Here is a simple test for that separate section lots of people (including me) want to include in empirical articles, but doesn't fit in lit review, theory, methods, results, or conclusions. If it could have been written unchanged regardless of the results, it belongs elsewhere. This one is classic.
