Unfortunately h2odev.law.harvard.edu, the machine that runs my media research scripts, was down for over a week – as a result, I’m very far behind on my research comparing international news in the blogosphere to news in the mainstream American media. I hope to have some solid results to post early next week. In the meantime, the scripts referenced in this post should be working, with the caveat that they have very little data to reference.
Research I’m not really ready to share yet… but hey, it’s Bloggercon and I’ve got to have something to show off…
For almost a year, I’ve been tracking the appearance of a set of keywords on some major media websites: BBC, CNN, Google News, etc. The keywords try (to the best of their little boolean abilities) to find international news stories that mention each of 187 countries. I’ve been using this to build maps of media attention, which you can find on my research site.
With the help of the good folks at Intelliseek, I’ve been polling the Blogpulse engine for the last week or so to build maps of international reporting in blogspace. Here’s today’s map:
Countries in red are getting a lot of mentions in the blogosphere – the darkest red represents countries that have 3.2% or more of the international news stories I’m retrieving. Blue countries are getting few mentions, down to the deep blue ones, which are generally seeing single mentions. This map measures the last two weeks as spidered by Blogpulse – I’m also running scripts on 90-day data sets – results of both are available online here and here.
The main reason I started building these maps was to test whether the blogosphere is more or less interested news from developing nations than the mainstream media. It’s a question I don’t have an answer to, but I’m starting to have some theories.
Much of the work I’ve done on attention compares actual distribution of stories to models which posit a relationship between distribution and an outside factor, like national GDP, population, surface area, etc. My research suggests that story distribution, on most media I’ve studied, is tightly correlated to GDP and loosely, if at all, correlated to national population. The result – rich, small nations get covered a lot more frequently than big, poor nations.
The blogpulse data I’ve got so far seems to fit a similar pattern – the data fits a GDP model pretty tightly (R^2=0.68 for the 90-day blogpulse data, as compared to an 0.70 fit for Google News data.) It fits a population model even less tightly than mainstream media (R^2=0.43 versus 0.48).
I’ve also built a tool that allows me to compare two media sources on a given day. (There’s a web interface to this tool – it’s an alpha, so don’t be surprised if it breaks often and in ugly ways.)
Here’s a map of the last two weeks of blogpulse, versus two weeks of Google news.
Countries in red were better represented in the blogosphere than in Google news; in blue, were less well represented. The maps suggests that, for the most part, folks in the blogosphere talk about Africa, Central Asia, Eastern Europe and parts of Latin America less than the media sources covered by Google News do.
I don’t claim to fully understand the data yet. I see some reasons for optimism, though – while Africa is, for the most part, a sea of blue, there are about three times as many mentions of Rwanda in the blogosphere, proportionately, as in Google News. One theory suggests that folks in the blogosphere may pick up and amplify timely conversations about the developing world – Rwanda has been getting lots of press attention because of the 10th anniversary of the genocide, and these conversations have extended into the world of blogs. Is it possible that when the mainstream media focuses on a developing nation in a big way that the blogosphere picks up the ball? I’ll be trying to answer that question by watching this sort of data closely.