I kind of feel sorry for them: humans make mistakes, and retrospective digging into any human endeavor is going to uncover them. That wouldn't be a problem if their marketing hadn't been "This is the infallible Voice of Science: the only people capable of understanding it all have already signed off on it. Implement our recommendations or be cast into a fiery pit." Mr. Wizard always looks less impressive from behind the curtain.
I have half a mind to take a week or two this summer, grab one of the old papers in the field and the associated data sets, and just try to reproduce the math. The reason I have only half a mind is that I know I'm likely to discover that the process is not covered in sufficient detail publicly to reproduce it.