Over the course of the last year or so, I’ve been looking at the way people ask and answer questions on Facebook. Much of this work happened with the phenomenal Merrie Morris and Jaime Teevan (haystack alum!) at Microsoft Research.
I’ve been interested in the ad hoc way people ask questions as their status messages (not using the Facebook Questions app), and how it signals some unmet needs in the search (and perhaps social) space.
We’ve identified the types and topics of questions (people ask recommendation, opinion, and factual knowledge questions about things like technology, entertainment, and home & family), motivations for asking questions (in order: trust, subjective questions, belief a search would fail, specific audience, connecting socially, and faster answers), how phrasing affects answers (the more explicitly you ask, and the more you scope your answer to a specific group or eve anyone, the more and better responses you get), and how asking questions in status messages compares to searching (it’s not as fast, but it’s a great supplement).
Lately, I’ve been looking at who answers facebook questions. For instance, in the question I posted above, how close are these friends?
The thing that prompted this direction was work done by famed sociologist Mark Granovetter. He’s known, among other things, for his seminal work on the strength of weak ties, which identified that a lot of useful information came from weak ties, acquaintances, rather than strong ties, or close friends. I’ve been interested in finding out if that’s the case when asking Facebook questions as well.
Thus far, I’ve done a very interesting pilot study – results from the study indicate that, instead of weak ties giving you more useful and helpful responses, strong ties do. We’re currently in the process of testing this hypothesis with a bigger and better study – please email me if you’re local and are interested in participating.
One of the interesting topics in designing this has been understanding what makes an answer to a question useful or helpful – in some of our prior work, people thought answers that were unconstructive were helpful, even though they don’t seem to be on the surface. An example of answers like this for the question I linked to above would be something like “Oh Joanna Newsom! Cool!” – it’s supportive, and on-topic, but certainly not an answer to the question I asked.
We’ve thought about whether or not you might have seen the answer before as a marker of helpfulness, how on topic it is, how much you trust it, how careful you’ll be about verifying it, and other things like that. What makes an answer useful or helpful to you?