shades: you're implying that your standard of 'fact' would ever be possible. It wouldn't. At the level of certainty you're asking for there would be very few facts.
Really? I disagree.
If an inductive reference from the observations is justified, it should meet three criteria:
1. The number of observations forming the generalization must be large
2. The observation must be made under a variety of conditions
3. No accepted observation statement should conflict with the "derived law" (or universal principle)
I would say that numbers 1 and 2 are the most important, and that two is missing in this case. You have a small sample from one tax office in one state of the US. Not a variety of conditions. I don't think that it is unreasonable to suggest that KFCs observations may not lead to the "fact" that Catholics give less.
It is possible to arrive at a false conclusion based on true observations. Take the story of the turkey often credited to Bertrand Russell. The turkey is feed at 9am. As days go by, each morning, the turkey is fed at the same time. The turkey concludes that each day he will be fed at the same time. However, on Christmas Eve, the turkey is killed. The turkey's conclusion (that he will be fed at 9am every day) is a false conclusion that was based on true observations.
All that said, KFC, I apologize for the extremely off topic hijack. If anyone is interested in this philosophy of science stuff, I'm reading "What is this Thing called Science" by A. F. Chalmers. It's interesting--if dense.
PS. KFC-I'm sure the law of averages comes in somewhere, but I haven't reached that chapter yet. I'll have to get back to you.