The chart on “quality of question” vs “strength of evidence” in Roger Peng’s blog post about the goals of data science works pretty well for humanities questions too.