Stavrositu 2013: Online Political Discusson: Evidence from a Content Analysis of Blogs
Volume 23 Number 3, 2013
Online Political Discussion:
Evidence from a Content Analysis of Blogs
Carmen D. Stavrositu
University of Colorado
Colorado Springs, Colorado, USA
Abstract: Available research examining the role of the Internet in reshaping political talk, in general, and cross-ideological political discussion, in particular, provides conflicting evidence regarding the medium’s contribution to democratic health. The present study seeks to weigh in on this debate by providing insights from the under-studied realm of the political blogosphere via a quantitative content analysis of partisan political blogs. Results suggest that, in general, the political talk occurring in partisan political blogs is characterized by high levels of ideological agreement and minimal levels of disagreement (i.e., cross-ideological discussion). Further, when disagreement does occur, it appears to occur to a greater extent in conservative, rather than liberal blogs. Theoretical and practical implications are discussed.
The levels of political participation and deliberation in the U.S. are less than ideal (Eveland, Morey, & Hutchens, 2011; Mutz, 2006), despite the fact that such citizen engagement is crucial to democratic health. With the growing reliance on the Internet as both a broadcasting and an interpersonal medium, social commentators and scholars alike have turned their attention towards understanding the role that the Internet might play in reshaping fundamental democratic processes.
Noting that the Internet offers increased access to a variety of political information and perspectives, as well as more opportunities to discuss politics with ideologically-dissimilar others across platforms (e.g. discussion boards, news sites, social networking sites), a recent body of scholarship has examined the Internet’s potential contribution to invigorating online political talk in general, and cross-ideological political discussion in particular.
Despite discrepant findings and a vigorous debate on the matter, there is only limited insight available from examinations of political blogs. Centered around providing political information and commentary, and equipped with several feedback tools allowing for two-way communication, political blogs can serve as ideal venues for assessing naturally occurring political talk. Indeed, several studies suggest that thanks to their embedded feedback functions – the most notable of which is the commenting function – political blogs often serve as virtual communities or town hall meetings, where exchange of political ideas and opinions is made possible (McKenna & Pole, 2004; Wallsten, 2008). Moreover, as notoriously partisan online environments, political blogs make for a relatively straightforward setting for examining cross-ideological political discussion, that is, the extent to which exchanges of p
olitical ideas and opinions cut across lines of ideological difference (Mutz, 2006).
The primary aim of the present study is to determine the presence and extent of cross-ideological political discussion occurring in partisan political blogs. To this end, it employs a content analysis of liberal and conservative blogs by closely examining the conversational dynamics manifest in blog comments. The next section of the paper reviews relevant literature pertaining to online political discussion first, and to political discussion in the context of blogs next. After describing the content analytic methods and the study results, the paper discusses the obtained findings by highlighting their theoretical and practical underpinnings.
Political Discussion in the Internet Era
Online Political Discussion
Ever since its inception, the Internet has been hailed as a powerfully democratizing force in contemporary society. According to many, the Internet’s ability to afford users increased access to a wider diversity of political information and perspectives has been well positioned to invigorate political discussion, which is widely recognized as the backbone of a healthy and effective democracy (Gastil, 2008; Wojcieszak & Mutz, 2009).
While there is little disagreement regarding the Internet’s role in providing better access to a more diverse pool of political information and viewpoints (e.g., Hargittai, Gallo, & Kane, 2008; Iyengar, 2001; Sunstein, 2001), there is substantial debate over the presence and extent of cross-ideological political discussion in online environments. Defined here as political talk cutting across lines of ideological difference (Mutz, 2006), political discussion has been shown to serve in disseminating political information, stimulating participation, and informing voting behaviors (Muhlberger, 2004; Mutz, 2002; Scheufele, 2000). With so much at stake, and with online forums for interpersonal discussions becoming increasingly diversified and popular, it becomes impe
rative that scholars determine the ways in which the web may be reshaping this fundamental democratic process.
Some scholars argue that the Internet stifles political discussion by fostering the emergence of so-called “echo chambers” (Sunstein, 2001). That is, the Internet fosters environments conducive to the pursuit of agreeable information and interactions among like-minded individuals. These environments, Sunstein argues, are a direct result of the “filtering” affordances embedded in a wide array of web interfaces (from search engines to portals to blogs), and serve to exacerbate a well-documented human tendency towards selective exposure to information (i.e., the tendency to purposefully seek out information congruent with pre-existing attitudes and beliefs, and avoid conflicting perspectives). In line with this position, several studies have empirically documented evidence of selective exposure or ideological filtering in media use — conservatives and liberals alike have been shown to prefer ideologically congeni
al news when the news source is salient (Iyengar & Khan, 2010; Stroud, 2010), and attitude-consistent political messages when the news source is not salient (Knobloch-Westerwick & Meng, 2011). This trend, in turn, has been shown to severely limit cross-ideological political discussion and lead to attitude reinforcement (Stroud, 2010), polarization (Iyengar & Khan, 2010; Knobloch-Westerwick & Meng, 2011; Stroud, 2010), and extremity (Kim, 2009).
