How to perform science without using "methodological naturalism."

Speaking of which, this is a favorite: