ok, i went through this in 2016 and i'm adamant about actually learning from the process rather than making the same mistakes.

you don't want to pay attention to media polls as a measure of public interest, not even if they have good methodology; there is a conflict of interest at play, and they've been clear enough about what they want. you need to take a gramscian analysis to this: the media is not an unbiased arbiter that has an interest in accurately measuring public opinion, but rather an active participant in the process that goes out of it's way to create public opinion for the purposes of maximizing profit.

and, the media has been clear that it thinks a candidate driven by what it is being called "identity politics" (but is too crude to even be that) will drive ratings. so, it wants a female candidate because it knows that gender rivalry is good for ratings, or it wants a black candidate because it knows that race conflict is good television. it doesn't care about actual policy.

so, i'm standing here saying "you know, a corporatist democrat like harris isn't really that different than trump on most things, may be a little better on some things, and may be a lot worse in a lot of important ways.", whereas the media is saying "sanders and trump are the same candidate because they're both white men, and that's bad for ratings. we need somebody that looks different than trump to maximize ratings."

so, they'll flat out publish fake polls to try and create what they want.

but, there's a caveat: what they do is actually effective on a large number of voters. that is, if the media decides that x candidate won the debate, and broadcasts it for days or weeks, you'll start to see that coverage reflected in real polling.

wait for polls done by non-media sources.

but, the media is being crystal clear on the type of candidate it wants to support, namely somebody that looks different than trump, and it's just another hurdle for the left to overcome - albeit a very substantive one.