Quote (InsaneBobb @ 30 Sep 2020 02:17)
That's funny. Honestly funny. Arizona, for instance, going by a quick look at 538 is in a dead heat with Biden. So, what aggregate is the above chart even from?
Again, totally convinced polls win elections.
Here's a link to their prediction for Arizona, click on popular vote:
https://projects.fivethirtyeight.com/2020-election-forecast/arizona/They predict Biden to win Arizona with 50.4% of the vote against Trump's 48.3%, a margin of 2.1%. So if we applied the 1% polling error in Trump's favor from 2016 that you brought up, then we'd still end up with a narrow Biden win.
I agree with you that polls dont win elections, but this doesnt mean that they carry no information whatsoever and can be discarded entirely. There was a big polling error in Trump's favor in 2016, at least in the crucial swing states. As I showed, even if he overperforms his polls by the same magnitude again (which seems a rather unlikely scenario to begin with since many pollsters learned from their mistakes), then he'd still lose. He needs an even bigger polling error than last time, while a big polling error actually seems less likely this time around.
This post was edited by Black XistenZ on Sep 29 2020 06:52pm