Other scholars suggest that the Internet is in fact very likely to revitalize democracy by stimulating, rather than stifling, political discussion. Most notably, a study conducted by Pew Internet Research as part of the American Life Project (Horrigan, Garrett & Resnick, 2004) has found that at least in an electoral context, Internet users appear to exhibit greater exposure to political information — both agreeable information, as well as information conflicting with their pre-existing attitudes and beliefs on key issues. Furthermore, Garrett (2009) has shown that while Internet users might actually favor opinion-reinforcing political information, they do not necessarily avoid conflicting information. Finally, some studies have found evidence for political conversation involving diverse viewpoints and perspectives among Usenet, real-time chat, and message board users, in
particular (Stromer-Galley, 2002; 2003), with most of these users reporting that they did not typically engage in face-to-face political discussion but relied on online venues for that purpose instead (Stromer-Galley, 2002).
In addition to the striking lack of consensus, the evidence supporting either position is still limited (Wojcieszak & Mutz, 2009), and quite possibly plagued by its reliance on self-report data. As Eveland et al. (2011) note, this reliance is problematic for several reasons. Most notably, measurement error stemming from the potential discrepancy between the researcher’s and respondents’ definitions of “political” is likely to impact the validity of the typical political discussion frequency measures assessed in such studies. In addition, such studies also rely on respondents’ unclear notions of what constitutes “discussion,” as well as their potential lack of awareness regarding the frequency and nature of discussion they engage in. These shortcomings call for taking advantage of publicly accessible Internet content archives and employing unobtrusive
methods of inquiry that allow us to observe and analyze naturally occurring political discussion groups (e.g., discussion boards, blogs). By allowing researchers to analyze available content for the presence and extent of political discussion consistent with their own conceptualization of “political discussion,” this methodological approach does not require reliance on possibly idiosyncratic notions of “political” or “discussion” across participants, which may be at odds with each other or with the researcher’s notions of the terms.
Blogging and Political Discussion
With this in mind, the present study attempts to contribute to the inconclusive body of research detailed above via a content analysis of political blogs. Political blogs are frequently updated websites centered around providing political information and commentary on issues ranging from the elections, to foreign policy, to the economy or healthcare, among others (Hargittai et al., 2008), and have garnered a lot of attention in the past decade for their influence on mainstream media, politics, and the public sphere (Dreznel & Farrell, 2008; Wallsten, 2007; Xenos, 2008). Noting their notoriously partisan nature — with political blogs often espousing an explicit political ideology, as well as the technological affordances that make blogs veritable forums for discussion and debate — via their embedded commenting funct
ions, trackbacks, and blogrolls, some have proclaimed political blogs as important emerging spaces for mediated political participation (e.g., Gil de Zúñiga, Puig-i-Abril, & Rojas, 2009) and deliberation alike (e.g., Xenos, 2008). As a result, political blogs have quickly become an ideal testing ground for the echo chamber vs. political discussion debate.
In this newer testing arena, one line of research has addressed the debate by examining the ideological selectivity (or lack thereof) in the coverage of political news in political blogs vis-à-vis mainstream media. These studies suggest that both conservative and liberal political blogs exhibit a strong tendency towards one-sided coverage of news based on their own ideological bent (e.g., Baum & Groeling, 2008; Xenos, 2008), at the expense of creating opportunities for shared reference points that cut across lines of ideological difference.
Another line of research has weighed in on the debate by taking a closer look at the linkage patterns in the political blogosphere. For example, studies have examined the linking dynamics within and between liberal and conservative blogs (e.g., Ackland, 2005; Adamic & Glance, 2005; Hargittai et al., 2008) — that is, the degree to which conservative and liberal blogs link to other blogs that are either on the same, or the opposite side of the ideological spectrum. Results of these studies indicate that both liberal and conservative blogs appear to link primarily to other ideologically similar blogs. Nevertheless, evidence for some degree of linking to blogs that are on the opposite side of the ideological spectrum emerged as well.
