Personally, I don't mind lecturers "telling" the answers straight away. Obviously give motivation and lead into it rather than just delivering it with no introduction, but I've found a lot of problems are simply too large or too non-obvious to take a straw poll from the class.

The best teachers I've had have been:

(i) Really bloody good at their subject. My probability lecturer (the Field's medallist Tim Gowers), when presenting the proof of Stirling's approximation, said something along the lines of "now, the normal method for proving this bit is a bit complicated, and on my cycle in this morning I thought 'There must be an easier way', when the following struck me. Now, I've never written this down before, so hopefully it works...", and of course laid out a simplification of a famous proof that worked perfectly. Inspirational. (See http://gowers.wordpress.com/2008/02/01/removing-the-magic-fr...)

(ii) Incredible elucidators. The ability to take a complex concept and pick exactly the right words or metaphors to explain it is a powerful ability.

I feel these criteria are far more important than asking the audience questions. You need both -- I've been taught by plenty of talented mathematicians who had no idea how to teach, and certainly a handful of great teachers who simply didn't have the aptitude to really grasp the more complex topics.