I like the citation from 1997 they put up as proof that the raw number of times innocuous words were tweeted is proof of misogyny.

Taking as its starting point the category text (as opposed to clause, collocation span, sentence etc.) this paper proposes and illustrates a method of identifying key words in text, and leads from this to the notion of key key words (words that are key in many texts). A key key word is shown to have associates: words that are key in the same texts as a given key key word. Finally, these associates can be grouped together in clumps, which themselves are revealing about text schemata and stereotype. The number of clumps in which any key key word participates relates not to ambiguity but to a diversity of perceived roles; the members of the clump describe and illuminate these stereotypical roles.

Translation - we think there are keywords that the very presence of, regardless of context, indicate a certain stereotype is believed in. Sometimes these words don't actually correlate with the stereotype we say they do, so we call those key-key words and rather than admit our methods are ambiguous and unreliable we'll call them diverse because we just fucking can, and now they're diverse no-one can criticize what we say these vulnerable words of color mean.