May 25, 2004

The population of political blogs

Posted by Eszter

I resisted temptation for a while, but have finally launched into a project on blogs with two graduate students in our Media, Technology and Society PhD program, Jason Gallo and Sean Zehnder. We are focusing on political blogs in particular. This raises a whole set of methodological questions. A big one has to do with sampling. We have decided that we would not focus on a random sample of blogs, not only because that is just about impossible to achieve, but more importantly because that is not our focus. We are interested in the most widely read political blogs. (Yes, there remains the question as to what counts as a political blog in the first place, that is just one of the many questions we are grappling with.)

One way of finding prominent political blogs (or prominent blogs of any type) is to look through the links of prominent blogs we already know about. However, since linking is one of the questions we are interested in, it seems problematic to rely only on that method to find blogs relevant for our study. The same concern applies to using Technorati as a method for finding prominent political blogs. Another idea is to run searches on certain political topics and “blog” or “weblog” and see what we find. Of course, in such cases we are left wondering how widely read the particular blogs are, especially if they do not have comments turned on (and in any case, number of comments is a very limited measure of how widely read a blog may be). Other methods we have thought of is to look at directory listings (such as Yahoo!’s) of political blogs for ones we may have missed using the other methods.

So to sum up: What other approaches should we be using to identify political blogs? What methods do people recommend for identifying “top” political blogs?Also, if people know of political blogs that don’t seem to get mentioned here much, please feel free to post away. I realize this method of collecting information mirrors many of the shortcomings mentioned above, but hopefully by using all these approaches, we can get a reasonable sample (or dare I say population) of the most widely read political blogs. Thanks!

Posted on May 25, 2004 04:50 PM UTC
Comments

You could do a lexis-nexis search to find the blogs that receive the most print references.

This has the advantage of being data that is external both to the blogs themselves and their delivery medium (unlike counting links or looking at portal references).

It has the obvious disadvantage of reflecting only the blog-reading habits of a specific subgroup (ie, print journalists). You could argue either side of the question of whether this is a misrepresentative subgroup or a superrepresentative subgroup.

Posted by globecanvas · May 25, 2004 05:09 PM

You could use tools like Technorati or Lexis-Nexis to narrow your list of potential blogs down then contact the authors themselves for their site statistics.

Posted by mallarme · May 25, 2004 05:25 PM

I’d start with the “Truth Laid Bear” ecosystem Top 100. It ranks by links to, and also by traffic.

Start at the top and work your way down. Decide what counts as “political”.

Posted by Richard Bellamy · May 25, 2004 05:32 PM

There are, of course, a number of well-known indices-to-blogs. For example, The Truth Laid Bear tracks blogroll references between blogs: this gives you a fairly good popularity measure.

There are certainly errors in his measure (BlogShares is not a blog, for example), but there’s no question that his top ten blogs are certainly within the “top” twenty of all blogs.

http://www.truthlaidbear.com/ecosystem.php

I use the word “top” fairly loosely, though. I realize you’re asking a slightly different question, that of “most read”. That’s a tough term—I really don’t know how to operationalize it.

Clearly, hit count won’t cut it: I read CT with Bloglines via RSS. Which means that you record far fewer than one hit for me (because everyone on bloglines who reads CT shares the RSS hit); it means that you record too many hits (because Bloglines pings your feed fairly frequently).

Just as clearly, link-based methods don’t cut it, because the populations of “linkers” is different than the population of “readers.” As is, I suspect, the population of “journalists,” as globecanvas points out.

Posted by Danyel · May 25, 2004 05:34 PM

Maybe you could do a search for political weblogs with google, Eszter. Multitasking.

Posted by Motoko Kusanagi · May 25, 2004 05:58 PM

I think the main TTLB Ecosystem suffers from the same problem as Technorati — it’s based on numbers of links (and thus is mostly sensitive to how many bloggers read you). But the traffic rankings at TTLB seem like a good semi-independent measure of popularity.