While both lines of research reported above appear to show greater support for the echo chamber position than for the opposite stance, it is worth noting that these studies focus on bloggers’ actions only — i.e., choice of news coverage and linking activity, while ignoring blog readers’ activity. Consequently, it can be argued that they provide more insight into whether bloggers create an environment that is conducive to political discussion in their blogs, and use that as a proxy for the actual presence and extent of political discussion taking place in the blogosphere. However, blogs are spaces that are only in part created by the bloggers. In fact, while bloggers indeed have full control over the structure of their blog and the content they themselves create, it is blog readers that play the most important role in the actual discussion that may emerge, via commenting. Recognizing the importance of blog readers in shaping blog discussion, a recent
study by Lawrence, Sides, and Farrell (2010) shifted focus away from the bloggers and placed it on blog readers. Despite the commendable focus on blog readers, nevertheless, the study by Lawrence et al. (2010) offers yet another proxy, rather than direct evidence, for the existence and extent of political discussion in the political blogosphere. More specifically, their study relies on readers’ self-reports regarding the types of blogs they most commonly read. While both conservatives and liberals were shown to read predominantly blogs that match their own ideological orientation, it is highly arguable whether this pattern alone is a clear testament to whether readers engage in political discussion or not.
Based on the above discussion, the present study attempts to bring a contribution to the echo-chambers vs.political discussion debate in three ways: (1) the present study focuses on blog readers and their blogging behavior; (2) this study seeks to provide direct evidence of political discussion by examining blog comments, as past research has called for (e.g., Hargittai et al., 2008); and (3) it does so unobtrusively via a content analysis of partisan political blogs, in order to avoid the types of concerns highlighted by Eveland et al. (2011). More specifically, the present study addressed two main research questions. First, what is the prevalence of political talk, in general, and cross-ideological discussion, in particular, in U.S.-based partisan political blogs? Second, what is the relationship between the political ideology espoused in U.S.-based partisan political blogs, the topics discussed,
and the extent of cross-ideological political discussion?
The two research questions were addressed via a content analysis of 174 partisan political blogs randomly selected from a publicly available political blog directory. Despite that some concerned scholars cautioned early on against the rapidly changing online content and the specific challenges posed to the process of data collection and coding (e.g., McMillan, 2000), blogs should not be prone to these threats given their typically open content archives. While new blog posts might change the overall content of a blog, the posts sampled for content analysis remain intact. For this reason, they are indeed comparable to traditional media for purposes of content analysis.
Partisan political blogs analyzed in this study were sampled from a publicly available blog directory listing blogs in several political/ideological categories (http://directory.etalkinghead.com/). The choice of blog directory was informed by the following considerations. First, in order to more accurately assess the political ideology espoused in a blog, the blog directory had to list blogs by ideological category (i.e., conservative blogs, liberal blogs). Second, the blog directory had to have a sufficient number of blogs listed in both ideological categories of interest. Third, the blog directory had to list blogs authored by “lay” individuals, rather than large media organizations with specialist staff and a mass audience (e.g., Huffington Post) or “A-list” blogs — highly prominent, well-linked and well-trafficked blogs (Adamic & Glance, 2005; Trammell & Keshelashvili, 2005) (e.g., Instapundit, Daily Kos).
At the time of the sampling, the number of conservative blogs listed was 156, while liberal blogs amounted to 167. In order to randomly sample from this pool, the study employed a web-based random number generator tool (www.random.org). More specifically, a list of 323 blogs (156 conservative and 167 liberal blogs) was compiled first. Next, using the integer generator function of the random number generator, the software was asked to generate 200 integers, ranging in value from 1 to 323. The emerging list of 200 blogs was then examined in order to ensure its suitability for the subsequent content analysis, as follows. Political blogs listed in the eTalkinghead directory were accompanied by a short description, which was used for coding blogs as conservative or liberal. For example, the description of one of the conservative blogs read “Converted blue state democrat who has found the bright shining light of Conservatism through an authentic and personal journey,&rdqu
o; while the description for a liberal blog stated “Political blog, left leaning, looking to point out the hypocrisy and unreasonableness of the Republican party and all politicians that put self ahead of country.” Using this procedure, coding of the ideological orientation of blogs was quite straightforward in most cases. However, some blogs had vague descriptions, despite being listed in either the “liberal” or the “conservative” category (e.g., the description for a blog listed in the “liberal blogs” category read “A political blog for the fair and balanced individual”). When faced with vague descriptions such as this, coders took a few additional steps. Upon opening the specific blog site, they first looked for unambiguous indicators of political affiliation — political party banners, buttons or banners for candidate endorsement (President, Governor, Senator, etc.). Coders also checked the “abo
ut the author” and the “blog mission” sections of each blog, looking for verbal statements regarding bloggers’ political allegiance. If none of these additional steps succeeded in clarifying the political ideology espoused in the blog, the blogs were discarded. Further, blogs that did not focus on politics were discarded — at times, bloggers appear to list their blogs in certain web directories simply to gain visibility without paying attention to the criteria for self-categorization, and so were blogs that appeared to be wrongly listed in a political ideology category (i.e., conservative or liberal). Lastly, blogs that did not focus primarily on U.S. politics or had not been updated in six months or longer were excluded from further analyses, as well.
