This blog post about an article shows why you need a theory. You can't just assume a particular correlation and then interpret your data naively. without looking at interpretative contexts. In this case, what the authors of the study were missing is a generic marker of lyric poetry.
Here, theory does not mean any particular theory, but a way of looking at things with more of a consciousness of what interpretative framework you are going to use. Some really shoddy social science work just assumes a naive empiricism and has at it.