First Monday

Getting political on social network sites: Exploring online political discourse on Facebook

This study explores use of the social network site Facebook for online political discussion. Online political discussion has been criticized for isolating disagreeing persons from engaging in discussion and for having an atmosphere of uncivil behavior. Analysis reveals the participation of disagreeing parties within a discussion, with the large majority of posters (73 percent) expressing support for the stated position of the Facebook group, and a minority of posters (17 percent) expressing opposition to the position of the group. Despite the presence of uncivil discussion posting within the Facebook group, the large majority of discussion participation (75 percent) is devoid of flaming. Results of this study provide important groundwork and raise new questions for study of online political discussion as it occurs in the emergent Internet technologies of social network sites.






The emergence of the Internet, not unlike other technologies before it, created a popular avenue for discussion of political and social issues (Holt, 2004; Puopolo, 2001). Many scholars have explored the role of computer–mediated communication (CMC) in political discussion (e.g., Brundidge, 2006; Davis, 1999; Holt, 2004). In recent years, new Web technologies, popularly referred to as social media, have opened up possibilities for rich, online human–to–human interaction unprecedented in the history of Internet communication. Of these new Web technologies, social network sites (SNS) in particular have created unique arenas for online discourse.

Sites such as are representative of the phenomenal growth social network sites have seen in recent years. Facebook was launched in 2004 and in April, 2008 surpassed 70 million active users (Facebook, 2008b). With the increase in popularity of social network sites, the potential for individuals to engage in online discussion about social and political issues has grown rapidly in a few short years.

Due to the explosive growth of social network sites, scholars have little understanding of the nature of online political discourse as it is occurring in these new social spaces. It is unclear what might influence online political discourse on social network sites compared to traditional computer–mediated communication. With the continued expansion of online applications and networks, CMC studies must seek to unpack political discussion in new and emerging online arenas. This study applies computer–mediated discourse analysis (Herring, 2004) to explore the potential of social network sites to serve as locations for online political discussion.

Potentially, this study has important implications for scholars of social network sites and social media scholars in that it provides support for a growing body of literature that demonstrates the political utility of social network sites (e.g., Gueorguieva, 2008; Hayes, 2008; Williams and Gulati, 2007). This study is also significant for scholars of computer–mediated communication broadly, and online political discussion specifically, in that it investigates the potential for elements of social networking to affect civility and diversity in online political discussions.

This study concludes that continued investigation is needed into the aspects of emerging Internet technologies that may both enable and inhibit exposure to diverse viewpoints. In order to encourage civility among online discussion participants, further analysis is need as well to understand those factors that could improve online political discussion, ultimately serving democracy.

Social networking sites

In recent years, social network sites (SNS) have become some of the most popular destinations for online traffic (comScore, 2007a; comScore, 2007b). Social network sites are:

“[W]eb–based services that allow individuals to (1) construct a public or semi–public profile within a bounded system, (2) articulate a list of other users with whom they share a connection, and (3) to view and traverse their list of connections and those made by others within the system.” [1]

These sites have become the focus of a recent wave of scholarly research. Initial research into SNS has focused mostly on networks and their structures, privacy issues, and the management of friends and online impressions (boyd and Ellison, 2007). is one of the most popular SNS. According to Baron (2008), Facebook–related research began to emerge late in 2006 (e.g., Acquisti and Gross, 2006; Ellison, et al., 2007; Golder, et al., 2007; and Vanden Boogart, 2006) and it continues to be studied across diverse disciplines (Baron, 2008; boyd and Ellison, 2007).

Evidence pointing to the potential of social network sites to serve in the arena of political discussion is emerging, indicating a need for further exploration into the political aspects of social network sites such as Facebook. For example, a survey conducted during Spring 2008 by the Pew Internet and American Life Project found that 10 percent of all Americans who use social network sites use them “for some kind of political activity.” [2] The report states that, of adults under 30 who have a social network profile, half use social network sites “to get or share information about the candidates and the campaign.” [3] A prior study by Pew conducted in December 2007 found that seven percent of all respondents and 27 percent of young adults reported using social network sites to get information about the 2008 Presidential election campaign (Kohut, 2008). Comparison of these studies demonstrates the growing political use of social network sites over the course of a few months, especially in the lives of young adults.

Social network sites such as Facebook are clearly becoming important in the political lives of their members. Facebook functions primarily to connect people virtually. But it serves other functions including as tool of political discussion. Although scholars are beginning to examine the political potential of social network sites (Gueorguieva, 2008; Williams and Gulati, 2007), there is currently a lack of knowledge about political discussion on Facebook. This study explores how Facebook is serving as an arena for political debate among members.

Political discussion on Facebook

An area of Facebook which has yet to receive scholarly study is the groups feature. Thus, prior to reviewing the literature on online political discussion, we offer below a brief overview of the Facebook groups feature. Citing a personal interview, Williams and Gulati (2007) state that there are several thousand Facebook groups with a political focus. In our research we were unable to find any source indicating just how many politically–centered Facebook groups exist.

Any Facebook account holder can start a group of nearly any kind, so long as the group does not attack a specific person or group (Facebook, 2008a). The group creator becomes the administrator who selects the title, description, and settings of the group account. Settings include designation of a group “type” which ranges from common interest to recreation and entertainment. These are then sub–divided into further classifications depending on the type. For example, common interest groups can be specifically labeled as political.

