A couple of you asked me some questions over the weekend that are relevant to today’s polling review. Please feel free to send me any other questions you may have.
Am I going crazy? The polling right now is so stable that Biden should win PA, MI, WI and VA that, essentially, it looks very, very, likely that he wins with at least 290 electoral college votes. It would require a substantial voter suppression effort to prevent that. Right? Or is there something in the polling that should make me more worried about those states?
The only worry I see is R new registrations outpacing D in all four of those states and new voters aren’t going to hit a LV screen that plus suppression in combo is still a worry for me.
A lot of pollsters screen for likely voters simply by asking if they will vote instead of looking at voting history these days. Ann Selzer "proved" that asking was more reliable than voter history because the latter missed new voters. That is how she accurately predicted Obama's victory in Iowa in 2008 caucuses. The polls at this point will have to be very wrong for Biden to lose - unless something changes in the next two weeks. However, as of this morning over 28 million people have already voted.* Every day that passes means millions more votes submitted, and that means any event that could change the race will have less ability to do so.
Georgia is now leaning to Biden. Florida polling still seems all over the place to me. Ohio is even, but Biden is starting to look like he has an advantage, but it is small. The Ohio polls are telling us how difficult the road to victory is for Trump.
The polling in Wisconsin, Minnesota, and Michigan looks out of reach for Trump. Pennsylvania does too, but I think the race could be a little closer there than Democrats hope it is and the GOP registrations there do concern me a little. Biden can win without Pennsylvania (not easily, but losing Keystone State for Biden is not the same as losing Florida is for Trump). However, if I had to project a winner in Pennsylvania based on current polling it would be Biden, and by five points or more.
It is true that the GOP has been winning the ground game in voter registrations in North Carolina and Florida. Cook thinks that Democrats are doing less well because of lack of college students on campuses and the because Biden campaign has not been investing in voter registrations as much as the Trump campaign has been. The net gain “pales in comparison” to the number of votes to be cast in those states. The polling leads in those states are more important than the voter registration race.
Source: FiveThirtyEight
Biden’s lead in the FiveThirtyEight polling average remains steady at plus 10.7 points. As we approach the final two weeks of the election season, we’ll keep an eye on the polling to see how close pollsters get to the final national result and see how close the state polling averages are to the final result in several battleground states. First up is a sample of pollsters organized by FiveThirtyEight ratings. Here are the most recent polls by pollster.
The Best: First Tier (A+) pollsters:
Siena College/New York Times Biden +8
Marist College Biden +11
Monmouth University Biden +5
ABC News/Washington Post Biden +12
Above Average: Second Tier (A to B+) pollsters:
IBD/TIPP (tracking) Biden +5
Fox News Biden +9
NBC/Wall St Journal Biden +11
SurveyUSA Biden +10
Average: Third Tier (B to B/C) pollsters:
YouGov: Biden +11
USC Dornsife (tracking) Biden +12
Ipsos Biden +6
Morning Consult Biden +9
Mediocre: Fourth Tier (C+ and worse) pollsters:
HarrisX Biden +7
Survey Monkey Biden +8
Rasmussen Biden +5
Zogby Analytics Biden +2
Next we will look at the polling averages by state (and Congressional district where applicable), according to The Upshot. For the states, we’ll focus on whether pollsters are collectively getting right the race in the battleground states instead of how individual pollsters are faring. Since some pollsters are only in a few of the states and others only poll once or twice in the entire election cycle, it’s better to stick with the polling averages here.
Arizona: Biden +5
Florida: Biden +4
Georgia: Biden +2
Iowa: Biden +3
Maine-02: Biden +5
Michigan: Biden +9
Nebraska-02: Biden +8
New Hampshire: Biden +11
North Carolina: Biden +3
Ohio: Biden <1
Pennsylvania: Biden +8
Wisconsin: Biden +8
Texas: Trump +2
There is one state pollster we want to keep an eye on because it is so out of step with every other pollster this year (and in the past). You may have heard about The Trafalgar Group. They are not doing national polling (from what I can tell so far), but they do plenty of state polling. Their polling makes it looks like Trump is doing much better than other pollsters – even other mediocre pollsters who favor the GOP like HarrisX and Rasmussen. Trafalgar asserts that there is a shy Trump voter that is not being captured in other polling because they won’t respond or will lie to the interview. I have noted before that Morning Consult has tested this concept and found little, if any, support for it. FiveThirtyEight interviewed many of the well-known pollsters out there recently and the only one that said that they were worried about a shy voter effect was Gravis Marketing, which is a Republican PR firm.
In a recent interview with the National Review, Trafalgar’s chief pollster refused to explain how their “proprietary digital technology” incorporates this shy voter impact. You’ll also hear a lot of Trump supporters on social media call Trafalgar as the pollster who “got it right” in 2016. But, that’s not true. Yes, Trafalgar had Trump winning Pennsylvania and Michigan – however, their polls were four to four and a half points higher than the result. They missed Florida (other than by predicting Trump the winner) by the same degree and had Trump winning Nevada by five points (Clinton won by 2.7 points). In 2018, Trafalgar had Brian Kemp beating Stacey Abrams in Georgia by 12 points. The actual margin was less than one point.
Let’s keep an eye on some of their state polling and see if they really have some special sauce this year that beats the other pollsters. There are four battleground states that Trafalgar is regularly polling. The polling is suspicious. For example, in their previous four polls in Pennsylvania Trafalgar had nearly identical results (all were Biden +2). Together with their shy voter weighting, which they will not describe, this suggests that Trafalgar begins with the result they want and they weights the data to get to it. Of course, I am assuming they are not fabricating the raw data but that cannot be ruled out until they start being transparent.
Florida: Trump +2
Pennsylvania: Biden +2
Michigan: Trump +1
Wisconsin: Biden +1
There are 15 days left until November 3rd.
* This was announced in a Cook Political Report briefing this morning.