This!
With all polls, extreme skepticism of the methodology should be our first (and often only) stance. The way they asked the question, to how many people, and the way they chose the audience is everything. You can get any result you want by tweaking those three things, intentionally or not. Surveys are essentially useless for this reason. The controls are far too weak and the error bars swamp the data.
I wouldn’t put any stock in results like this at all. We know the current state of media has some issues and we should work on them, but a poll like this is worthless.
