Response to Club Troppo Thread
Hi all. This continuing thread over at Club Troppo is discussing the merits (or lack thereof ;)) of our research. I had written a long reply post but it felt a bit rude to post something of that length on someone else’s blog, so I’ve stuck it here and tracked back to the thread.
Hi all - perhaps I should try to make it clear once more that we’re interested in the qualitative aspects of blogging, too (as broad and strange a comment as that appears to be when it’s written down). The piece under discussion is not the final word on the research we’re doing on the blogosphere. It reflects Axel’s interest in using mapping/tracking tools. (It’s a sole-authored paper, too, you’ll note: lots of comments are implying joint authorship of the paper. I like it but I didn’t write it.) I think that some commenters are still assuming that making maps is all we’ll be doing. I think, too, that people are upset because they see us as approaching this in a reductive way, or as missing certain important blogs, but (I get to this below) this particular paper is explicitly limited to examining a particular aspect of the Australian blogosphere.
Laura and Ken - I’ll try to deal with some of your points together if I could. I’d say to everyone though that a lot of the things you’re picking up are explained in the methodological section of the paper. Perhaps I should have linked to that, rather than just a chart. Here you go.
Here’s the abstract, by the way:
This paper explores methodologies for using the IssueCrawler research tool to map the interconnections of individual blogs in sections of the blogosphere. It uses the case of Australian–born Guantanamo detainee David Hicks as a case study, mapping the distributed discussions of this case in that part of the Australian blogosphere which is concerned with debating news and politics. Its findings indicate the presence of a strong and sustained engagement with this case by Australian political bloggers, and point to a tendency for discussions to cluster around a handful of sites which are defined by their political orientation. The network maps also suggest a lack of sustained coverage of the case by bloggers outside of Australia, and indicate only limited engagement between bloggers and the mainstream media.
So, nothing about left and right in this summary of the overall thrust of the paper - that’s offered, towards the end, as one explanation for the presence of links, and of the apparently greater propensity of some blogs to discuss the Hicks issue relative to other blogs. There’s nothing about “solidarity” up-front either, just the apparent presence of engagement with the issue, and one another, on the part of some blogs/bloggers. The question of course, is how to explain this? If not in terms of political alliegances, is there another explanation? (Can I assume that we all agree that there is some explanation available, somewhere?) When Ken writes
A possibly larger factor is that Bruns et al have picked an issue that is “owned” by “the left” (for want of a more sophisticated classificatory system). You would expect to see more posts and more links between bloggers who could be expected to be strongly interested in that issue than between those for whom it is mostly an “irritant” issue only discussed reactively. If you chose an issue “owned” by “the right” (e.g. the Jyllands Posten cartoons of Mohammed), I suspect you would see a very different clustering pattern. The clustering may well have more to do with the nature of the issue than the nature and extent of community.
This seems puzzling to me in two, related ways. First, what you’re saying here seems very, very close to what Axel said in the first instance, and which you seemed at first to be quite exercised by. You’ve been saying our conception of left and right is limited (that’s fine - no doubt it is), but then you say here that the issue in question is “owned” by “the left”, and that the bloggers in question could be expected to comment more on the issue for that reason. I think you’re absolutely right - the cartoons issue would turn up an entirely different group of closely-clustered blogs. But how to explain the difference between the groups, if, as you say, they’re likely to be different? If there are problems with Axel’s explanation, then I think yours could be vulnerable to the same objections.
Second, it seems that you’re assuming that “issues” and “communities” are easily separated. If blogs - or “that part of the Australian blogosphere which is concerned with debating news and politics” (NB a limiting gesture) are essentially about dialogue - how else would community form except around the discussion of issues? To be clear in the question: what does community on the blogosphere consist in, if not in dialogue around issues as they emerge? You’re right to point out that all of this is likely to be quite dynamic - particular issues will activate different communities of interest, and certain bloggers might be more involved in particular conversations than others, but that’s precisely the value of a case-study, or a series of case-studies. If you dispensed with the word community, there would still be links, an apparent attraction to a particular issue, and some apparent “congregating” as Laura puts it, left to account for. Isn’t it actually pretty fair to suggest (ahead of qualitative work that might nuance this in a satisfactory way) that these blogs constitute some sort of community of interest and debate? If so, what else do they share?
Laura writes:
Yes, what is that .com .net business? Everyone knows that having your own domain is supposed to be a sign that you mean serious business, but everyone also knows that if you host on Blogger you have no worries with service interruptions.
That’s something issuecrawler does when it’s drawing your maps for you. No value judgement implied on our part, though it may be something that the Issuecrawler people need to think about. I think it’s annoying too. Same with the poor legibility of some of the blogs in the chart.