The final sample consisted of 174 blogs (50% conservative and 50% liberal), and a complete list of these blogs is available from the author upon request. On average, blogs were updated 25.28 times per month (SD = 40.14), with a median of 15.50. 86.5% of the blogs were authored by “lay” bloggers, whereas 13.5% were authored by “professional” bloggers (i.e., journalists, academics, etc.). None of them was identified as an A-list blog.
Units of Analysis
The codebook contained variables and categories derived from multiple units of analysis: blog, post, comment, and blogroll. A partisan political blog was defined as a frequently updated website focused on covering political news and providing commentary on American politics. A post was defined as an entry authored and posted by the blog author. A comment was defined as any entry authored and posted by a reader or the author herself in reply to any of the blog owner’s posts or other readers’ comments. A blogroll was defined as a dedicated list of blogs the blog author recommends by linking. All 174 blogs included in the sample were coded for blog-, post-, comment- and blogroll-related variables, based on these four units of analysis.
Blog variables. The blog-level variables coded for pertained to the author profile (information about the author), author gender and location, author expertise (coded as personal — lay bloggers, or professional — academics, political commentators, journalists), political ideology espoused in a blog (conservative or liberal, as self-identified in the blog directory listings), frequency of updating (assessed by randomly selecting a month from the life span of the blog, then counting the number of posts written during that month), and the lifespan of the blog (the period from the month and year when blog was started to the most recent month and year).
Post variables. From the randomly selected month used to serve as an assessment of the blog’s frequency of updating, I further randomly selected one day and then one post per day (if multiple existed), in order to examine the following post-level variables: type of post (text, video or combination), post topic (economy, foreign policy, economy, or politics/elections) and length (number of words included in the post). Post topic was coded as focusing on economy when a post addressed issues related to the economic crisis, the budget, stimulus, etc.; it was coded as foreign policy when it addressed the war (e.g., Afghanistan, Iraq), diplomatic relations, issues related to national security, etc.; it was coded as social issues when it explored abortion, gay rights, civil rights, affirmative action, etc.; finally, it was coded as politics/elections when it dealt with broader political issues (e.g., the emergence and activity of the Tea Party, other political
parties), and more directly with the elections. Occasionally, multiple topics overlapped in the same blog post. In such cases, the topic was coded based on the dominant focus of the post. In sum, for each blog, I selected one month, then examined one post per month.
Blog posts were overwhelmingly text-based (97.2%), with a small minority being text + video (2.8%), and ranged from 10 to 3219 words in length (M = 569.96; SD = 589.20). Social issues were the most common post topic (34.9%), followed by politics/elections (31.4%), the economy (17.8%), and foreign policy (16%).
Comment variables. For each post, the number of comments received (i.e., number of entries posted in reply to the original post and/or other comments) was recorded first. In order to determine the level of cross-ideological discussion occurring in political blogs, disagreeing and agreeing comments were tallied next. Disagreeing comments were defined as “comments reflecting ideological disagreement with the original post (i.e., comment involves disagreement in perspective, that is, it is written from a different ideological perspective than the original post).” For example, in response to a conservative blog post favorably discussing American companies that have taken their operations to more tax friendly countries, a disagreeing comment read “For every blood and money sucking corporation out there that wants to leave our shores I say good riddance! You want to sell back to these shores it should be painful enough for them to think twice abou
t leaving. We should take our business to someone hungrier and more willing to remember that it is….’We the People’ not ‘ME – the individual.’ Your individual freedom rests upon the shoulders of many. It always has and always will. Support the SP-USA!” In contrast, agreeing comments were defined as “comments reflecting ideological agreement with the original post (i.e., comment involves agreement in perspective, that is, it is written from the same ideological perspective as the original post, even though it may involve disagreement with other ideologically antagonistic comments).” For example, in response to the disagreeing comment quoted above, one “agreeing comment” read “From the outset, it appears that your idea that “we don’t need them” is based on the assumption that we are so wealthy that we are needed by companies. It appears a very arrogant and close-minded position to think
that the 300 million people here are so essential to businesses that they dare not leave our shores. Many companies, like Philip Morris, make far more overseas than they do here already. PM makes more in Japan alone than in the states, and they maintain offices here more due to the costs involved in moving and the availability of tobacco than because of any dependency on the US consumer. I think you will find the US consumer not so important to companies that are large enough to go global. Good riddance, you say, but when all that is left to purchase is made in China, where will the corporate tax revenue come from?” This comment is in agreement with the original post, but in clear disagreement with the previously cited disagreeing comment.
