My advice on this is to resign when:
1) You believe your position is lost and
2) You believe your opponent sees and is skilled enough to convert the win and
3) You believe your opponent can win without difficulty and
4) You don't believe you can learn anything from watching the winning technique.
Do you guys think one learns more by continuing a losing match when things would depend on an unlikely series of blunders from the opponent, or is it a waste of time, you learn more by starting a new match? It doesn't preclude studying the resigned match to try to figure you more crucial mistakes, which may be the best way you spend your time in that match, rather than continuing it.
I'm inclined to think that resigning is better, but I may be wrong on that. What type of learning it could offer?
I think in a way it's just like a series of tactical exercises, but I wonder if it's the type of exercise that would prepare you for situations with a winning potential. The doubt regarding that is that it seems that much of the pattern-assimilation depends on a "holistic" view of the board, which may be fundamentally different on a winning versus losing context, even if tactic procedure is exactly the same.
It also seems intuitive that if we learn by repetition, the time we'd be using on losing games would have better use on more repetitions of the parts we get right.
There's also the question of ettiquette, addressed in another topic. When you're loosing just too bad is perhaps very annoying to drag it for too long, specially on correspondence chess. I guess it concerns much more correspondence chess than regular matches or even blitze, but I think it can start to get annoying even in a 10m blitz. Not so much a big deal if it's not a complete waste of time to continue, though.