I see the author of this interesting site is active in this thread. You may already know about this, but for onlookers I will mention that Uri Simonsohn and his

I see the author of this interesting site is active in this thread. You may already know about this, but for onlookers I will mention that Uri Simonsohn and his colleagues http://opim.wharton.upenn.edu/~uws/ http://opim.wharton.upenn.edu/~uws/ have published a lot of interesting papers advising psychology researchers how to avoid statistical errors (and also how to detect statistical errors, up to and including fraud, by using statistical techniques on published data).