In general, partisan blogs received an average of 4.84 comments (SD = 20.29) per post. A vast majority of blogs (84.7%) received 4 or fewer comments, with about half of all partisan political blogs examined barely involving any commenting at all (Md = .50), and the other half receiving comments ranging from .50 to 197.
Lastly, blogroll-level variables included blogroll size (number of other blogs listed on the main blog) and political leaning of blogs on the list. This was determined by counting the number of blogs for each of the following categories: conservative, liberal, ambiguous and other. Blogrolls in this study listed an average of 39.98 blogs (SD = 86.40), ranging from 0 to 700.
All coding was conducted by two coders trained in coding procedures and familiar with the code book. All blogs were independently coded by both coders so that intercoder reliability could be computed. This was done by employing the alpha macro developed by Hayes and Krippendorf (2007). Intercoder reliabilities ranged from α = .85 to α = 1.00.
The first research question pertained to the prevalence of political talk, in general, and cross-ideological political discussion, in particular in partisan political blogs. In order to address this question, a first step involved taking a close look at the amount of comments recorded in partisan blogs (as an indicator of general political talk), while a second step entailed examining the nature of comments in the political blogosphere (as an indicator of cross-ideological political discussion). In general, partisan blogs received an average of 4.98 comments (SD = 20.64) per post (MAgree = 4.07, SD = 18.93; MDisagree = .91, SD = 3.64), with a median of .50. Because all comment-related variables (e.g., overall number of comments, agreeing/disagreeing comments) were severely skewed and could not be meaningfully normalized, all subsequent analyses involving these variables employed no
nparametric statistical procedures.
In the first step addressing the first research question, a Mann-Whitney U-test — the nonparametric equivalent of independent-sample t-tests for non-normal data was conducted first, in order to determine whether the prevalence of comments is significantly different across liberal and conservative blogs. This analysis suggests that the amount of comments received on average was similar for liberal (M rank = 84.72, n = 86) and conservative blogs (M rank = 84.27, n = 82), z(168) = -.07, p = .95, indicating that political blogs exhibit similar levels of political talk, regardless of their ideological endorsements. Further, a Kruskal-Wallis H-test (nonparametric alternative to one-way ANOVA) was conducted to examine whether post topic dictated in any way the number of comments received in partisan political blogs. This analysis revealed no significant differences in the overall number of comments received as a fu
nction of post topic, with blogs focusing on social issues, politics/elections, economy or foreign policy eliciting about the same number of comments, H (3, N =163) = 1.33, p = .72. Together, this evidence points to moderate general levels of political talk across liberal and conservative blogs and different post topics.
In the second step addressing the first research question, a Wilcoxon rank signed test, the nonparametric equivalent of a paired-sample t-test, was conducted to compare the prevalence of agreeing vs. disagreeing comments in political blogs. This analysis revealed that partisan political blogs, regardless of espoused ideology or topic discussed, attracted significantly more comments in agreement (M rank = 46.78) rather than disagreement (M rank = 37.39), z = -5.42, p < .001. This particular finding points to generally limited levels of cross-ideological political discussion in the blogosphere, by showing that political talk in partisan political blogs entails in large part exchanges of political ideas and opinions from the same, rather than opposing, ideological perspective. The figures below display the proportions of agreeing and disagreeing comments in relation to the overall amount of comments in partisan political blogs across p
olitical ideology (Figure 1) and topic discussed (Figure 2).
Proportion of Agreeing and Disagreeing Comments Across Political Ideology
Proportion of Agreeing and Disagreeing Comments Across Post Topic
The second research question asked if the political ideology espoused in partisan blogs and the topics discussed had any influence on cross-ideological political discussion. That is, do liberal vs. conservative blog posts, or different post topics attract significantly distinct levels of disagreeing comments? The effects of political ideology and post topic on the level of ideological disagreement were determined by way of a Mann-Whitney U-test and a Kruskal-Wallis H-test, respectively. The first analysis revealed an effect for political ideology approaching significance, indicating that the number of disagreeing comments received by blog posts may depend on the political ideology espoused in the blog, with conservative blogs receiving a significantly higher number of disagreeing comments (M rank = 88.71, n = 82) than liberal blogs (M rank = 80.42, n = 86), z(168) = -1.59, p = .10 (see Figure 3
). The second analysis showed that while posts discussing social issues attracted the highest number of disagreeing comments (M rank = 88.72, n = 58), followed by posts on the economy (M rank = 84.71, n = 29), then by posts on foreign policy (M rank = 78.84, n = 25), and finally by those on politics/ elections (M rank = 74.37, n = 51), these differences were not statistically significant, H (3, N = 163) = 5.51, p = .13.
z(168) = -1.59, p = .10
Cross-ideological Discussion as a Function of Political Ideology
Lastly, additional analyses were conducted to examine what other factors might be related to the occurrence of cross-ideological discussion in partisan political blogs. To this end, a series of Spearman rho correlations (non-parametric alternative to Pearson correlation) was conducted to assess the potential relationships between the prevalence of disagreeing blog comments and frequency of blog updating, blog post length, and blogroll size. There was no correlation in the data set between prevalence of disagreeing comments in partisan blogs and either the frequency of blog updating or the length of the blog posts (Spearman rs = -.03 and .07 respectively, ns). However, prevalence of disagreeing comments was positively correlated with blogroll size (Spearman rs = .15, n = 168, p = .05).