Laura again:
I’ve been reading most of the blogs in the crowded part of the chart for three or four years. (the ones whose names I can decipher, anyway.) It’s no news to me (nor I should imagine is it news to any other denizen of this version of the Australian blogosphere - there are other versions, of course) that they link to each other in blogrolls and in posts, subscribe to each others’ feeds, comment on each others’ posts. Can this congregating be explained as an expression of leftist solidarity?
Well that’s the question we started with. How do we explain it? But the issuecrawler work offers to move the discussion from our shared anecdotal hunch that they congregate to a more concrete picture of the ways that they do link together, based on particular issues. That’s what I meant by this research enabling us to ask new and interesting questions - it provides new information and refines the question at hand, or at least gives us a new perspective on it. It needn’t be seen as a picture of solidarity - it could possibly be innumerable shadings of disagreement, but that’s where we’re pointed towards more qualitative evaluation. It’s not the last word, but is having the last word what doing research is all about? (By the way, you’ll see in the paper that blogrolls are specifically excluded - it’s a per-page analysis rather than a URL-based analysis.)
Laura again:
There are heaps of ‘lefty’ bloggers around who don’t link much and who don’t get many links. Many of them are just as readable as the ones on the chart. If dense interlinking is to be satisfactorily explained by the bloggers’ political tendencies then why are some lefty blogs not part of the linky network? I see that technorati has been asked to show blogs with some authority, which technorati measures by links, since it has that type of mind. Can Technorati also explain why some blogs don’t get into the linking business? And does this compromise their leftiness?
You’re dead right on the first point - there are inevitable limitations based on which URLs are seeded in the first place, there’s no way to capture every Australian blog in this picture, and some bloggers don’t engage in the kind of inter-blog debate and commentary that is the focus of this paper. But that’s all acknowledged up-front when the methodology is discussed. With respect, though, I think the second point is fallacious. Axel’s saying that the “linkiness” (nice word btw) of these blogs may be explained in terms of some shared political convictions, but not that linkiness on Hicks is a criterion of leftinesss. There are obviously some lefty bloggers who don’t link much, or didn’t discuss the issue, and clearly there are many lefties who don’t have a blog at all! The study is not framed to take non-linky blogs into account - a quantitative mapping of connections between blogs couldn’t be - but surely it shows us something interesting about a pretty prominent bunch of blogs out there.
You’re absolutely right in saying that there’s a lot that technorati can’t explain. It can’t tell us really interesting stuff about why people get into blogging, how they connect it with their own political beliefs, or how they inspire their readers and other bloggers. It’s unarguably true that lots of well-written, well-read blogs have an influence that may well be impossible to quantify, and which certainly isn’t susceptible to this kind of analysis. But that’s why we’re doing interviews, and working in other modes of research including textual analysis (and I guess, fwiw, that’s much closer to where my own interests in the project lie).
But for mine, one of the most important ways in which blogs differ from other forms of politcal information and debate is in their appeals to dialogue and community. Community happens within and between blogs, and communities, even dynamic ones, must have an outside as well as an inside. There are lots of different ways of approaching and thinking about that, but I think that this one is certainly legitimate, within its acknowledged limitations - and it could be that this discussion has gotten us a little further towards explanations of what it’s turned up.
I can’t speak for Laura. However, I don’t have a problem with your research per se, indeed I think it’s great that you’re doing it. Nor do I have a problem with the link mapping exercise as a first step in exploring patterns of interaction and development of “community” in the political blogosphere. My sole problem was (and remains) with the increasingly strident partisan claims in ABC articles and blog posts to the effect that left-leaning blogs are more like Daily Kos in size and influence, that right-leaning blogs are less “developed” in that sense, and that left-leaning blogs are better at community-forming (and for that matter your suggestion that perhaps this is because MSM op-eds are dominated by right wing pundits, a claim that I agree with Blair is just plain wrong) . Those conclusions are not open from your data as far as I can see, at least in a general sense beyond the specific Hicks issue that you’ve plotted, and for the reasons I’ve canvassed at length on the Troppo thread.
I note a passing reference in Bruns’ mapping paper referring to subsequent exercises claimed to be confirmatory of the results of the Hicks cluster mapping. However, even if that’s the case, the significance of such results may well depend on the nature of the issue being studied and mapped (as I suggested). One cannot validly draw general conclusions about strength of “community” or “influence” on a comparative ideological basis (as you have done with varying degrees of stridency in various documents) without taking account of the extent to which the issue you’re mapping is likely to be one that is “owned” by or engages the attention of a particular part of the ideological spectrum.
In that sense I agree that it’s unavoidable to an extent to engage in some left-right dichotomising. I don’t have a problem with that either (unlike some other Troppo commenters) as long as you keep the limitations of that dichotomy always in view. I think you all to some extent, and Axel specifically in his most recent blog post, tended to lose sight of those limitations.
