But what we really need (to convert this to science) is to measure the objective metrics. For example, measure the number of articles in [insert social network name] for claims that contradict established scientific laws (let's say, physics). I know it'll be more difficult to do for economy or sociology, but at least with physics we might have a large body of well established and experimentally confirmed facts.
What constitutes the top N pop songs each year is becoming more distilled and refined as the industry matures, which means they're all converging onto each other and losing complexity. I also contend if you chart the volume of sales for the whole music industry vs. this tight cluster, the cluster's ratio of single or streaming sales decreases.
Meanwhile there's artists in an increasingly large tail making what you'd probably call "pop" in style but just without marketing push, which are more experimental and less refined, and these become underrepresented in the study.
1. Counts of unironic/unacknowledged logical fallacies - discussing psychology of loss aversion wouldn't be an example or even a strategic examination of loss aversion's performance and concluding it say has a niche. But stating it is better to not make an additional $10k without extra work or opportunity costs and pay a higher marginal tax rate on it? Flagged. 2. Claims which when examined closely are vacuous (having no meaning). 3. Unsupported claims which upon closer examination do not justify assertions but merely reassert more furiously. 4. Evasive chosen definitions that refuse to stay consistent as either specific or general in classifications.
One gap with it is a matter of hypotheticals and implications - even a theoretical model contrary to reality can be a demonstration against a hypothesis.
A fuzzier/snarkier example at a field level is a "SCI/Gen test" where even experts cannot distinguish noise from data indicates a lack of knowledge - although in a more positive sense it could just be a field's complexity being high such as description of a drug which interacts with proteins. I am no expert in it but given some infamously whimsical names they might not realize that a drug's complex organic compound that binds to the "Romero" prions wasn't a real specific subtype of prion without looking it up - let alone that the drug wouldn't bind with it let alone without unacceptable side effects.