In sum, findings indicate that conservative and liberal blogs exhibit similar levels of political talk in general, consisting primarily of ideological agreement. At the same time, conservative and liberal blogs differ to some extent in the levels of cross-ideological discussion. That is, conservative blogs appear to host more political discussion stemming from ideological disagreement than liberal blogs. Post topic had no influence on either the overall level of political talk, or the level of cross-ideological discussion.
The present study represents an effort to contribute to extant political discussion research by investigating naturally occurring political discussion in partisan political blogs. More specifically, a content analysis of partisan blogs was performed in order to assess the presence and extent of political talk in general, and cross-ideological political discussion in particular, by closely examining the commenting activity on partisan political blogs.
In general, the findings reported here offer more direct support for the echo chamber hypothesis with regards to the political blogosphere than other previous studies relying on proxy indicators of political discussion (e.g., Adamic & Glance, 2005; Hargittai et al., 2008; Wallsten, 2007; Xenos, 2008. The main findings of this content analysis suggest that, overall, there are at best moderate levels of political talk taking place in non A-list partisan political blogs, characterized by high levels of agreement and minimal levels of disagreement.
This pattern of results presents a series of implications for democracy, and two of its normative outcomes – participation and deliberation. Participatory democracy is centered around direct political action and discussion among those who share congruent political views and have similar political goals, as opposed to deliberative democracy, centered around discussion among people who have conflicting ideological beliefs and different political goals (Mutz, 2006). Unfortunately, the present findings echo previous scholars’ concerns regarding the tension between participatory and deliberative democratic ideals, as well as the difficulty of achieving multiple democratic ideals at the same time. Political discussion research has spotlighted this difficulty in both traditional networks (e.g., Mutz, 2006) and online environments, such as blogging (e.g., Lawrence, Sides & Farrell
While citizens’ political participation may differ across online and offline settings in their tactics and mechanisms, blogs and their ensuing conversations have been recognized in recent scholarship as important online participatory tools (e.g., Gil de Zúñiga, Puig-i-Abril, & Rojas, 2009). The levels of agreeable political conversation present in partisan political blogs, as documented in this content analysis, corroborate the notion that blogs offer Internet users a viable venue for online political participation and suggest that, at least by the normative standard of political participation, the non-A-list political blogosphere is fairly adequate. The vast majority of the political talk in partisan political blogs, conservative and liberal alike, appears to consist of conversations among like-minded individuals.
On the other hand, the pattern of results revealed in this study seems to indicate that by the normative standard of political deliberation, the political blogosphere is rather lacking. The minimal levels of cross-ideological disagreement uncovered here provide further support for Sunstein’s (2001) notion of “echo-chambers,” undermining the very nature of deliberative democracy. Political talk in partisan political blogs appears to function primarily as a safe nest for “birds of a feather,” while failing as a forum for expressing political difference and diversity.
Interestingly, when disagreement does take place in partisan blogs, it is more likely to occur in conservative blogs rather than in liberal blogs. This may be indicative of liberals’ greater willingness to engage with those holding opposing political views, hence the higher prevalence of comments stemming from ideological differences in conservative blogs. In fact, some evidence exists to support this possibility. More specifically, this evidence suggests that tolerance is a chronically accessible characteristic in liberals (Greenberg, Simon, Pyszczynski, Solomon, & Chatel, 1992). In the present context, this particular finding may translate to liberals being less critical of ideologically dissimilar others, and more likely to engage them in political discussion and debates. At the same time, it is possible that conservative bloggers are more successful at structuring their blog in such a way that it becomes conducive to cro
ss-ideological exchanges than liberal bloggers. While documenting blog structure was beyond the scope of the present study, previous studies show that at the very least, conservative bloggers seem to link more densely to other blogs, including liberal blogs (e.g., Adamic & Glance, 2005; Hargittai et al., 2008), possibly inviting more traffic from ideologically dissimilar others as well as encouraging more discussion involving ideological disagreement.
Another interesting yet not surprising finding pertains to the relationship between prevalence of cross-ideological political discussion and blogroll size. This finding appears to indicate that bloggers’ connectedness with others in the blogosphere may dictate to a certain degree the volume of cross-ideological discussion. With this in mind, political bloggers may want to focus on maximizing their own activity and involvement with the rest of the blogosphere, if their goal is to contribute to enlivening the online conversational environment.
