen media on their blog news reports. Evidence of such hyperlink influence would indicate that elite traditional media entities are using their blog to produce a more participatory form of journalism.

Table 11 presents data on the blog linking practices of both media entities across all three issue periods under consideration in this study. As the data highlight, links to other traditional media entities still account for the lion's share of these elite blog links, with only four of the top 20 sites representing citizen media. Two of these four citizen media sites are to left-leaning blogs in this sample (Talking Points Memo and Huffington Post). Interestingly, both of these left-leaning blogs were founded by bloggers who were once traditional, professional journalists.

Table 11.  Top Media Blog Domains for the New York Times and the Washington Post

Across all issues, both the New York Times and the Washington Post dedicate the majority of their links to other traditional media entities (see tables 12 and 13). The data also reveal that for each elite entity, its top source is itself with its second top source being the other. The cohesion of the traditional media social network is evidenced by the fact that the second highest linked-to source for the New York Times is the Washington Post and vice versa. Like the aggregate data for both media blogs, there are few blogs in the top 20 for each individual traditional media entity. The Washington Post has four citizen media outlets in its top 20, with two of these representing left-leaning blog authored by individuals who were once journalists. The New York Times has three citizen media sources in its top 20 sources. One of the two blogs in the top 20 links for the New York Times is to the right-leaning blog Captain Quarter's. The other is to Robert Stein, former chairman of the American Society of Magazine Editors and well-known media critic.

Table 12.  Top Domains for the New York Times' Blogs
  1. *bolded sites represent citizen media

Table 13.  Top Domains for the Washington Post Blogs
  1. *bolded sites represent citizen media



Addressing the applicability of agenda setting theory to the new media environment, McCombs (2005) noted that, “whether the basic agenda setting effects of news media continue in much the same fashion as the previous decades or eventually disappear because of the changing media landscape, measuring these effects will remain high on the research agenda for at least the near term” (pp. 546). Blogs have brought the press to the people (Rosen, 2006) and have been lauded for their ability to restore the balance of power between citizens and traditional media in the creation of citizen news reports (Armstrong & Moutlisas Zuniga, 2006; Reynolds, 2006). Agenda setting theory, formulated during the time when traditional media had monopoly power over the tools of content creation and distribution (Lippmann, 1949; McCombs, 2004), allowed traditional media to be the central gatekeeper over the passive news reading public. With the growth of more interactive Web applications and the adoption of the blog by traditional newsrooms, this study sought to examine the role of citizen and traditional media in the setting of news agendas. This study also sought to examine whether social influence theory could predict agenda setting influences among and between these two media players.

This study's findings highlight that traditional media's agenda setting power is no longer universal or singular within citizen media outlets: The independent blog platform is redistributing power between traditional media and citizen media. Traditional media agenda setting is now just one force among many competing influences. Unlike traditional media platforms, independent blog networks are utilizing the blog tool to allow citizens more influence and power in setting news agendas. Across all blog networks, there were insignificant differences in traditional-to-citizen media links across all three issue periods combined (t(34) = −1.49, p > .05). In two of the three networks (right leaning network and moderate network), this finding was supported on an issue-by-issue basis.

Traditional media's loss of agenda setting monopoly power in the political blogosphere can be well explained by shared status homophily. The independent political bloggers' outside status from “beltway” or “mainstream” media has afforded them greater freedom to utilize other citizen media sources when building critiques of traditional media's news reports (Meraz, 2008). Unlike traditional media outlets that must rely on bureaucratic, routine sources in their effort to produce reliable, credible, and predictable journalism (Gans, 1980; Shoemaker & Reese, 1991), independent bloggers are bound by no such codes. Citizen media gain their strength to critique traditional media's news reports through the dynamic, real-time assemblage of relevant perspectives and opinions shared by other citizen media outlets.

How much power does citizen media have in diluting traditional media's singular agenda setting influence? The New York Times and the Washington Post maintained relative dominance as top sources within each blog network's top 20 media choices for two of the three issues under investigation in this study. In terms of agenda setting theory, elite traditional mass media entities are more likely to exert their agenda setting power at the “short head” of the long tail of media choices while citizen media influence aggregates agenda setting power down the “long tail” of media options (see Figure 1). Such a finding suggests that though traditional media's agenda setting power is no longer the sole influence, its influence still remains a driving, “A-list” force in the creation of blog agendas. As predicted by long tail media theory (Anderson, 2005), citizen media's efficacy is in its aggregate effect, an effect which is able to blunt traditional media's singular agenda setting effect.

Of particular interest was the inefficacy of issue characteristics in determining media agenda setting power in conservative and moderate blog networks. Closer inspection of this study's three selected issues highlights a negative tilt towards the conservative agenda. The impact of this negative tilt was evident in right-leaning blogosphere's low hyperlinking volume through all issue periods. This finding provides another window into how agenda setting and gatekeeping operate in the new media environment. The conservative bloggers' lower hyperlink volume suggests that these bloggers made a conscious choice to ignore media agenda-setting power. In the new media landscape where hyperlinks fuel story popularity and saliency, silence on issues can be a powerful strategy to circumvent media agenda setting power and ultimately, story popularity on Web portals, search engines, and news aggregators.

As predicted by social influence theories, elite traditional media entities have hijacked the blog form, a tool designed for outward, networked conversations, to maintain internal elite, conversations within their network neighborhood of other trusted, traditional media entities. Interestingly, for both the New York Times and the Washington Post, the second most-linked-to source is the other, and for both media entities, traditional mass media predominate in their hyperlink choices. The only independent, citizen media bloggers that are able to gain traditional media's attention were once journalists or maintain traditional media ties as in the case of the blogs Talking Points Memo and the Huffington Post. It is apparent that for traditional media entities, agenda setting influence is filtered by the sociopolitical boundaries of the press pass or knowledge of journalism norms and traditions.

Declining newspaper circulation rates, eroding network television audiences, and declining credibility of traditional media news outlets among Web users all suggest that traditional media is in desperate need of reinvention. It is questionable how long traditional media can erect walled gardens in tools that build wealth through networked conversations and hyperlinked transparency. The empowered Web audience, familiar with the participatory freedom of independent blogging, may find little utility in a newsroom blog that ignores the wisdom of the crowd.

This study had several limitations. The blog sample was limited to 18 top independent political blogs and two elite traditional media entities in an effort to capture dominant influence. Future studies should expand the independent blog sample down the ‘long tail’ of media options and be more inclusive to both elite and nonelite traditional and citizen media entities. Greater work also needs to be done in replicating this study across diverse issue periods and time frames to further explicate the role of issue characteristics in mediating agenda setting influence between traditional media entities and citizen media entities.

Future work could attempt to parse traditional-media-to-citizen media influence across first-level and second-level agenda setting to examine whether traditional media function as a greater influence at the informational versus the interpretative level. Future studies can also isolate order of appearance of sources in an effort to determine whether such visual cues and markers in traditional journalism indicate relative agenda setting power in the networked, new media environment.

About the Author

Sharon Meraz is an Assistant Professor in the Department of Communication, University of Illinois, Chicago. Her research focuses on the theoretical and practical impact of networked, social media technologies on citizen political engagement, citizen journalism, digital democracy, and mass media evolution.

Address: 1007 W. Harrison, Behavioral Sciences Building MC 132, Chicago IL, 60607.