If the two groups (Alzheimer's and non-Alzheimers) were not controlled for all other variables (for instance, age) then the results are meaningless.
This is the problem called "hidden variables" or "confounding factors".
Sounds like this was at work in the coup and FISA warrants.