Posted by Stentor · May 25, 2004 06:02 PM

I liked this political blogmap idea a lot, if it had grown and was updated, I guess it could be some use.

I agree on the Technorati suggestion, but you should also check Daypop, it lists the top blogs by citation which, combined with traffic ranking, is probably the best method to see which websites are more influential.

But why do you have to be so elitist and stick to the top ones? :)

I hope at least you won’t be nationalist and include non-US blogs too.

Posted by pepi · May 25, 2004 06:22 PM

Maybe you could take into consideration also the Guardian picks.

Posted by pepi · May 25, 2004 06:30 PM

Thanks for the suggestions. Just fyi, the TTLB traffic rankings seem to depend on sites using SiteMeter, which means that a bunch of quite popular blogs are not listed. Nonetheless, both that and the Ecosystem are helpful additional sources.

Re Mallarme’s suggestion about contacting bloggers, for now we’re staying away from getting in touch with people. We may explore that later in the project, but not right now.

Good idea about Lexis Nexis as well.

The decision to include the more popular blogs is not about elitism, but about the questions we are asking. And to be sure, by “most popular”, we don’t just mean ones read by tens of thousands of people per se. Also, regarding US vs non-US blogs, this also has to do with our research questions. We are pretty sure we won’t be restricting the group of blogs to US blogs, but for now we are mainly interested in commentary about US politics. This could be expanded. Like many other projects, the more we think about it the more possibilities arise.

We welcome additional suggestions!

Posted by eszter · May 25, 2004 06:47 PM

>Also, if people know of political blogs
>that don’t seem to get mentioned
>here much, please feel free to post away

An excuse for a shameless plug I can’t resist: perfect.co.uk">perfect.co.uk

Posted by Robin · May 25, 2004 06:54 PM

Google: “Bush Lies” and “Kerry Flip-Flops”.

How many do you want,dozens,hundreds? It wasn’t clear. Do you want non-American political blogs?
Political Theory blogs? Political News Aggregators? Partisan chat-rooms?

Posted by bob mcmanus · May 25, 2004 06:56 PM

Good questions, Bob. We are not setting a limit on the number right now, first we want to see what we have. We’re interested in political commentary for the most part about US politics (for the most part).

Posted by eszter · May 25, 2004 07:00 PM

How are you defining “political blog”? I ask, because I know a number of bloggers who regularly post on political topics, but either they do not do so exclusively, or the topics are political but not in the form of “breaking news.” (e.g. feministe)

Posted by Rana · May 25, 2004 07:02 PM

As I mentioned, the definition of what counts as a political blog is also up for discussion. It is certainly not restricted to blogs discussing “breaking news” only, nor is it restricted to blogs that only talk about political topics. So, for example, even though there’s much non-political content on CT, CT would still count because there are plenty of posts having to do with political issues. We haven’t yet decided what proportion of posts would have to do with political topics (again, broadly defined) in order for the blog to count as a political blog for our purposes.

Posted by eszter · May 25, 2004 07:44 PM

Have you tried searching on vivismo? Their clustering methodolgy may help you a bit.

Posted by decon · May 25, 2004 07:58 PM

1)

Posted by David Weman · May 25, 2004 08:59 PM

I don’t now much about their methodology, but Blogstreet has a “100 Most Influential” list. That would probably be another good place to start the weeding process.

Posted by kevin · May 25, 2004 09:56 PM

Maybe you could ask people to nominate their “top ten” political blogs, and cull from that group?

Posted by Rana · May 26, 2004 12:11 AM

Separate the two parts of the problem: (1) find large numbers of blogs with claims to be political (2) decide on rules for deciding what counts as “political” for your purposes, and why, and then worry about how to count popularity, probably looking at several methods.

If you initially aim for comprehensiveness, you’ll keep looking, using all these suggested techniques, until eventually you realize you’re only getting pointed to blogs you’ve already counted. Then with the universe covered, you can figure out how to decide which ones are political. And if you keep track of the myriad referrals, from people, email, links, blogrolls, etc. then you’ll have a lot of information on popularity already, and can do other things like ask the owners for additional data. It doesn’t distinguish hypothesis formation from data collection, but as pioneers your first task is defining the problem and universe to examine.

