Study: Commenters on CNN.com may call you an idiot; on msnbc.com they tell you why you’re an idiot

Tim Libert Writing
Tim Libert, a Web developer for NYU’s journalism department, conducted a detailed study of about 2,400 article comments posted on six news sites to learn how commenters view the hacking group LulzSec. In the process, he noticed some differences among the sites’ commenting communities. He looked at three general-interest news sites (CNN.com, msnbc.com and The New York Times) as well as three technology sites (The Register, based in the U.K., Wired.com and ZDNet). He discovered that commenters across five of the sites disapproved of LulzSec’s actions — except when the stories were about the hacking of the News Corp.-owned The Sun website. Some of Libert’s conclusions:

• CNN.com had the most comments on its stories, but also more “name-calling, off-topic comments and what I suspect were fake accounts (aka ‘sock puppets’),” as well as more violent comments. “For example, there were at least 8 different calls for cutting off hands or fingers of LulzSec members.”
• The lowest level of discourse among the sites was on Wired.com. The highest was on the Times.
• Conspiracy theories were prevalent on msnbc.com. “A number of people seemed to think that Rupert Murdoch, rather than LulzSec, had hacked the Sun website for some sort of poorly defined personal gain.”
• Comments on msnbc.com were much longer and written at a higher level than CNN.com. Libert attributed the difference in length “to the fact that on CNN somebody may call you an idiot, but on MSNBC they will take the time to tell you why you are an idiot.”
• On the Times site, comments “tended to be more nuanced, thoughtful and balanced. Calls for violence and rape found on CNN were virtually absent on the New York Times.”
• Another difference in comments on nytimes.com: “While comments on most other sites appeared to be people expressing already held beliefs, comments on the Times website gave the impression that the commenter had used critical reasoning when reading the article and were responding to specifics.” Fewer Times comments were off-topic or contained personal attacks.

Libert explains his methodology and notes the many caveats. For instance, moderation policies vary among different sites, and the number of comments varied among stories and sites. The point of Libert’s study was to judge perception of “techno-vigilantism,” so if that’s your thing, the whole article is worth a read.