> why these criteria are nonsensical I don't conduct surveys or other research so it's not immediately obvious to me. I would guess that it's some combination

> why these criteria are nonsensical I don't conduct surveys or other research so it's not immediately obvious to me. I would guess that it's some combination of: 1) Humans making subjective calls like "generally contain claims to support" (and even just the subjectivity of the word "support"); 2) Most opinionated users (read: people who reply 1,000 times) on social media are going to "entirely, or almost entirely support one candidate"; 3) Most people are going to post things which "generally contain claims to support their arguments". But I'm not really sure if those are (to a researcher who ought to know) the obvious problems with these criteria. Is there another problem that is seen which is not listed here? (Are these suggested "problems" even problems in this case?)