Comments on: Cross-ideological conversations among bloggers

By: jorge schaulsohn

jorge schaulsohn — Sat, 28 May 2005 21:35:00 +0000

In my experience people with different persectives do not interact very much when it comes to politics. In a familiy reunion, for example we tend to ignore conflictive issues of politics or values. Same thing at the work place. Who do we talk about this subjects? Friends for the most part and with them we have a tendency to share a basic common ground. Why should blogging be any different?

By: Josh

Josh — Sat, 28 May 2005 15:41:28 +0000

Eszter—
I won’t be making it to ICA this weekend so can’t ask—have you found a statistically significant difference between conservatives and liberals at all? The E-I ratios seem to have a little bit of difference.

I’m curious as to whether there is a difference based on belief/opinion, or just whether those who hold a belief, whatever it may be, tend to stick to those who agree with them.

Thanks.

By: Ginny

Ginny — Sat, 28 May 2005 07:29:00 +0000

An aside: while political blogs make clear the politics of the writer, the voice comes out with unclear gender, race, age, ethnicity, religion, place of origin, current address, education, etc. While the right and the left do group themselves to some extent, those groups also dimiish other factional divisions.

By: Allen K.

Allen K. — Fri, 27 May 2005 17:52:00 +0000

How was the graph embedding on the right created? Was it by using the first two eigenvectors of the Laplacian of the graph?

(How odd to have a Post button with no Preview. I haven’t been here before.)

By: Seth Finkelstein

Seth Finkelstein — Fri, 27 May 2005 17:08:31 +0000

What he said, here.

By: murky

murky — Fri, 27 May 2005 16:58:36 +0000

The trackback links on Typepad sites multiply a visitor’s link-out options for any given post. I wonder what effect that has on the diversity of sites a person frequents. I imagine Typepad counts trackback link clicks.

By: Michael Yuri

Michael Yuri — Fri, 27 May 2005 14:57:06 +0000

I have a few questions about the study—
1. Suppose liberal blogger A posts something, and conservative blogger W responds with a detailed counterargument which thoroughly engages with A’s post. Then conservative bloggers X, Y, and Z link to W, saying “Take a look at W’s refutation of A.” Further, A could respond to W in a substantive way, after which liberals B, C, and D link to A’s reply.

As I understand it, this gives us six links that will count in your study as interlinking among same-ideology blogs, while only two links count as crosslinking. But in some sense, those six interlinks were hitchhiking on the crosslinks, pointing their readers to a real cross-ideology dialogue.

In this way, it seems like a simple link count could underestimate the amount of crosspartisan interaction that’s actually going on.

2. On the other hand, it seems to me that limiting the study to the top 20 blogs on each side could underestimate the divide between the two sides.

The top liberal bloggers likely read a number of the top conservative blogs. Similarly, the top conservative bloggers read a number of the top liberals. However, it would not surprise me if the further down a blogroll you go (in terms of traffic) the more one sided the list becomes.

Of the very popular blogs that a conservative blogger reads (and therefore links to), there is a decent representation of liberals. However, when you look at the more obscure blogs that a conservative reads, they may be much more skewed to the right.

If the study ignores any links that aren’t to the top 20 on each side, it may skew the results more in the direction of parity. I suppose if this has a significant effect it may be visible in the 120 blog study.

I’m interested in your thoughts on these issues.

By: eudoxis

eudoxis — Fri, 27 May 2005 14:39:27 +0000

What is the relationship between cross-linking and cross-ideological conversation?

By: Henry Woodbury

Henry Woodbury — Fri, 27 May 2005 13:32:47 +0000

To expand the analysis, I would be interested in how often bloggers of one political stripe link to news sites of, ostensibly, another (just tracking who links to which New York Times op-ed columnists, and why, could be pretty interesting).

A significant amount of blogging is the attempt to rebut a news story or editorial with which the blogger disagrees. This provides active readers with both sides of an argument, even if they are predisposed to agree with the rebuttal.

By: Ugh

Ugh — Fri, 27 May 2005 12:52:51 +0000

Sunstein’s book got absolutely ripped in a review titled philippic.com in the California Law Review. It made Sunstein look quite foolish.

By: b

Thu, 26 May 2005 22:48:19 +0000

I wonder about the premise of Sunstein’s argument (though I haven’t read the book). I think we’re being a bit too presentist. Throughout American history, have most voting citizens really had access to opposing viewpoints? I’d say this was perhaps true since the early 20th c., but not before.

By: Keith, Indy

Keith, Indy — Thu, 26 May 2005 22:03:32 +0000

It’s a small very diverse world, this web world.

I came here via Michelle Malkin, who mentioned a thread over at Freakonomics, who has a thread on “A Freakonomics Roundtable”…

To me this is the value of interlinking. Finding new sources of information. I’m now curious enough to pickup the Freakonomics book. And this interlinking study has me intrigued enough to keep track of this website. I think it’s a good first cut at measuring what has only been guessed at before.

Hi thibaud – I know that name from Roger Simons blog… See what I mean about a small world.

By: thibaud

thibaud — Thu, 26 May 2005 20:27:06 +0000

What Chris Clarke said. Another issue is that your model seems to have brushed aside the crucial question of how and to what end people use political blogs.

I’d posit that people who read them regularly derive several benefits, including two that most likely decrease political polarization:

1) encyclopedic: basic news and info drawn from a wider range of information sources than can be found on the reader’s preferred legacy media sites. Especially true of blogs that focus on locally-sourced international news, also true of more idiosyncratic blogs that cover non-political topics of the blogger’s choosing (e.g. Reynolds and commercial space travel).

2) entertainment value: amusement via short, snappy soundbites and put-downs;

3) the virtual pub: shooting the breeze, seeing old friends, enjoying the human spectacle as it passes by etc

4) solidarity in the struggle: camaraderie and affirmation with fellow partisans

For those readers seeking # 4 solidarity, and to a lesser extent, #2 artful put downs, the blogosphere may increase polarization, though it’s equally likely that their readers were completely polarized to begin with.

I’d suspect that a far, far higher number of visitors is more keen on benefits # 1 and # 3, the encyclopedic and virtual pub explanations. Most intelligent people are not anywhere near so political or partisan as the readers of CT or Powerline or Kos. They do not turn to the web for solidarity or for ammunition against their enemies. For these folks, a blog is worth returning to if it broadens their world or offers specific, useful information of a sort they don’t find anywhere else.

By: Chris Clarke

Chris Clarke — Thu, 26 May 2005 19:16:39 +0000

For me, the interesting doubt - I don't know that I'd go so far as to call it a concern - lies not in how one divides up the category "political bloggers" into neat packages but in definition of the category "political bloggers" itself. Would my blog qualify? Or PZ Myers'? Or Rana's? I certainly understand the need to limit one's sample, and I think the initial group of blogs chosen is a fine sample. But is the intent to determine the effect of linking on the political thought of people who blog only about politics? Or are you interested in us folks who may have very strong political beliefs indeed, but who also ~~have lives~~ write about other subjects as well? Because, you know, there are more bloggers like Rana than there are like Atrios or Insty. And trends that take place only among the putative trendsetters are no trends at all.

By: David Foster

David Foster — Thu, 26 May 2005 18:36:28 +0000

Interesting and valuable work. I second the concerns about the simple liberal vs conservative dichotomy. An alternative might be to use the typology recently developed by Pew Research, in which the categories actually emerged from the data, using some form of cluster analysis.