While the present study showed that the partisan political blogosphere does see moderate levels of political talk in general and minimal levels of cross-ideological discussion in particular, future research might explore whether political blogs other than just conservative and liberal (e.g., independent, green, anarchist, etc.), or even non-partisan or apolitical blogs, might witness a livelier political discussion scene. Some ethnographic research suggests that most political talk does not arise in interactions where politics are a central focus, but rather emerge incidentally in everyday social interaction practices, such as chatting (Walsh, 2004), finding replicated in the context of apolitical online discussion groups more recently (Wojcieszak & Mutz, 2009). Consequently, it could be that online political discussion is more prevalent in blogs where politics is an incidental, rather than a central focus.
In addition, the findings reported in this study should be interpreted in light of a few limitations. First, the blogs analyzed here were randomly sampled from a particular political blog directory listing a limited collection of blogs rather than the entire population of political blogs, raising concerns related to sample representativeness. Second, the study analyzed only one post per blog. While the posts were randomly selected from the sampled blogs, and thus representative of the political blogs employed in this study, the analyzed posts are not necessarily representative of the entire political blogosphere. These two concerns, of course, pose threats to the external validity of the study and limit the generalizability of its findings. As others have noted before (e.g., Wallsten, 2007; Hargittai et al., 2008), in order to obtain a truly representative sample, one would need to have access to the
entire population of political blogs. Unfortunately, at the time this study was conducted there was no known Internet-wide, comprehensive index or directory that would make such access possible. As a result, the present study mirrored the sampling techniques previously employed in blogging research (see Adamic & Glance, 2005; Wallsten, 2007; Hargittai, et al., 2008). That said, there is little reason to believe that the political blogs randomly selected from the blog directory previously mentioned are systematically different from the rest of the political blogs in the blogosphere, at least in ways that are relevant to the scope of the present study.
Another potential issue of concern pertains to the coding of agreeing and disagreeing comments. Agreeing comments, in particular, consisted of replies which explicitly endorsed the position taken in the main blog post, as well as of those replies which did not overtly reject this position. It was assumed that comments not overtly rejecting the stance taken in the main post indicated agreement. Given the general partisan nature of political blogs, this is a reasonable assumption. It is possible, however, that some middle voices might have been subsumed in the “agreeing comments” category. Future research may do well to more clearly demarcate not only agreeing from disagreeing comments, but both of these types of comments from middle voices. Consideration of these voices may provide a more nuanced account of the political blogosphere, in particular, and political discourse, in general.
Finally, a note of caution. The present results should not be interpreted as evidence of selective exposure to partisan political blogs. While selective exposure may well be an accurate explanation for the limited cross-ideological political discussion observed, it is not possible to tell from the present analysis whether people do not read political blogs that are at odds with their own political views, or whether they merely restrain from commenting and engaging in political discussions online. It may well be that blog readers do expose themselves to ideologically opposing views in political blogs as a means to inform their offline deliberative actions. Untangling this possibility may provide another fruitful area for future research. Indeed, follow-up research could help shed light on the potential for cross-media or cross-platform effects of exposure to partisan political blogs.
Ackland, R. (2005). Mapping the U.S. political blogosphere: Are conservative bloggers more prominent? Retrieved from http://voson.anu.edu.au/papers/polblogs.pdf
Adamic, L., & Glance, N. (2005): The political blogosphere and the 2004 U.S. election: Divided they blog. LinkKDD ’05: Proceedings of the 3rd International Workshop on Link Discovery. New York, NY: ACM Press.
Baum, M. A., & Groeling, T. (2008). New media and the polarization of American political discourse. Political Communication, 25(4), 345- 365.
Eveland, W. P., Jr., Morey, A. C., & Hutchens, M. J. (2011). Beyond deliberation: New directions for the study of informal political conversation from a communication perspective. Journal of Communication, 61, 1082-1103.
Garrett, R. K. (2009). Politically motivated reinforcement seeking: Reframing the selective exposure debate. Journal of Communication, 59, 676-699.
Gastil, J. (2008). Political communication and deliberation. Los Angeles, LA: Sage.
Gil de Zúñiga, H., Puig-i-Abril, R., & Rojas, H. (2009). Weblogs, traditional sources online and political participation: An assessment of how the internet is changing the political environment. New Media & Society, 11(4), 553-574.
Greenberg, J., Simon L., Pyszczynski, T., Solomon S., & Chatel, D. (1992). Terror management and tolerance — Does mortality salience always intensify negative reactions to others who threaten one’s worldview? Journal of Personality and Social Psychology, 63, 212-220.