In addition to type, groups can be set as “open”, “closed” or “secret”. An “open” group allows anyone to join and to invite others to join. All group information, discussions, and posted items are available for viewing by persons with a Facebook account. A “closed” group requires approval from the group administrator(s) in order to join. Anyone can see the group information but all discussions, posted items, members, etc. remain private. A “secret” group does not appear in group search results or in members’ profiles and requires an invitation from an administrator to join. Only members are able to view group information.

A group administrator has no binding responsibilities to the group they form. Administrators are not moderators of content per se and the group will continue to exist if an administrator abandons the group. Administrators can, and often do, partake in the discussions of the group and they can delete members or content, though it is not their sole responsibility as is found in some other online forums. Members of groups can, at any time, delete any comments they made to the group though they cannot delete information posted by others. We now shift focus to a review of research in online political discussion.

Engaging opposing viewpoint

Online discussion of political and social issues has occurred over many interactive channels, including of note Usenet (Davis, 1999), bulletin boards (Brundidge, 2006; Holt, 2004), e–mail, and chat rooms (Brundidge, 2006). Within online discussion research, some scholars argue that Internet communication serves as an excellent portal for debate among persons of varied opinions and beliefs. Holt (2004) states:

“The ability of the Internet to unite those of disparate backgrounds has great potential for fostering debate and discussion of issues in the civic arena. In many cases, differences of opinion about, for example, political issues arise from lack of familiarity with the perspectives of other people.” [4]

Research has shown that online political discussion does serve to expose participants to non–likeminded partners (Brundidge, 2006). Yet, despite the potential of the Internet to bring opposing camps together in a common (cyber)space and provide exposure to different ideas, some evidence suggests that this may not necessarily be occurring.

The structure of the Internet has been found to offer conditions particularly conducive to selective exposure to media content (Bimber and Davis, 2003). Research has found that individuals favor news and information that aligns with their own views (Best, et al., 2005; Bimber and Davis, 2003; Mutz and Martin, 2001). Along this vein, some evidence indicates that selective exposure also occurs in online political discussion arenas which may lead to political polarization. For example, scholars have noted that virtual communities are fairly homogeneous in terms of values and viewpoints (Dahlberg, 2001). Participants in online discussions often hold comparable political perspectives (Wilhelm, 1999). Davis (1999) found from his seminal study of Usenet that persons engaging in online political newsgroup discussion tended to “gravitate to groups agreeing with their own views.” [5] Davis concluded that “Usenet discussion must be viewed for what it is — a means for reinforcing preexisting views rather than a virtual community where people can freely express their differing views with each other … .” [6] Thus, online communities may serve to insulate groups from those with whom they disagree.

Despite evidence suggesting it is unlikely that online political discussion enhances exposure to different viewpoints, there is potential for online discussion to lead persons of contrasting viewpoints to engage in online discussion. For example, studying online political discussion and heterogeneity among participants’ discussion networks, Brundidge (2006) concluded that individuals use the control the Internet affords them to “seek out likeminded partners for political discussion but to [sic] do not comparably avoid non–likeminded partners when they are inadvertently exposed to them.” [7]

Unique aspects of social network sites such as Facebook may serve to bring disagreeing parties together in online political discussion. Foremost, the Facebook community itself is enormous. The site reportedly reaches over 31 million people within the United States (Quantcast, 2008), 85 percent of the four–year college market and is growing most rapidly among those over 24 years old (Facebook, 2008b). The enormous reach of the site translates to a greater potential for interaction among disagreeing parties to occur on Facebook than was possible on previously existing online discussion networks. For example, the Usenet population that Davis (1999) studied was estimated in the ballpark of 180,000.

Beyond reach, novel features of social network sites may be conducive to bringing together individuals of divergent perspectives. Unlike traditional arenas of political discourse where individuals often must seek out debate communities, Facebook is a medium where people learn about the activities of others via feeds with embedded links to site content such as profile pages. These feeds provide the user updates of friends’ changes to their profiles therefore enabling immediate access to an array of information upon logging into the site (Baron, 2008). The introduction of the feeds feature in September 2006 reportedly led to a 40 percent increase in page views from September to October (Baron, 2008), indicating that indiivduals are reading and following the feeds to linked content within the Facebook network.

Because Facebook is an online space where political expression occurs on profiles through friend associations with candidates and associations with groups and political applications (Williams and Gulati, 2007), there is great potential for users to receive political information about friends in their network, such as their political stances and affiliations as well as membership in political Facebook groups.

These unique push features of Facebook may lead to increased inadvertent exposure to dissimilar viewpoints and ultimately lead disagreeing parties to interact in a common discussion space. Such an occurrence may enhance the prospects for online political discussion to move beyond its described role of “reinforcing preexisting views.” [8]

Given the emerging role of Facebook as a political utility for its users and the unique format of the site, this study seeks to explore whether engagement in political discussion occurs between persons with opposing viewpoints in a Facebook group. Thus, the following research question is proposed:

RQ1: To what extent are participants who express opposing viewpoints engaging in political discussion in a Facebook group?

Civility in online discussion

How do individuals interact once they have come together in an online discussion forum? Past research into online political discourse has illuminated a theme of uncivil behavior by disagreeing participants. For example, Davis (1999) concluded that political discussion on Usenet favored brazen behavior, resulting in “vigorous attack and humiliation” [9] on the less assertive. Davis found that the polite and respectful become discouraged from participation and the discussion becomes controlled exclusively by the belligerent. Such acts of attack have become known as flaming, a term used to describe posts made in online discussion that personally attack another poster or the poster’s ideas (Davis, 1999).

