Totally feasible, though bound to have wavering accuracy – since there are so many different types of Quora user. I'm actually working on this at the moment as an experiment myself… I think the results will be best for those with a large number of answers, though large number of question will also be useful, given the tagging system.
If you wanted to really nail this, though, you'd need a pretty large test group and data corpus, to help tease of the subtleties of edits and voting and personality type.
Is it feasible to do a MBTI (or big 5) study of Quora users?
deano.com 