An excellent post by my colleague John Nielsen-Gammon, the Texas State climatologist, can be found here. An excerpt:

… consider a pot of water. Somebody you don’t know claims that they have invented a new way to heat water. They demonstrate this technique to you by placing the pot of water in a separate room and showing you a temperature readout that indicates a rising temperature. A true skeptic will not immediately believe that the temperature readout is coming from a measurement of the temperature of the pot without some corroborating visual indication of steam or the like. A skeptic will also not believe that the pot is being heated by some new, exotic technique until he or she has been convinced that it is not being heated by any conventional technique.

This description may fit most of us, and I assert that it also fits the typical global warming skeptic by analogy.

Now consider a different scenario. Somebody you don’t know claims that they are about to heat of pot of water. You watch them place the pot of water on a stove and turn the burner on. You watch the burner turn red. You can even feel the heat coming from the burner. If you are a skeptic, you may still have some doubt that the pot of water is warming, at least until you see steam start to rise from it or until you can dip your finger in it to measure its temperature, but you will proceed under the assumption that the pot is warming. Indeed, if you have no information whatsoever regarding the temperature of the water, you will probably be much less likely to test the water’s temperature by dipping your finger in it the longer the pot remains on the hot burner.

In this scenario, your null hypothesis has shifted. Because you can see that heat is being input to the pot, your baseline assumption is that the water is becoming hotter, and you would require affirmative evidence to the contrary before you would be willing to believe that the temperature of the water is not actually changing.