New Economists/YouGov poll, electability question: Biden 65%, Warren 57%, Sanders 52%, Harris 47%

[thesquanderer]

16. People actually KNOW who they are likely to vote for. They have NO way of knowing...

...who other people will vote for (unless they are parroting other more valid polls, in which case you might as well just go look at the more valid polls yourself). So their answer to that question is largely a result of their own biases and what they have read in articles about, well, who people would vote for, which just creates a circular argument (i.e. they think X would beat Y because they read that X would beat Y, which really tells you nothing).

I just had an exchange in another thread with someone who said Warren can't beat Trump because she would lose the key midwest states that cost us the election last time. And then I showed that every poll of Warren vs. Trump in those states shows Warren winning. So what's more accurate, asking people in Michigan, Wisconsin, and Pennsylvania who they would vote for if the choice was Trump or Warren? Or asking a bunch of random people across the country who they think people in Michigan, Wisconsin, and Pennsylvania would vote for? The former tells us how the vote would likely go if the election were held today. The latter just tells us what people's biases are and what people have heard/read other people saying. What's the point of asking what you think others will do, when you could actually ask those other people what they would do? I understand what you're saying, some people may lie about what they would do. But some random person in some other state has no way to provide more accurate info than what you'd get asking the people involved. The info can only be less based on fact, not more.