A few times I have been asked to give a talk to students and rather than just perform magic for them, I took one trick and — without offering any explanation — ask them how they would go about trying to figure out how it was done. It wasn't about the answers, it was about asking the right questions.
Gradually, they realized they would want to have different students watching from different places, that would want to videotape it to watch it multiple times, possibly in slow motion, and most importantly, to have me perform it again under a variety of different conditions. (Can we shuffle the cards? Can we choose a different one? Can we bring the deck of cards?)
Essentially the students re-invented the scientific method on the spot.
Along the way, they had offered up suggestions as to how the trick might have been accomplished. After all, in science you need a hypothesis to test. One of them was most of the way to being right, although I never told her at the time.
Recently Michael Shermer, writing for Scientific American, recently took the same approach to a more everyday topic: the resurrection of Jesus. How do you evaluate the truth of claims you can't investigate for yourself? If you're a skeptic, how can you know anything is true.
By way of introduction:
When I was speaking to the students, I eventually moved on to the more difficult question:
If this were accomplished by magic (instead of some sort of trick), how would you know?
Most people don't know that it's impossible. You can't use a lack of an explanation to justify an explanation. The "I don't know, therefore magic" argument is invalid. Shermer's full piece does a fantastic job of explaining why.