I've learned a lot more about the economy from introspection than I have from statistics. If someone shows me statistical evidence that people buy more chocolate when its price goes up, my reaction will be "I've bought lots of chocolate, and I would buy less if the price went up. Wouldn't you? Come on."
But introspection, like statistics, has its problems. A case in point: When you're in a store, introspection typically reveals that you are not very price sensitive. If you go to the store to buy some carrots, how high does the store's price have to get before you cross carrots off your list? The usual answer is: Pretty high. If you repeat this exercise for your whole grocery list, you're likely to conclude that your demand curves are highly inelastic, and markets are a lot less competitive than most economists would have you believe.
The problem with this exercise: It assumes you're already at a particular store. Yes, if you get to a store and see high prices, you're probably pay. But are you going to return to that store? Or - let's introspect again - are you going to think "That store's a rip-off, I'll go somewhere else"? At least that's my reaction - how about yours?
The lesson: Just as the solution to sloppy statistics is better statistics, the solution to sloppy introspection is better introspection.