Forgive me for rambling here a bit, but I'm still really trying to figure out when to do it and when not.
A few experiences at the $200 tables as background: There are some tables that play pretty tight, rarely actually super-tight PF, but, it you yourself are playing tight, it's very hard to get any action at all on big hands.
Then, there are these high action tables where it's really hard to say where people are at. Sample hand from last night: I re-raise AAds. Flop 973 with 2 spades (to which I have the nut draw as well as a 7). 4 callers to the re-raise. Guy pots it to me, I re-raise to almost all-in, he moves in with KJTs3s (great hand for calling a re-raise, huh?), turn a T, river a 3... So, I lose $250 on that one, although I think the play was right in that particular case--I'm actually pretty far ahead and playing for stack.
Anyhow, the tight tables are also rather trappy at times. One thing I normally like about betting out the draw is that if you get a caller, they're supposed to be on a draw rather than a made hand, so you can take down the pot if you both miss. But you have to at least be able to put them on the draw rather than a slowplayed top set (which I think is a bad play generally but it can sometimes be difficult to deal with).
Another aspect: If you get raised, do THEY have yet another draw? Then you have to decide whether to call, come over the top, or whatever. If the action is on the flop, coming over the top, I think, is often the way to go if your draw is pretty serious, because it gives you 2 cards to make it. Otherwise, it's almost impossible to see the river if your opponent bets aggressively.
Actually, I'm still at a bit of a loss to really formulate anything coherent about when or when not to bet draws. So, I guess I'll leave it for the moment at just a formulation of the problem.