To limit the project, I suggest choosing a time-slice, so that at some point you’re open to taking retrospective data without counting every new blog added each day. And of course, keep us posted periodically please!

Posted by vivian · May 26, 2004 12:30 AM

You could try this Alexa page or others - Alexa claims to estimate traffic by measuring the behaviour of people who use their add-on software (it reports - benignly - back to HQ).

Of course it isn’t rigorous but it’s the only source of public traffic data I know of that measures the traffic of low traffic sites. Though you could ask people like Comscore if they have any data.

On a similar subject, how would people try to sample the much broader field of home pages and weblogs in the UK - to be more precise “sites that are not primarily in furtherance of professional goals (eg online CVs, galleries of art from artists etc), are not explicitly temporary, are substantially the work of a single individual, and are not closed to the public either explicitly (through a password) or implicitly (for example collections of photos from an event without an accompanying narrative that are only meant to be accessed by a small group for a short time even if they are openly available online).”

I thought about sampling randomly from directories compiled by Geocities or Freeserve/Wanadoo but I looked and it seems they no longer index their pages. Using Yahoo or DMoz would introduce obvious biases because submission is not automatic. Tripod still does have directories of its UK users but how representative would Tripod users be of all users? Searching for “personal home page uk” in Google gets me nowhere.

How should I balance blogs with home pages? Using the stats from Pew suggests I should include about one blog for every four home pages…

Posted by David Brake · May 26, 2004 10:33 AM

While I think about it - are you tacitly wanting commentary about US politics by Americans? If so how do you plan to separate them from weblogs from elsewhere commenting on US politics?

P.S. There used to be a master directory of Blogger weblogs. Is there still? I couldn’t find it on their site.

Is there any up to date info on the relative popularity of the various weblogging platforms?

Posted by David Brake · May 26, 2004 10:58 AM

what is your research question?

Posted by david · May 26, 2004 04:00 PM

Alexas is decent for somethings, but it doesn’t help at all with blogger blogs.

Posted by decon · May 26, 2004 11:06 PM

Take a look at BlogLines, which is a terrific web-based blog reader service. It has a lot of user info that’s shared with other users. It’s mostly reflecting subscriptions to RSS feeds (as opposed to traffic measured from the blog’s perspective by hits or links). I’d bet the founder, Mark Fletcher, would be quite interested in discussing your research program. His personal blog is wingedpig.com.

Posted by Claudia M · May 28, 2004 12:58 AM

In connection with the point about how much # of comments means, consider this absurdity:

http://www.laze.net/fait/archive/2002/07/28/maurys_blooper.php

Doesn’t seem that even the people who leave comments read that website!

Posted by w · May 28, 2004 01:22 AM

This Memorial Day weekend’s sudden attention to the letter by Sharon Underwood about her gay son would make a good test case for following a meme around the echo chamber. I’ve done a little playing with this idea at http://home.pacbell.net/mabjo/martha.html#490 .

Posted by Martha Bridegam · June 1, 2004 12:51 AM

Some blogs that are less widely read are relatively influential. Check out Blogstreet, which weights blogs according to the prestige of those who link to them.

http://blogstreet.com/

Posted by Ileana · June 1, 2004 04:17 AM
Followups

→ Which blogs are popular with the media?.
Excerpt: Nothing spurs forward progress in research like competition. First Henry Copeland has his blog survey. Now I read that Eszter Hargittai is starting her own project on blogs and the media, and she's looking for a "way of finding prominent...Read more at Daniel W. Drezner
→ Which blogs are read by the media?.
Excerpt: Nothing spurs forward progress in research like competition. First Henry Copeland has his blog survey. Now I read that Eszter Hargittai is starting her own project on blogs and the media, and she's looking for a "way of finding prominent...Read more at Daniel W. Drezner

This discussion has been closed. Thanks to everyone who contributed.