Tautology is not scientifically valid, yet I see examples of tautological argumentation nearly every day in my field's scientific journals.
Here is an abstracted example to help you think this out:
Scientist has some evidence to suggest that most As are also Bs. C is invented as a way to describe an attribute of A. Scientist then concludes that most having condition C are also B, so condition C becomes a way to diagnose B and spurious, oftentimes ideological connections are assumed to exist between all three.
Tags: