This is a long-developing trend, but it gets a solid treatment from the neuroscience blog Mind Hacks:

Yesterday’s New York Times had a fascinating piece about online social networking tools, discussing how different forms of social relationships are being formed through the use of ‘broadcast to subscriber’ tools like Twitter and Facebook.

These articles pop up quite frequently, discussing how young people live in a ‘post-privacy’ world, or how our personal lives become increasingly public to our friends and acquaintances, but they rarely mention the ways in which these social networks can be used to reveal and exploit the dynamics of social power.

Sociology gets a bad rap in science as being ‘wooly’ or ‘vague’, but it’s often not to do with the methods its uses, but with the way of gathering data.

Using social networks for data-driven sociology raises interesting issues around self-selection. The only people who user twitter are…well, twitter users. That means that sociological insights drawn from twitter are really about…well, twitter users. It’s hard to say exactly what distinguishes twitter users from the population at large, but its certainly a distinct and probably relatively unique subset. How does a researcher generalize their work based on social media data so it applies broadly?

The question isn’t that far from the one business have to ask about such tools–how can they generalize insights based on consumers who use twitter so that they apply to consumers as a whole? Neither sociologists or marketers can assume that twitter represent everyone else–after all, some people use it and some don’t, so there must be some difference there. It’s an open question, I think, and one that the academic community might make progress on first…