The presence of flaming in online political discussion has been reported at varying levels by scholars, with some research pointing to pervasive and disruptive levels of flaming (Davis, 1999; Hill and Hughes, 1998) and other research attributing less prominence to flaming behavior (Rafaeli and Sudweeks, 1997).

One culprit of uncivil discussion is anonymity. Researchers have noted that anonymity affords users a level of freedom and power to act in an uncivil manner as well as avoid being held accountable for their statements (Barber, et al., 1997; Davis, 1999; Streck, 1998). Facebook removes the level of anonymity previously enjoyed by online deliberators with its profile feature.

All Facebook members create distinct profiles which consist of an ever–growing array of features (boyd and Ellison, 2007) and which are personalized by displaying photos, personal descriptions and much more. Once a Facebook user joins a group, the affiliation is noted in their profile (unless the group is secret or unless the user chooses not to disclose that affiliation). In addition, their profile image appears in the “group members” section of the group profile.

Depending on the level of privacy members have selected for their individual profiles, group members may then view other members’ profiles. If an entry is made to the group discussion, a member’s profile image, along with full name and network affiliation, is displayed. Network affiliation is most likely a high school or university, but since being opened up in 2006 to anyone with an e–mail address (Baron, 2008), this can also feature employer, country or city. This is unlike other forms of asynchronous chat online where identifying information is often restricted to merely screen names or icons. The names, images and networks displayed in Facebook chat could be false or only partially realistic. However, the amount of information avalable about individuals participating in discussions may affect discourse in this unique space.

Knowing that participants in a group discussion can access at the very least the name, photograph and network affiliation of a messenger may produce a sense of accountability. Because exchange in group discussion incorporates profile information, and not simply an anonymous screen name, persons engaging in discussion may feel a greater sense of interpersonal interaction. On the other hand, identifying information may be used as a source for ad hominem attacks, particularly in instances where the discussion has devolved to uncivil arguments. To explore civility within online political discussion on Facebook, the following research question is proposed:

RQ2: What are the characteristics of discourse between members of a political group on Facebook?




Due to limited scholarly research into Facebook discourse, a qualitative method is appropriate for in–depth exploration. This study is a computer–mediated discourse analysis (Herring, 2004) of political discussion within a Facebook group. The group is focused around the current debate over U.S. policies regarding torture and was formed in response and opposition to the U.S. Military Commission Act passed by Congress in October 2006. According to a White House press release, this Act “[w]ill allow the CIA to continue its program for questioning terrorists.” (White House, 2006). The group’s stated views are that the law is an attack on habeas corpus and essentially legalizes the U.S. government to use torture. The group opposes the use of torture to obtain information from suspected terrorists.


The group under study was chosen at random from 10 pre–selected groups that met the researchers’ criteria: (1) public group type; (2) topic of interest being torture; (3) membership of more than 100; (4) active discussion postings within the past 30 days; and, (5) total discussion postings exceeding 50. Only public Facebook groups were selected due to the availability for anyone with a Facebook account to join them and the ability for the researchers to access information.

The researchers wanted to analyze discussion of a serious and timely global issue in which many people of varying backgrounds, education levels, classes, etc. would be potentially participating. The topic area of torture was selected for two reasons. The first reason was that torture was a popular topic discussed in the news globally during the time of this study, providing timeliness and relevance. State–sanctioned torture is a global issue that has garnered a great deal of media attention and therefore attracts a diverse population of communicators. The second reason was that torture is an issue that is contentious in nature. Given that this study sought to focus on the engagement of persons with varying perspectives in an online discussion space, it was important to choose a topic for which people would hold varying opinions. Further, it was important to select an issue for which people would hold strong enough opinions that they would be motivated to translate their opinions into the behavior of engaging in online discussion. A group with a membership of more than 100 increased the likelihood that there would be greater diversity among participants in the discussion. A lasting discussion with current activity was desired in order to successfully analyze the evolution of arguments and the ongoing participation — or lack thereof — from members.


At the time of data collection, the group under consideration was made up of over 800 members from around the world. The data set comprised all posts present on the discussion board from its inception up until the date of data collection — 15 April 2008. This does not include any posts that were deleted by any member of the group and thus not available at the time of data collection. The data set consisted of 176 total discussion postings spanning an 18–month period. When printed, the data was 49 single–spaced pages of text displayed with a 12–point font. To ensure privacy, the names of group members were replaced with initials and the title of the group was withheld.

In total, 66 individuals participated in the group discussion. This constituted eight percent of the group population. Of the 66 participants, only 10 posted five or more times. Their posts combined totaled 103 and made up 59 percent of the entire discussion. The highest number of posts by an individual was 22. For the purpose of this study, we will focus primarily on the trends found within this group of frequent communicators. Though characteristics of these participants will be discussed throughout the analysis, an overview of top 10 participants is provided in the Appendix.

Data analysis

A computer–mediated discourse analysis, applying the coding and counting approach (Herring, 2004) was used to analyze asynchronous discussion within a Facebook group. The process of analysis was derived from Glaser and Strauss’ (1967) constant comparison model, which entailed repeated analysis of the data to allow for coding categories to emerge from the analysis process. Analysis of participation, in which “phenomena of interest are number of messages and responses and message and thread length,” [10] was employed.

