Being a doctrinaire frequentist (ie., non-Bayesian), I nevertheless found a very interesting piece about the potential danger of (over)relying on p-values to make statistical conclusions:
Go here: http://www.causeweb.org/wiki/chance/...Chance_News_71
Page down a bit to the section "Bayesians and Bem's ESP Paper" which shows a great example of how a p-value could be misleading.
As an aside, there's also this link to a great little book called "What is a p-value anyway?"
If you're really interested, you can subscribe to the newsletter -- there's usually some great stuff to read and think about.