Bowling Off-Topic: rethinking the echo-chamber

Image by drrt licensed under a Creative Commons Attribution 2.0 Generic License.

Interactions in online forums don’t fit so easily into Cass Sunstein’s arguments about the internet as echo-chamber. This argument is basically the oft-quoted claim about the internet fostering cyberbalkanization. Others have built on this claim by citing the lack of hyperlinks connecting left and right wing blogospheres as evidence for a decline in the ability to communicate with civility across ideological barriers (Adamic and Glance, 2005).

But when we expand scope of inquiry to include models of online interaction outside of the typical political blog platform, the notion of internet as echo-chamber becomes less clear. A variety of counter-examples from online fandom and other modes of interest-driven participatory culture suggest that Sunstein’s argument about the limitations of online communication may need to be re-evaluated.

One interesting area that emerged in our research over the past year is the role of off-topic subforums within larger interest-driven forums. These off-topic subforums provide a space for diverse conversation on the sidelines of an otherwise fairly homogeneous forum. In this sense, they seem to parallel the function of what Ray Oldenburg calls ‘third spaces’ — spaces where a heterogeneous group of people come together (in cafes, bars, and other establishments) to have casual interactions outside of work and domestic experience. For our purposes, the contrast to professional and domestic space is perhaps less salient, but the idea of “thirdness” is still relevant insofar as these subforums are neither the primary focus of the group nor are they explicitly coded as sites of civic action. Instead, they are something in between.

The decline of this kind of diverse community engagement seems to be part of what Robert Putnam laments in Bowling Alone (2000), and he uses the example of bowling clubs to talk about how membership in local institutions once provided the context for a whole host of parallel civic activities. Combining arguments of Putnam and Oldenburg, we can think of the bowling club as a pretext for interaction between folks who wouldn’t otherwise communicate were it not for the common ground of a shared recreational hobby.

Perhaps contrasting with Putnam and Oldenberg’s narrative of civic decline, though, off-topic subforums seem like a modern-day equivalent of the classic third-space. And like the bowling club, forum contributers share a core-interest which in turn provides the pretext for a safe foregrounding of the group’s diversity in off-topic subforums. For example, in a forum for diamond aficionados called Pricescope, an off-topic subforum known as ‘Around the World’ became a site for heated political debate during the run up to the 2008 elections. Our colleague Lana Swartz observed one of the participants remark: “I hate your politics, but I love your diamonds.” In this sort of expression, it seems that allegiance to a shared group identity provides a kind of buffer for the friction of intragroup differences to play out safely. Perhaps the social investment that members have in the larger forum means that they have a stake in promoting civil discourse in the off-topic areas.

In light of these kinds of observations, perhaps the question of whether the internet creates echo-chambers needs to be reframed. If we accept that cyberbalkanization is not always the case, then the more interesting question is what exactly is happening when it’s not.

This question has emerged as a key area of inquiry this summer. Currently, I’m working with Kevin Driscoll to research and design scraping tools for analyzing message boards. The idea was to design something that could augment qualitative strategies by equipping a researcher with the capability of asking sophisticated analytical questions about thread structure, group practices, user practices, and interactive patterns. Kevin conceived of the tool as part of a research paper he wrote this past spring, and this summer he’s been doing heavy lifting in Python to develop a thread scraping tool that will eventually be compatible with all vBulletin forums. I’ve been learning from Kevin, and following several forums this summer as we examine individual threads and try to think through the relationship between the analytical tools and potential research aims.

In designing the scraping tool, one of our tasks was to identify the kinds of questions that researchers might be most interested in asking, so that we could build the tool to better address those lines of inquiry. With this goal in mind, we started looking for threads that seemed to demonstrate the potential of the message board platform for facilitating productive civic discourse.

This first stage of our research was fairly ad hoc — we didn’t know exactly what we were looking for but we had a sense that we would know it when we saw it. We looked for threads that suggested potential conflict or friction, and the idea that off-topic subforums were a key site for this kind of interaction began to crystallize over time.

After these exploratory investigations, we broadened our focus to ask more basic questions about the quotidian features of a forum. For example, we were interested in questions about the technical affordances of a particular forum structure: Do you need to be logged in? Is one-button quoting technically enabled? Likewise, we were interested in understanding the intersection between these affordances and the norms of user practice, for example: when and how often do users post to a thread? What are the norms for going “off topic”? How do users let each other know who is the intended addressee? How does the group regulate these norms? Do they use particular key words (such as ‘hijacking,’ ‘trolling,’ etc.) to identify breaches? Does the group have different norms for off-topic subforums verses primary topic forums?

The answers to these sorts of questions can be markedly different depending on what kind of forum one is investigating. For example, a forum like WrongDiagnosis often has anonymous users posting one-off questions to an audience they may not have interacted with previously. In a forum called threads tend to cover a wide range of topics and so the idea of an off-topic subforum seems less applicable. Interestingly, the answers to these larger questions about the communicative practices and technical affordances of a particular forum work together to form a kind of gestalt fingerprint that is starting to help us to identify what is unique about a particular forum culture.

The answers to these sorts of questions can also provide a kind of background for asking more specific questions about individual users. For example, is a user unusual compared to group norms? Likewise, in a particular thread, is a user doing something unusual given their own track record? This is the kind of information that would be interesting to know about the participants in Pricescope’s ‘Around the World’ subforum.

Our hypothesis is that the overarching allegiances to group identity encourage participants to find diplomatic solutions to this kind of intragroup friction. If this were the case, we’d expect to find that those participants who share frequent interactions in a core topic forum will also demonstrate a different (more productive?) mode of interaction in the subforums. This difference might be manifested as a register-shift towards more diplomatic interactions, or alternatively, it might manifest itself as more overt disagreement — i.e. since the security provided by a shared sense of group identity might actually license more, rather than less, direct conflict.

Another implicit assumption here is that a group that appears homogeneous along one dimension may actually demonstrate heterogeneity along another dimension. In other words, a single group of online participants could be considered either heterogeneous or homogeneous depending on the context of interaction. A forum like the Extreme Skins forum for Washington Redskins fans might be described as homogeneous in terms of its football loyalties, but the off-topic subforum area called ‘The Tailgate’ is, by contrast, ideologically heterogeneous. In this sense, features of homogeneity and heterogeneity are performed and enacted in specific scenarios rather than something inherent in a groups’ make-up.

And yet, these communicative processes are undergirded by the norms of a particular forum community. In this way, strategies of agreement and disagreement point to that elusive bridge between micro and macro levels of social experience. This makes the multidimensionality of an online community a hard thing to demonstrate and categorize. But by harnessing analytical tools that enable us to peer beyond the micro-level context to compare and contrast multiple threads and multiple instances of participant-to-participant communication, we may be able to start building a fuller picture of the kinds of online interactions that complicate claims of cyberbalkanization.


  1. Sorry I’m only getting to this now, but this is really fascinating. I wonder whether it would be helpful to build a typology of forums according to characteristics such as homogeneity/heterogeneity, narrow/broad range of discussions, as well as technical affordances such as the one-button quote or the need to log in. Such a typology can then be helpful in thinking of the sorts of forums that are most condusive to civic discussion, or potentially civic discussion.


  1. […] the threat of balkanization in online media. While I’ve raised questions about this premise elsewhere, even if we grant the argument about balkanization, it still seems uncertain to me whether future […]

Speak Your Mind