The process of analysis consisted of six steps. First, each researcher individually read the data set twice to establish familiarity with the data. We then individually examined and identified emergent trends based on the research questions through repeated analysis. Third, we compared our individual analysis to identify categories and classify operational definitions. Fourth, we individually coded the data based on the operational definitions. Fifth, we compared our individual analyses. For the purpose of validity and reliability, inconsistent results were identified and discussed to reach a consensus on data coding. Sixth, we returned to the data to confirm our findings. As a result of analysis, we arrived at the following results.




Research Question 1

Presence of opposing viewpoints

Research question one sought to address to what extent participants who express opposing viewpoints engage in political discussion within a Facebook group. We assessed the content of each post to determine whether the views expressed in each were in support or opposition to those of the group. After categorizing each individual post, we used these initial categorizations to identify how many participants were in support, opposition or remained neutral to the group’s stance. Through analysis we identified three types of posters: Support, Opposition and Neutral.

An individual’s posts were always clearly either in opposition or support, never mixed. That is, we experienced no members who flip–flopped in their stance on the issue. However, individuals’ posts may have included a mixture of neutral posts. Where this occurred, they were categorized based on which stance was exhibited. For example, those who had both posts supporting the group stance along with some neutral posts were categorized under “Support” despite the number of neutral posts evident. Therefore, those categorized as “Neutral” never posted any discourse in support or opposition of the group stance. The categories are further explained in detail below.

“Support” posters were selected for their characteristic disagreement with the U.S. Military Commissions Act of 2006. Commonly, these posters specifically communicated opposition to the use of torture to obtain information from detainees but they also broadly expressed opposition to torture in general for any purpose as well as opposition of government infringement on personal rights such as habeas corpus. An example of a post characteristic of a poster in “support” of the group’s stated interest is clear in the following post:

“I don’t give a damn if torture WAS useful to get information. It still goes against the Geneva Convention, and is a blatant assault on human rights. What is this world coming to? We’re SUPPOSED to be the good guys, but the way I look at it we aren’t doing a very good job of showing it. Torture is wrong. 100 percent. No justification.”

“Opposition” posters were distinguished by discourse showing favor for the use of torture. Commonly these posters communicated beliefs that torture was justified against terrorists or suspected terrorists and that torture was an effective means of extracting information from suspects. A clear example of a post made by a poster in “opposition” to the group is found in this post:

“bunch of liberal pansies … do you really feel sympathy for the same people that blew up a part of our country? Do you think these ‘people’ (and I used that term loosely) would stop to consider what they are (and did) to our prisoners? … I say whatever happens to them is much less than they deserve … .”

A category of “Neutral” posters was required due to the presence of posters who made no post indicating either clear support or opposition of the group’s stance on the Military Commissions Act of 2006. These posters made posts that were usually either unrelated to other discourse taking place or posts in which no bias was communicated.

Of the Top 10 posters, seven members aligned with the group and three opposed. In total, of the 66 individual posters, 48 aligned with the group (73 percent), 11 were in opposition to the group (17 percent) and seven were neutral (10 percent). This confirmed that the individual group members were diverse in their views concerning the topic despite the fact that the group itself had an established stance on the subject. Although expression of opposing viewpoints was present within the discussion indicating a level of interaction among disagreeing parties, the large majority of posters expressed likeminded perspectives.

Research Question 2

Research question two sought to investigate the characteristics of discourse in a political Facebook group. After repeated analysis of the individual posts and the overall discussion, four types of posted content were identified and categorized: Informational, Productive Argument, Unproductive Argument and Miscellaneous. Each category is explained below.

The first category is the “Informational” category and includes posts purely informational in nature. These were posts that shared links to external sources, dates of events or other items of interest, or referenced bodies of information pertaining to the topic of state–sanctioned torture. An example of an informative post is the initial group post which provides a link to a page within the Federal Research Division of the Library of Congress. The linked section is labeled “Military Commissions Act of 2006” and provides further links to files containing the actual language of the Act at different stages of the process as it was introduced, debated, amended, approved and reported on by the U.S. House of Representatives, U.S. Senate and their respective committees.

This post provides group members with the original documentation and information regarding legislature that the group opposes. The post itself does not contain any statement of stance or opinion regarding the Military Commissions Act of 2006. This post was followed by a series of informative posts. In fact, eight of the first 10 posts to the discussion board included links to informative Internet sources. These included links to news articles from sources such as the New York Times, Washington Post, Register, and Yahoo News.

Also categorized as informative posts were those providing information on how members could take action or extend their Facebook group affiliations into real life events. Such posts encouraged people to vote, attend protest marches or ask their Congressional representatives to support certain legislature.

One post in particular, made by member GK, urged group members to call their representatives and ask them to support the Restoring the Constitution Act. The call to action was followed by three links for further information. The first led to an overview piece at, the second to an article declaring support of the act on the Human Rights Watch Web site and the third to a government tracking site which provides an overview and summary as well as a tracking of all amendments and actions regarding the Restoring the Constitution Act. This site also informed viewers as to which representatives co–sponsored the bill and other related legislature histories.

In a group where members’ opinions and knowledge of torture issues varied greatly, the legitimacy of discussion taking place ranged from unjustified personal attacks and name–calling to well–structured arguments complete with concessions and well–supported facts. Overall, posts tended to display a favored style of argument. For example, those whose posts displayed profanity, name–calling and/or personal attacks tended to show a trend in these types of posts. Isolated events were uncommon, especially among the most active communicators.

