NYT/Siena Polls of 6 Battleground States

[Annatar]

Siena was accurate in 2018 because they more than other pollsters made the biggest effort to ensure that their samples matched up with voter file data and were representative of the electorate. I trust Nate Cohn is still doing that and that’s why I think they’re reliable. Even on trump approval when other trash pollsters were markedly off on trump approval state by state, Siena were the closest to the exit polls. I see no reason to believe why they would have changed their methodology to be wrong.

[Annatar]

Im not gonna pick apart Siena, especially after 2018. Its not surprising that Trump is holding up better in the midwest compared to the other regions of the country.

I also wished they polled his job approvals. He was at 45/54 in 2018 exit polls, that seems about where he still is in these states too. Maybe a few points better than that.

In the swing states in the 2018 exit polls he was at 51% approval in Florida, 50% in Arizona, 48% in Wisconsin and 45% in Pennsylvania & Michigan.

[Annatar]

I'm looking at the crosstabs, and noticing that this poll's sample deviates pretty significantly from the 2016 exit polls.  Examples:  In PA, 2016 exits for education level showed 22 HS diploma or less, 30 some college, 29 bachelors deg, 19 postgrad.  This poll's sample shows 37-28-21-14.  In MI, 2016 exits showed some college or less 58%, bachelors degree or higher 42%.  This pol's sample has those numbers at 69-30.  Also in MI, 2016 exits showed party ID at 40 Dem, 31 Rep, 29 Ind.  This polls sample is 29-24-44.  Not saying that these numbers are necessarily wrong, but it would be interesting to hear the pollster's explanation as to why they think the 2020 electorate will look so much different than in 2016.

The 2016 exit polls got education levels wrong, they corrected that error in 2018, if you want to know the actual makeup of the electorate in 2016, look at this analysis from Center for American Progress.

https://www.americanprogress.org/issues/democracy/reports/2017/11/01/441926/voter-trends-in-2016/

[Annatar]

Nate Cohn reckons the swing states are about 5% more R than the nation.

"Over the summer, we did an analysis of what we thought the President’s approval rating was in every state. And in these states we thought the President’s approval rating was five points better than nationwide. So, if Biden is up two in these states, I would say maybe that he is up seven nationwide. That would make Warren up three nationally."

https://www.newyorker.com/news/q-and-a/the-times-nate-cohn-on-elizabeth-warrens-odds-against-donald-trump

So this poll would suggest a Biden lead nationally of around 7%, comparable to Obama in 2012. Since Trump won these states by 2% in 2016 and lost the NPV by 2%, it would also suggest his advantage in the EC/NPV split has increased from 3% in 2016 to 4% today.

[Annatar]

Republicans ran 1.5% ahead of trump in the Virginia senate and basically even in the Virginia house, those results would suggest this poll is to pessimistic on trump’s numbers.

[Annatar]

Republicans ran 1.5% ahead of trump in the Virginia senate and basically even in the Virginia house, those results would suggest this poll is to pessimistic on trump’s numbers.

What in the world?

https://twitter.com/ryanmatsumoto1/status/1191948895926063105