That said, I would be interested in seeing the results of a link mapping exercise of an issue putatively “owned” by the right (like the ones I suggested). However, even if it showed a less tight clustering among a relatively small group of Australian right-leaning blogs than was the case with the Hicks issue (as it may well), that would not allow a conclusion that the “right” blogosphere has a less developed sense of community. It may just mean that the “community” for those bloggers is more of a global village of the like-minded. It was the value-laden adjectival characterisations and assumptions in your article/s and posts that got Blair’s hackles up, I think, and I agree with him to some extent.
I think those general claims (influence and better developed community) are both dubious and counterproductive, given that further interpretation and pursuit of your research will require constructive interaction with inter alia the bloggers who have just been gratuitously sledged. I should, however, note that the really counterproductive blog post to which I reacted and posted about, was written by Axel Bruns and not yourself or Barry Saunders. I’m interested in the fact that Axel has not to date participated in this discussion. Does he accept that his post was at the very least ill-advised?
So, can we agree:
- You have no data at all to suggest that lefty blogs have any influence in Australian politics, let alone a lot of influence like Kos clearly has in the US.
- You have no data of any worth to suggest that lefty blogs tend to be more community-based than righty ones?
If we agreed on that this would go much more smoothly, I think.
Also, in response to this:
But how to explain the difference between the groups, if, as you say, they’re likely to be different?
I would say that you would explain it the way Ken has - by pointing out that on the net as in real life different communities form around different issues as a function of different preferences. Ie nothing to do with inherent predilections based on political preferences.
In relation to which, I actually thought that the academic research tended (with limited reliability imho) to show that lefties were more likely to be sad loners whilst righties were far more into community.
This is generally presented as: lefties are more likely to be independently-minded critical challengers of the status quo whilst righties are more likely to be conformist sheep, but the gist is the same regardless of which lens you use.
You are not changing lenses, you are positing a completely different thesis. in relation this research see eg Tyler Cowen.
As always Ken, you’ve made some fair points. Not sure how Axel feels about his post - I’ll leave that one to him. I, though, am grateful for the way you’ve pushed us on this stuff, and I’m thinking more clearly about these issues as a result.
One thing that it’s made clear for me is that one must be careful about the connections one makes between research and other forms of writing that one does. You’re right to point out that we ought perhaps to have taken greater care in all of this. I think this issue is particularly challenging for us because so much that we do is mixed up with the area we’re actually studying. But I’ll certainly be keeping it in mind in future.
The main reason I wrote this long post was to clarify some of the underlying questions that animate the research. I think that you’re in sympathy with most of the questions, and elements of the methods we’re using to address them, but think we’ve drawn some hasty conclusions at times. I’m prepared to own that, but I’ll also say that the conversation on Troppo has revealed how deep some of the difficulties are with answering the questions we’re interested in - something else I tried to address in this post.
Anyway, once again, I hope you’ll continue to follow the research with interest. Thanks for your positive comments on its value.
The big point that has occurred to me, which I mentioned on the troppo thread but you haven’t taken up so far on this one, was that it may well be that the interaction of feed reader subscriptions and comment box participation is much more critical to “community” formation than hyperlinks in primary posts. It would no doubt be a lot more difficult to research than link mapping (inter alia because you would need to assess not just the quantity but the quality of comment box interaction - in terms of comments engaging with and being responsive to each other rather than just abusive trolls or obsessive barrow-pushers shouting past each other in the dark), but I suspect it would be a lot more fertile and meaningful in the long run.
For example, to what extent do commenters (and bloggers) engage with each other mostly as part of an exercise in confirmation bias? Is there any significant Habermas-ian “communicative rationality” element? Does formation and maintenance of strong “community” necessarily exclude or place major limits on widely differing values/ideologies among group members, in cyberspace as in meatspace?
Feed reader subscriptions are a big thing, and certainly an important measure of readership at least. Who subscribes to which feeds, and what subscribers have in common with each other and with the blogs/bloggers they read is also important, but as you say, difficult to assess without some large-scale audience research. That’s certainly something that we can think about, though it’s not on the horizon right now.
Of course, readers come out in the open when they link from their own blogs, or when they comment on blogs. Links… well we’ve covered that in enough depth for the moment. Comments, though, we could approach in various ways.
