Here's a little blurb from the Atlantic on how eliminating comments at the National Journal improved site traffic by a considerable margin. The author did a little, let's call it an experiment that seemed to show that reading negative comments associated with a post would leave the reader with a more negative impression of the story's quality than just reading the story alone.
In their particular exercise the comments didn't seem to have an impact on mood, but over here at Gawker personal anecdata skews to the contrary sometimes, so I'd be interested to learn more about under what conditions they might. Back when they were appearing three and four a day, Zimmerman posts and associated commentary, for instance, seemed to have a palpable impact on mood for a number of us.
It brought back memories of all Denton's kvetching about how negativity in the comments at Gawker impacted perception of the site for the worse (and hurt the writers' feelings). At the time it seemed like the obvious solution was, well, write better articles then, and the commenters will like them, QED, but we see how that panned out (he hired Neetzan lol jk brb). For a while they were doing that "zero tolerance for criticism of Gawker" thing which was apparently a John Cook-led initiative to ban anyone who ragged on the author of a post—and that also seems to have fallen by the wayside. I guess the current strategy at Gawker is to host comments that are SO MORONICALLY, DISTRACTINGLY STUPID you completely forgot you even read an article.
Of course that made me curious about all the different ways the comments associated with a story might have the potential to impact the reader's takeaway. I want to see some comparisons of the impressions readers tend to form of posts with favorable comments, those with interesting discussions, and other sorts of interactions to see how each might affect their responses to the content.
Something that goes unaddressed by the article but has maybe been studied elsewhere is the question of how does actually commenting on something affect one's reaction to the content? I assume content generators would care less about that since it's always going to be a very small slice of people that actually comment, but just from a learning perspective how does your participation enhance (or detract from?) your experience of a story?
When you read an absurd and factually inaccurate piece of crap, for instance, are you better off composing a thoughtful correction or just going on your way?