It looks like a lot of people here vote on questions not based on the question formulation itself, but based on the answers that are given. That means they do not vote until they see the answers, then they vote.

Is this correct?

If yes - how is this behaviour justified?

What I would think - question is estimated on the formulation itself - how clear is it? does it shows research effort? (These are the criterions you see when you put mouse over voting-arrows). Otherwise how would user, who ask (and **who doesn't know the answer**) can estimate what is good question and what is not?