Respectively, those members who posed opinions but were also seeking or willing to listen to other opinions tended to maintain this style throughout. These posts were placed in category two, “Productive Arguments.” This category includes posts (regardless of stance on the issue) which pose an argument or respond to an argument but are devoid of any harsh judgment, personal attacks, flaming, etc. This type of discourse is strongly exemplified by group members MR and RP. These two members participated in a discussion consisting of nine posts directly engaging one another that are each categorized as “Productive.”

Of note in these exchanges are the willingness each member has in considering other opinions, posing genuine and provoking questions, and the ability to recognize their own misconceptions. For example, in one post, MR makes such statements as, “… I would like to here [sic] a rational response to my argument. Cause maybe im [sic] wrong …” and “… k [sic] maybe im [sic] exaggerating the issue a bit, but I dunno …” He also states, “… (of course there are the exceptions …” and even moves to unify himself with group members by asking, “… what does [sic] everyday people like us have to concern over matters of this nature?” MR maintains this style of discourse. He admits in one post, “… (this is just my view of things btw, not saying it has to be right … .” In direct response to content in RP’s posts MR makes other productive moves such as restating a point MR makes and finishing it with, “… I can’t agree …” and, “I see your point though … .” RP’s discourse is similar and also considered productive. Much like MR, RP is willing to point out the good and bad in the opposing argument. In one post RP states, “Also, you make a good point about …” and in the following post, “… I still don’t find your defense of torture persuasive.”

Category three is the “Unproductive Argument” category and includes posts that contain some elements of argument but feature prominent use of personal attacks and insults to the group at large or individuals. These posts range from mild to downright offensive but all of them contain some discussion of the topic and are therefore relevant to the overall discourse. For example, group member MW posts this category three response:

“It is not anyone here’s [sic] job to convince you that you’re wrong, but if you’re here to challenge the GROUP’s opinions, back it up. Oh, and cut down on the language — AND learn how to spell. I guess what it comes down to is to THINK before you write, instead of writing out your ass. Maybe I’m asking for too much … .”

This is a mild personal attack but the emphasis is put on the attack rather than furthering valid discussion or respecting opposing viewpoints as seen in our discussion of “Productive Argument” posts above. Other examples of “Unproductive Argument” are much harsher. Two members, SJ and ZB, exchange more than 10 messages of this type. Distinctive and qualifying discussion elements in these posts include comments to individuals and the group such as,

“… you disagree, and your [sic] an ignorant FUCK.” And “… you people make me want to vomit and eat it.” Following a synopsis of their viewpoint, ZB ends a post with, “In conclusion, go to hell, you sadistic chickenhawk [sic] fuckwit [sic].” The profanity and personal attacks are numerous between the two members. Such posts are in stark contrast of those we categorized earlier as “Productive Argument.”

The final category, the “Miscellaneous” category, encompasses posts which were unclear in their purpose, unintelligible, or completely non–pertinent or off–topic. These were often short, solitary posts from members not actively communicating. Examples include, “You do understand that we are all probably on a watch list by now?” and “Jack Bauer can torture you into giving up information you do not possess.” These were not posted in direct response to any posts going on at the time and were the only posts by these members. Others included posts like this, “Thanks for the PDF John.”

Some posts were deleted by members leading to some confusion in our analysis. One post indicated that some contentious material was deleted so some posts are unclear and seemingly unrelated to a larger discussion. Examples of this kind that were categorized as “Miscellaneous” are, “FUCKED UP” and “Hitler was an atheist.” Most likely these posts are in relation to deleted posts and therefore we could not adequately evaluate them.


To analyze trends within discourse and explore the level of civility displayed among participants, existing categories were further grouped according to the presence of either positive or negative styles of discourse. Posts falling in the informational and productive argument categories were labeled “Civil” content posts. Civil posts represent posts that focused on issues or argument and were executed in a generally respectful manner. Posts in the unproductive category were labeled “Uncivil” content posts. Uncivil posts were those that contained flaming and thus were a potential threat to respectful discussion. Posts that fall into the “Miscellaneous” category are considered “Neutral” content posts.

Each member of the Top 10 was evaluated to see which style of communication they tended to utilize in their postings. Seven of the Top 10 (the majority) engaged in predominately civil posts and were categorized as “civil” communicators. Only three of the Top 10 participants made predominately uncivil posts. These individuals, SJ, RW and ZB, were categorized as “uncivil” communicators. 68 posts (66 percent) from the Top 10 participants were categorized as “civil”, 31 (30 percent) were categorized as “uncivil” and four posts (4 percent) were categorized as “neutral.” Table 1 shows the number of posts in each category as well as the percentage it represented of each individual’s posts.


Table 1: Post type by Top 10 participants.
ParticipantPost type (percent)
MW18 (82%)4 (18%)0
KN11 (65%)6 (35%)0
GK10 (91%)1 (9%)0
NC7 (58%)3 (25%)2 (17%)
SJ3 (33%)6 (67%)0
MR7 (78%)1 (11%)1 (11%)
RP5 (83%)1 (17%)0
RW2 (29%)5 (71%)0
MKW4 (80%)01 (20%)
ZB05 (100%)0
Total68 (66%)31 (30%)4 (4%)



Given the unique structure of discussion in Facebook groups, a final area of interest was whether identifying elements of a Facebook user’s profile were the subject of discussion. Analysis of the data revealed that discussion posts were largely absent of elicitations for identifying information provided by the Facebook group discussion format. Over the life of the discussion there were only seven instances where personal attributes were brought up within the discussion. All but one of these instances involved group member SJ. One instance occurred within a post made by ZB in which ZB comments on SJ’s profile image in order to insinuate that SJ was of low class and intelligence. Another such post occurred during an argument between ZB and SJ. The post came from someone who claimed to know SJ in real life, though it was unclear in what capacity the two were acquainted. This member first stated an acquaintance with SJ and insisted on debating SJ’s previously stated viewpoints. The member insinuated that since she knew SJ her argument would be more effective than others in disagreement with SJ. A third such post brought forth another unique aspect of Facebook group discourse. Within this post SJ stated, “… not that you know anything about me.” In certain online discussion forums this could be true but in Facebook ZB may actually have had access to quite a bit of SJ’s personal information depending on how SJ managed a profile.

