The discussion over the definition of blogging is as old as the practice itself. For some all a blog is is a publication mechanism - thus any use of that mechanism is blogging; for others it is a certain publication and interaction behaviour through the web. One aspect of the application of social media infrastructure that I'm becoming more aware of is the level of engagement. For example, a typical blogger may write posts that link to other bloggers, and is likely to follow up with comments posted on their own blog. In addition, such a blogger may well respond to posts that link to their blog via the comments on that other blog or via posts on their own blog. Such an individual is engaged in the blogosphere.
At the other extreme, we have those who write blog posts that never link to other bloggers and, though they may receive a large number of comments, don't respond to these comments via their own commenting system. Such an individual is, we might say, a non-engaged blogger. Another example of this being the tweeter who has plenty of followers but who never issues an @'d tweet.
While the definition of blogging may still be in debate, the behaviours above can certainly be determined from pretty clear signals automatically. I'm guessing that someone has already done this analysis - anyone know of a paper?
An area of social media research that this measure has impact on is social network analysis. Typically, when inducing a social network from blog data, researchers look for reciprocal links. However, many political bloggers, while being of the non-engaged type, catalyze discussion in other blogs, or even simply within the many comments that each of their posts receive. Thus, one might argue, the simple notion of a tie between nodes should be abandoned for a model that can capture the different types of behaviour precipitated by different types of applications of social media publication technology.
I've long be suspicious of the wholesale adoption of real world social network analytics applied to social media, and blogging in particular (just as I am skeptical of the use of terms like 'conversation' when applied to this data). The above ideas, to me, seem to capture something of the reason for this discomfort.
