On a recent thread about condensation forming over the wing on landing, one of the (now deleted) answers was entirely believable, was grounded in an accurate observation of physics, but happened to be simply wrong (cold fuel in the wings). Other answers and comments gave the correct answer and (somewhat) pointed out why this particular answer was not actually the correct mechanism behind the condensation observed in the original question.
To my mind, that's a good learning point for others -- to say that yes, the wing being cold DOES cause condensation when the aircraft is sitting, but it's not that but this other dynamic at work creating the visible condensation during approach and landing. Of course, that learning point is lost if the answer is deleted. (A good comment on why this particular mechanism wasn't at work would have been helpful for the learning process too -- more so, I'd have thought, than deleting the answer.)
What's the best way to balance out the desire to have accurate answers, with the learning that can occur when people can see a good thought that happens to be wrong in this case, coupled with the explanation of why "that" isn't actually correct?