Participants in Facebook group discourse may forget how much information is actually available to their fellow communicators. Towards the end of the argument, SJ started to make attempts to repair the discourse and move toward constructive argument. During this time SJ requested that ZB stop making personal attacks and insisted that the profile picture in question does not represent SJ. Although they were few, all of these cases highlight the impact of the unique structure of Facebook on discussion.




This study expands upon previous research into the political utility of social network sites (i.e., Gueorguieva, 2008; Williams and Gulati, 2007) by demonstrating that Facebook is a legitimate location for discussion of political issues. The discussion under study consisted of participants with opposing viewpoints. To some extent the discussion appears to have succeeded in overcoming polarization of online discussion that has pervaded online political discussion in the past (e.g., Davis, 1999). However, representation of viewpoints within the Facebook group was highly skewed in favor of discussion among likeminded participants with 73 percent of all posters aligning with the group stance on state–sanctioned torture. Only 17 percent of posters were in opposition to the group’s stance. This indicates that the unique social media network structure of social network sites such as Facebook may play a minor role in facilitating participants’ exposure to disagreeing parties and heterogeneity in online political discussion. However, because this study did not contain any interviews with discussion participants, conclusions cannot be drawn as to how or why disagreeing parties came to know about the group and what motivated them to join it.

Although it may facilitate diversity in political discussion, Facebook could enable individuals to insulate themselves from interaction with those they disagree with. In this study we explored discussion within one Facebook group about a specific issue, state–sanctioned torture. However, the fact that there are numerous torture–related groups means that Facebook users with an opinion about the issue of torture have the ability to seek out a group with likeminded members. Any member who does not find a group that conforms to their perspective has the ability to establish a new group to promote their view and recruit likeminded members.

Despite a few ‘bad apples,’ civil argument and information exchange characterized much of the discussion. Our findings confirm past research showing that online political discussion tends to contain a significant level of uncivil discussion (Davis, et al., 1998; Rafaeli and Sudweeks, 1997). However, our results contradict the conclusion drawn by some scholars (e.g., Davis, 1999) that online political discussion is dominated by belligerence that discourages participation by the polite and respectful. That 30 percent of posts by the most active 10 participants were uncivil in nature means there is room for behavioral improvement in online political discussion. Yet, seven of the 10 most frequent posters were predominately positive in their discourse styles. Overall, 130 of the 176 posts (75 percent) were civil and represent a willingness of participants to engage in a discussion group despite the presence of uncivil or even aggressive communication. On the whole, our results provide some support for Brundidge’s (2006) conclusion that individuals use the control the Internet affords to “seek out likeminded partners for political discussion but to [sic] do not comparably avoid non–likeminded partners when they are inadvertently exposed to them.” [11]

Although each country has unique characteristics, we discovered a number of common features among participants throughout the five countries. In particular, we found a wide geographic distribution of participants from inside and outside the region, supporting our argument that chat and forum participation enhances the sources of information available both to the participants and to their off–line communities. We also found that the content of the discussion, rather than the domain in which the site is located, is a major factor determining the popularity of the site. We find this especially interesting because we had hypothesized that individuals might find the use of chats and forums in a domain outside of the reach of a repressive government a safer environment in which to engage in public debate. Nevertheless, censorship and self–censorship practices within certain sites are easily observable. In addition, across all countries, Russian or English are the primary languages for site administration and substantive discussions. Rather than using individual national languages, the preference for Russian and English demonstrates the ties to a broader online community of participants from various geographic locales. Finally, we found that, even when using non–national language in the very global medium of the Internet, Central Asian chat and forum participants create and use avatars and emoticons, nicknames, images, and symbols that reflect their national or ethnic characteristics. In this way, online communities are imbued with local culture. These findings are discussed in more detail below.

By analyzing the discourse trends of the most frequent posters over the lifetime of the discussion, this study highlights unique patterns of behavior that offer an alternative explanation for trends of flaming in online political discussion. In all forms of discussion there will likely always be those who are brazen and boisterous and those who are civil and considerate. In all such cases the persons who yell the loudest would appear to draw the most attention. In an online forum, the loudest person is established by frequency or audacity of post rather than volume. Yet this study shows that the loudest persons are not necessarily the person drawing widespread attention from the group. Rather, the group discussion board served as a forum for argument for a concentrated type of poster with certain discourse characteristics whereas the majority enacted a predominately civil discourse style by engaging in constructive interaction through argument and exchange of information related to an issue of common concern, despite their stance on the issue.

This study also explored how unique aspects of Facebook, namely the presence of identifying information via the profile, may influence political discussion. Given that personal attacks were relatively few and civil behavior was fairly prominent, we can infer that the presence of identity attributes within the Facebook discussion may have a positive impact on the nature of online political discussion. It may be that these attributes of Facebook engender a more personalized, human–to–human feel that promotes civil interaction in online discussion. Further research is needed to explore this possibility and the potential role identity attributes may play in fostering more civility in online discussion. On the other hand, elements of the Facebook profile can become fuel for those types of communicators who engage in malicious online discussion. This unique aspect of Facebook may make engaging in online political discussion via Facebook particularly troublesome for those who become targets of aggression. Persons engaging in political discussion on Facebook should take caution and be proactive in establishing a comfortable level of access to their profile information by adjusting their privacy settings.

Limitations of study

Due to the focused nature of our study on one Facebook group, our results cannot be generalized to other Facebook groups or to political discussion on social network sites in general. Of note, our results cannot be generalized to “closed” or “secret” Facebook groups. Despite this drawback, this study opens new questions and avenues for computer–mediated communication scholars and social media scholars alike. The results of this study may be skewed towards lesser diversity of opinion given that the group had a stated position about the issue. This may have caused the group to be comprised of a high proportion of persons who support the group’s stated position resulting in over representation of individuals in support of the group’s position within the group discussion. Keep in mind one must be a member of the group in order to post messages. Had the group under study not taken a polemic stance about the issue, it is possible that a greater representation of diverse viewpoints would have been present.

Suggestions for future research

Future study should seek to expand upon the results suggested in this study that the unique structure of social network sites may serve to help bring together individuals of diverse and opposing opinions into a common online discussion space. The alternative possibility, that Facebook may serve to perpetuate isolation, must also be explored. Interviews should be conducted with discussion participants to examine the courses traveled by discussion participants that led them to enter a particular Facebook group discussion so as to explore what motivates and facilitates exposure to disagreeing parties in online discussion spaces.

The presence of identifying attributes within Facebook online political discourse is an issue that is in need of further investigation. Future scholarship should explore in–depth the role of social network site identity within online political discussion, and online discussion broadly. Future research could survey or interview Facebook users who utilize group discussion to see how access to profile information may, or may not, influence participation in group discourse.

Lastly, beyond text–based discussion, even richer forms of interaction are becoming possible through inexpensive media production equipment such as personal video cameras, editing software, and free online hosting. There is a rich, unexplored space of political discussion occurring through multimedia on sites such as Scholars interested in online political discussion should begin investigating this emerging arena in order to understand the rich mediums that people are playing with now to express themselves and engage others.




The results of this study are significant in that they demonstrate the capability of social networks to afford persons of different perspectives the ability to coalesce and engage in political debate. Furthermore, they demonstrate that people are seeking beyond recreational use of these social media and are harnessing the capabilities of these technologies to engage in political discussion and express their views about issues they care about. Although online political discussion was around well before the advent of social network sites, the richness and popularity of these technologies, along with the decreasing cost of Internet access, may be bringing larger populations and new participants into the foray of online political discussion. Furthermore, as this study indicates, these sites may promote civility in discussion and interaction among disagreeing parties. We are in the midst of a rapidly expanding population of citizens who are using social media as a political utility. New, creative political uses of social media will continue to emerge, yet the fundamental need for citizens to discuss issues will remain. Despite the drawbacks of online political discussion, use of the Internet remains a popular medium for people to connect and debate issues they care about.

This study also inserts some optimism into discussion of online political discussion. Although flaming remains a problem in Facebook, it does not dominate nor does it drive out those who seek to interact in a civil manner. With that said, the Internet is by no means a fully democratic medium where the people come together to politely and rationally discuss viewpoints and arrive at the best possible conclusion, à la technological utopianism. Be that as it may, the Internet may be the best hope we have. Scholars and practitioners alike shoulder the duty to study and critique emerging Internet technologies, not so as to condemn them but, rather, to create opportunities to continuously improve their democratic potential. It is the hope that this study provides a launching pad for research into the potential of social media in the civic process. End of article


About the authors

Matthew J. Kushin (MA, University of Miami) is a Ph.D candidate in the Edward R. Murrow College of Communication at Washington State University. His research interests are in new media.
Direct comments to matthew [dot] kushin [at] email [dot] wsu [dot] edu

Kelin Kitchener recently earned her MA in English from the University of Idaho. This paper served as part of her overall thesis project which focused on displays of political rhetoric. She can be reached at kelin [dot] kitchener [at] vandals [dot] uidaho [dot] edu.



An earlier version of this paper was presented at the annual convention of the Western States Communication Association, Phoenix, Arizona in February 2009.



1. boyd and Ellison, 2007, p. 211.

2. Smith and Raine, 2008, p. ii.

3. Ibid.

4. Holt, 2004, p. 14.

5. Davis, 1999, p. 162.

6. Ibid.

7. Brundidge, 2006, p. 21.

8. Davis, 1999, p. 162.

9. Davis, 1999, p. 163.

10. Herring, 2004, p. 358.

11. Brundidge, 2006, p. 21.



A. Acquisti and R. Gross, 2006. “Imagined communities: Awareness, information sharing and privacy on the Facebook,” paper presented at the Sixth Workshop on Privacy Enhancing Technologies (PET), Cambridge, U.K.; version at, accessed 18 October 2009.

B. Barber, K. Mattson, and J. Peterson, 1997. “The state of electronically enhanced democracy: A survey of the Internet,” report for the Markle Foundation, Walt Whitman Center for Culture and Politics of Democracy. New Brunswick, N.J.

