3 ms·
That would seem likely. But then the polls should undercount a population that is more likely to vote for the Dems and so the polls should have shown more supp
by tomjen3 6y ago
That would seem likely.
But then the polls should undercount a population that is more likely to vote for the Dems and so the polls should have shown more support for the republicans than what happened in the elections.
Maybe we should just have taken the lesson of Dewey defeating Truman and not trusted polls is the first place.
- esoterica 6y agoThe polls were not great this cycle, but the 538 forecast, which is mostly polls-based, still correctly predicted 48/50 of the states in this election. If you went an entire cycle without looking at a single poll, or reading a single piece of media informed by a poll, do you think you could do better than 48/50?
- icedistilled 6y agoCorrectly predicting is not a badge of honor if the numbers are way off. As an extreme hypothetical, if he'd predicted California to be Biden+40 and it was actually Biden+10 that would be correctly predicting who won but it would also be a massive polling failure. If the numbers land outside the predicted margins, it's bad polling regardless of whether the binary decision was right.