Theories explain relations among data.
The word "fact" is only used colloquially among scientists doing science.
A scientific paper follows a formula: question, methods, procedures, results, conclusions.
I am becoming tempted to write a paper as follows: Do self-identified creationists use more or fewer ad hominem arguments than self-identified non-creationists when posting to Free Republic?
proposed method: Take three consecutive crevo threads with more than 10 responses. Count ad hominems (using a standard definition) Analyse both per thread and per poster so that one poster will not unduly influence results. In order to maximize objecivtity, have 2 each of self-identified creationists and non-creationsits identify the ad hominems. Results tallied separately.
Threads must be at random however!