N. Baron, 2008. Always on: Language in an online and mobile world. Oxford: Oxford University Press.

S. Best, B. Chmielewski, and B. Krueger, 2005. “Selective exposure to online foreign news during the conflict with Iraq,” Harvard International Journal of Press/Politics, volume 10, number 4, pp. 52–70.

B. Bimber and R. Davis, 2003. Campaigning online: The Internet in U.S. elections. New York: Oxford University Press.

d. boyd and N. Ellison, 2007. “Social network sites: Definition, history, and scholarship,” Journal of Computer–Mediated Communication, volume 13, number 1, pp. 210–230, and at, accessed 18 October 2009.

J. Brundidge, 2006. “The contribution of the Internet to the heterogeneity of political discussion networks: Does the medium matter?” paper presented at the International Communication Association, Dresden, Germany, at, accessed 25 June 2008.

comScore, 2007a. “Social networking goes global,” at, accessed 15 March 2008.

comScore, 2007b. “comScore Media Metrix releases top 50 Web ranking for July,” at, accessed 15 March 2008.

L. Dahlberg, 2001. “Computer–mediated communication and the public sphere: A critical analysis,” Journal of Computer–Mediated Communication, volume 7, number 1, and at, accessed 14 July 2008.

R. Davis, 1999. The Web of politics: The Internet’s impact on the American political system. Oxford: Oxford University Press.

N. Ellison, C. Steinfield, and C. Lampe, 2007. “The benefits of Facebook ‘friends’: Social capital and college students’ use of online social network sites,” Journal of Computer–Mediated Communication, volume 12, number 4, pp. 1143–1168, and at, accessed 18 October 2009.

Facebook, 2008a. “Create a group,” at, accessed 20 April 2008.

Facebook, 2008b. “Press room,” at, accessed 20 April 2008.

B. Glaser and A. Strauss, 1967, The discovery of grounded theory: Strategies for qualitative research. Chicago: Aldine.

S. Golder, D. Wilkinson, and B. Huberman, 2007. “Rhythms of social interaction: Messaging within a massive online network,” paper presented at the Third International Conference on Communities and Technologies (CT2007), East Lansing, Michigan; version at, accessed 18 October 2009.

V. Gueorguieva, 2008. “Voters, MySpace and YouTube: The impact of alternative communication channels in the 2006 election cycle and beyond,” Social Science Computer Review, volume 26, number 3, pp. 288–300.

R. Hayes, 2008. “Providing what they want and need on their own turf: Social networking, the Web, and young voters,” paper presented at the National Communication Association Annual Conference, San Diego, Calif.

S. Herring, 2004. “Computer–mediated discourse analysis: An approach to researching online behavior,” In: S. Barab, R. Kling, and J. Gray (editors). Designing for virtual communities in the service of learning. New York: Cambridge University Press, pp. 338–376.

K. Hill and J. Hughes, 1998. Cyberpolitics: Citizen activism in the age of the Internet. Lanham: Rowman & Littlefield.

R. Holt, 2004. Dialogue on the Internet: Language, civic identity, and computer–mediated communication. Westport, Conn.: Praeger.

A. Kohut, 2008. “Social networking and online videos take off: Internet’s broader role in campaign 2008,” Pew Research Center for the People and the Press (11 January), at, accessed 8 April 2008.

D. Mutz and P. Martin, 2001. “Facilitating communication across lines of political difference: The role of mass media,” American Political Science Review, volume 95, number 1, pp. 97–114.

S. Puopolo, 2001. “The Web and U.S. senatorial campaigns 2000,” American Behavioral Scientist, volume 44, number 12, pp. 2,030–2,047.

Quantcast, 2008. “,” at, at July 2008.

S. Rafaeli and F. Sudweeks, 1997. “Networked interactivity,” Journal of Computer–Mediated Communication, volume 2, number 4, at, accessed 18 August 2009.

A. Smith and L. Raine, 2008. “The Internet and the 2008 election,” Pew Internet & Life Project (15 June), at, accessed 8 June 2008.

J. Streck, 1998. “Pulling the plug on electronic town meetings: Participatory democracy and the reality of the Usenet,” In: C. Toulouse and T. Luke (editors). The politics of cyberspace: A new political science reader. New York: Routledge, pp. 18–48.

M. Vanden Boogart, 2006. “Discovering the social impacts of Facebook on a college campus,” unpublished master’s thesis, Kansas State University, Manhattan, Kansas.

White House, 2006. “Fact sheet: The Military Commissions Act of 2006,” at, accessed 30 April 2008.

A. Wilhelm, 1999. “Virtual sounding boards: How deliberative is online political discussion?” In: B. Hague and B. Loader (editors). Digital democracy: Discourse and decision making in the information age. London: Routledge, pp. 154–178.

C. Williams and G. Gulati, 2007. “Social networks in political campaigns: Facebook and the 2006 midterm elections,” paper presented at the annual meeting of the American Political Science Assocation, Chicago, at, accessed 30 June 2008.


Appendix: Characteristics of Top 10 participants


Note: Participants are listed in the order in which they joined the discussion. *Administrator.



Editorial history

Paper received 28 August 2009; accepted 15 October 2009.

Copyright © 2009, First Monday.

Copyright © 2009, Matthew J. Kushin and Kelin Kitchener.

Getting political on social network sites: Exploring online political discourse on Facebook
by Matthew J. Kushin and Kelin Kitchener.
First Monday, Volume 14, Number 11 - 2 November 2009