Hargittai, E., Gallo, J., & Yale Kane, M. (2008). Cross-ideological discussions among conservative and liberal bloggers. Public Choice, 134, 67-86.
Hayes, A.F., & Krippendorff, K. (2007). Answering the call for a standard reliability measure for coding data. Communication Methods and Measures, 1, 77-89.
Horrigan, J. B., Garrett, K., & Resnick, P. (2004). The internet and democratic debate. Washington, D.C.: Pew Internet and American Life Project. Retrieved from http://www.pewinternet.org/Reports/2004/ The-Internet-and-Democratic-Debate.aspx
Iyengar, S. (2001, Summer). Cynicism and choice: A response to The Daily We. Boston Review. Retrieved April 30, 2002 from http://www.bostonreview.net/BR26.3/iyengar.htm and on June 5, 2013 from http://web.archive.org/web/20031130040340/ http://bostonreview.net/BR26.3/iyengar.html
Iyengar, S., & Khan, H. S. (2010). Red media, blue media: Evidence of ideological selectivity in media use. Journal of Communication, 59, 19-39.
Kim, Y. M. (2009). Issue publics in the new information environment: Selectivity, domain specificity, and extremity. Communication Research, 36, 254-284.
Knobloch-Westerwick S., & Meng, J. (2011). Reinforcement of the political self through selective exposure to political messages. Journal of Communication, 61, 349-368.
Lawrence, E., Sides, J., & Farrell, H. (2010). Self-segregation or deliberation? Blog readership, participation and polarization in American politics. Perspectives on Politics, 8(1), 141-157.
McKenna, L., & Pole, A. (2004). Do blogs matter? Weblogs in American politics. Paper presented at the annual meeting of the American Political Science Association, Chicago, IL.
McMillan, S. J. (2000). The microscope and the moving target: The challenge of applying content analysis to the World Wide Web. Journalism and Mass Communication Quarterly, 77, 80-98.
Muhlberger, P. (2004). Access, skill, and motivation in online political discussion: Testing cyberrealism. In P. M. Shane (Ed.), Democracy online: The prospects for political renewal through the Internet (pp. 225-237). New York, NY: Routledge.
Mutz, D. C. (2002). The consequences of cross-cutting networks for political participation. American Journal of Political Science, 46(4), 838-855.
Mutz, D. C. (2006). Hearing the other side: Deliberative versus participatory democracy. New York, NY: Cambridge University Press.
Scheufele, D. A. (2003). Talk or conversation? Dimensions of interpersonal discussion and their implications for participatory democracy. Journalism & Mass Communication Quarterly, 77, 727-743.
Stromer-Galley, J. (2002). New voices in the public sphere: A comparative analysis of interpersonal and online political talk. Javnost: The Public, 9(2), 23–42.
Stromer-Galley, J. (2003). Diversity of political conversation on the Internet. Users’ perspectives. Journal of Computer-Mediated Communication. Retrieved from http://jcmc.indiana.edu/vol8/issue3/stromergalley.html
Stroud, N. J. (2010). Polarization and partisan selective exposure. Journal of Communication, 60, 556-576.
Sunstein, C. R. (2001). The daily we: Is the Internet really a blessing for democracy? Boston Review. Retrieved July 5, 2002 from http://www.bostonreview.net/BR26.3/iyengar.htm and on June 5, 2013 from http://web.archive.org/web/20031130040340/ http://bostonreview.net/BR26.3/iyengar.html
Trammell, K. D., & Keshelashvili, A. (2005). Examining the new influencers: A self-presentation study of A-List blogs. Journalism & Mass Communication Quarterly, 82(4), 968 – 982.
Wallsten, K. (2007). Agenda setting and the blogosphere: An analysis of the relationship between mainstream media and political blogs. Review of Policy Research, 24(6), 567-587.
Wallsten, K. (2008): Political blogs: Transmission belts, soapboxes, mobilizers, or conversation starters? Journal of Information Technology & Politics, 4(3), 19-40.
Walsh, K. C. (2004). Talking about politics: Informal groups and social identity in American life. Chicago, IL: University of Chicago Press.
Wojcieszak, M., & Mutz, D. C. (2009). Online groups and political discourse: Do online discussions facilitate exposure to political disagreement? Journal of Communication, 59, 40-56.
Xenos, M. (2008). New mediated deliberation: Blog and press coverage of the Alito nomination. Journal of Computer-Mediated Communication, 13, 485–503.
Copyright 2013 Communication Institute for Online Scholarship, Inc.
This file may not be publicly distributed or reproduced without written permission of
the Communication Institute for Online Scholarship,
P.O. Box 57, Rotterdam Jct., NY 12150 USA (phone: 518-887-2443).