I genuinely hope we are close to moving past the era of “passively polling” the internet and treating it anything meaningful.
As @EylerWerve pointed out on the Little Mermaid racist review bombing thread, it is the weakest form of data gathering, and we have actual refined tools for the job of polling, Twitter News™ is the same lazy, cheap-ass non-method of arm-chairing social science. Particularly when the headline is “Twitter reacts to” or “Twitter is in an uproar” as if it’s a collective entitiy.
Unfortunately it seems like Twitter in some ways benefits from the bias that people have toward a “person saying something” (anecdote) vs anonymous data points in a survey, because each Twitter user can be perceived a “person” with a visual presence and a handle. It’s the worst of quantitative data combined with the worst of qualitative data delivered in a way that makes both readily available and also worse than meaningless.
