Now that the live stream is finished, here it is. The preliminary results for the Pennsylvania poll. Final results will be out tomorrow.
I'll reply to this post when I watch the whole thing.
Watched it. Keep in mind we will get final results tomorrow. Very, very, interesting stuff, worth the watch.
- When weighted, Biden is at a D+2.29. Keep in mind that in Baris's poll, Hillary Clinton was 3 or 4 points by October 24 in that state.
- 46 out of 54 counties in Pennsylvania are the cultural Appalachian region
- The final number will look different from what they are talking about today
- Currently at 1001 responses
- Reached out Independents around the states than the Sienna poll
- Raw electorate was D+9, D+10, or even D+11. 2016 in Pennsylvania was D+3 in exit polling, but Baris speculates it was actually D+4 or D+5
- Wants to get more responses from Erie, Bucks, and Beaver County
- Allegheny County: Biden 55.9% - Trump 35.3%
- Philadelphia Suburbs: Biden 50.4% - Trump 37%
- Barnes suggest the reason why the polls are so off for Biden - meaning he's not going to landslide in swing states - is because they use limited mode to poll
- Trump is up 55-44 with Luzerne, Lackawanna, and Monroe County
- Not a lot of response in Northeast Pennsylvania, but seems to be going for Biden
- Trump winning Greene County by only twenty points, much smaller than in 2016
- Berks County seems to be slipping away from Trump yet for some reason, is winning Chester County
- Bucks County is +7 Biden
- Baris can see Biden doing better in rural areas that were not friendly to Clinton
- He believes the sample is too moderate and too liberal
- Barnes speculates that the continuing decline in response mode and changes in modes and methodology may actually favor Biden
- Pollings have gotten even worse since 2018 and COVID-19 may have contributed to the increased bias in polls like YouGov which oversamples liberals
- Weighting control, which was a big thing under the 2012 election, is no longer a thing among many pollsters
- The sample Baris has so far is inconsistent, at least before weighting
- Barnes suspects that polls are inflated for Biden/Democrats because it is early voting and those who have voted early - especially on the Democratic side who will make up the majority of early votes - are more excited to tell who they voted for
- Democrats in general are not meeting expectations for voting by mail. Either they run a little below average (Virginia, Minnesota, and North Carolina) or awful (Michigan and Wisconsin). They are failing to get youth voters out there.
- Barnes wants to look at early voting for North Dakota and South Dakota. As they do not have any "Get out the votes!" (GOTV) efforts or competitive races that would put Republican incumbents in danger, he wants to see if Republicans are overperforming or underperforming where over 20% have voted
- South Dakota has a 19 point lead with Republicans while North Dakota has a 47 point lead, both states have high Norwegian populations which is indicative of how they will also vote in rural Minnesota and western Wisconsin. He speculates that polling errors is because the Norwegians are hard to reachs
- In Pennsylvania, Trump is winning Polish (53-40), Italians, French (60-40), German, Scot-Irish but losing non-Scot Irish Catholic to Biden by six points, English, and Dutch.
- Baris finds it odd that Trump is winning Chester County yet losing Bucks County by seven points, wants to reach out to Italians there
- Trump is doing great among economy and policing, but is getting killed in healthcare and coronavirus
- Baris and Barnes believe that Republicans totally screwed up by not preparing for Barrett's position on ObamaCare and pre-existing condition
- Confirms in the raw data that Trump is taking around 20% of the black votes in Pennsylvania
- Trump is winning among crossover votes compared to Biden who is only getting 8% of Republicans but needs to improve his margin among Indepedents as tied is not good enough
- Nearly half of the Pennsylvanian voters have no confidence in the integrity of this election and that's before weighting
- Most people are uncomfortable with Kamala Harris being president
- Fracking is overwhelmingly supported in Pennsylvania
- Leaners are going for Trump by three point
- Baris's 2016 poll did not show it going to Clinton+3 until October 24th (and might have been at Clinton+4)
- When Baris asked the question of if people were comfortable or uncomfortable with telling others about who they are voting for among family, friends, strangers, neighbors, co-workers, and pollsters, it is
significantly worse than in Florida, a far more Republican-friendly state
- Barnes explains that Gallup did in-person polling from 1936 to 1968, switched to landline from 1968 to 2004, and that pollsters moved to online during the 2000s. Half of the people with a landline were polled, which creates random sampling. With the move to online polling, it gets rid of random sampling, making polling purely an art rather than half art-half science
- Most pollsters have failed to correct their mistakes from 2016 as they take the wrong lessons from why they got the election result wrong. Like for example, they aren't adjusting both education and region, just education
- Philadelphia was grossly overrepresented in the raw data at 18.8%. Weighting should take it down between 12-14%, making it an unusually low turnout
- Barnes believes that voter registration is a more accurate predictor of electoral outcome than polls. Voter registration trends have shown whenever or not a state trends Republican or Democratic. Ex. in Pennsylvania, North Carolina, and Florida, Bush made gains in these states compared to 2004, they all dramatically shifted Democratic in 2008 when Obama had his landslide, shifted Republican when Obama had a closer victory, and all voted Trump when they shifted Republican again. All three states have trended Republican at an even faster rate than 2020.
- Baris plans to get 150 more samples
- Biden is three points ahead in Centre County
- Currently tied in Erie County, but that about 14% of the voters there are undecided (which neither Baris or Barnes believes)
- Nearly 2/3 of people believe there are "shy" voters who do not want to share they want to vote for Trump
- It's looking like Trump may flip Monroe County
- Trump is crushing Biden when it comes to working class voters 55-35