We could use the lens of individual blogs, take some sample of their active life, and look at the comments in relation to a series of posts. If you wanted to do simple content analysis, you could mark out a matrix including factors like ’supportive’, ‘critical’, ‘unrelated’ ‘reply to poster’ ‘reply to commenter’ etc. etc. You could also mark it by username etc. You could then even think about categorising it according to ite apparent ideological bias, but that’s when you’d risk the charge of some arbitrariness. Anyway, looked at from various angles (time, by post, by topic etc.) that might provide you with some kind of overview…
There are other things we’re working with like Leximancer, which actually analyses the frequency and concurrence of particular words and phrases across individual blogs, and between them, too. You never know, this kind of analysis might even address some of Laura’s criticisms - it can easily be applied to blogs that aren’t linky, but possibly show how their stuff has been picked up across other blogs.
There are other, probably more satisfying and finer-grained things you could do alongside that. Interviews with the bloggers themselves, and even with ’super-contributors’ like high-frequency commenters will always tell you a lot. But using your eyes and reading specific threads to find out how conversations develop is another thing. It’s hard to “write up” in any simple way, and it draws on the humanities skills of textual analysis etc…
All of this is do-able, and I think that we all see it as a logical next step. I’m jut happy that we have the opportunity to do this stuff at all.
Hi all,
just a quick comment from me by way of follow-up - mainly on Ken’s comment. Ken’s right, of course - I haven’t as yet responded in any detail to some of the comments and criticisms raised by him and others. Largely, that’s simply to do with the fact that we’re in the middle of week one of the new semester - and as the only one of us three who also has active teaching duties this year, this requires my full attention for the moment. I hope to follow up in some more detail next week or so (and in fact I’ve been planning to post something here on the benefits and limitations of the network mapping approach for some time).
For now, a few things that may be worth pointing out in reply to Ken’s comments: yes, as I mentioned I’ve done a number of follow-up crawls on various issues after the David Hicks study, on topics which could be considered as ‘owned’ by both sides of politics. In March 2007, for example, I looked at the Brian Burke controversy (an issue which at that time was driven very strongly by commentators on the right, but ultimately backfired); from March through to May, I did weekly crawls on WorkChoices (which I would say was hotly discussed on both sides of the fence).
I haven’t written articles about all of these (and can’t release some of the articles I have written before the print publications they’re in are published), but as I mentioned in my post on my own blog, the network maps resulting from these crawls can all be viewed on the IssueCrawler site, and I’d encourage people to view and interpret them for themselves. (You will have to work through the less than optimal IssueCrawler Web interface, though.)
Obviously, the ‘ownership’ of each issue that is selected for a case study necessarily skews the results; across the issues I studied, I can’t say I found major differences in my results, however (and I would consider at least the Burke controversy to be owned more by the right than the left at that point in time - others may disagree). There is perhaps an overarching problem in that (as Possum documented in his analysis of the Crosby-Textor materials a few months ago) during much of 2007, there were very few issues solidly owned by the right - and some of them didn’t lend themselves easily to the type of analysis we’re talking about (issuecrawling approaches require relatively distinct keywords, so terms such as ‘economy’ or ‘experience’ don’t work particularly well, while ‘David Hicks’ or ‘Brian Burke’ do). The change of government may provide us with an opportunity to see whether there will be a role reversal in the blogosphere as well. There are a few other questions around the network mapping methodology, of course (and I do raise some of them in the article published in First Monday), but I’ll address them in a more substantial post later. Jason has already indicated some possible solutions and extensions of this approach which add a qualitative dimension, anyway.
A couple of other points which I would like to make here, though: first, while rigour and accountability are important in each case, I would have thought everyone understood the differences in register between academic publications (such as my First Monday article on the David Hicks issue), opinion pieces (such as our Club Bloggery series for ABC Online), and posts on group and personal blogs. I don’t think there’s much wrong with being a little polemic in an opinion piece, and even more so in a personal blog post - and I would think that at least in our collectively authored series, we have sufficiently identified statements of opinion _as_ opinion (where we haven’t, readers of these pieces have disagreed, sometimes vocally). The lack of distinction on some comments here and elsewhere between our academic work (and the scholarly publications resulting from it), our Club Bloggery series on ABC Online, this group blog, and the post on my personal site which reignited this debate, has genuinely surprised me.
Second, then, to answer Ken’s final question - was my post on Snurblog ill-advised? In hindsight, yes, probably - ignoring the flaming, sleeping on it, and outlining more clearly the research upon which our views are based (and its limitations) would have been the smarter choice. That said, I don’t think I’ll be the first or last to respond in kind to being provoked - I think it’s been, oh, some 20 years since I was last involved in a flamewar, and I didn’t recognise the flamebait for what it was. To be honest, I don’t think it will much change what I say on Snurblog, and how I say it, but I’ll certainly be more careful in pointing out whether I speak on my own behalf, on behalf of us three, or on behalf of the wider research teams and organisations I belong to.
hi, this message chek for this forum, ya ya!
hi, this message chek for this forum, ya ya!
Hello my friends
