Polling Week in Review

August 29-September 4

“Far from being a source of weakness, the fact that Biden honed his political instincts over many decades is a source of electoral strength. And if Biden becomes the 46th president of the United States, it will not be despite his failure to understand (what most members of his own party believe to be) the spirit of the moment; it will be because he had the good political judgment to eschew it.” - Yascha Mounk, contributing writer at The Atlantic.

This was a very bad week for Trump and it comes right on the heels of a Republican convention that did nothing to help his approval rate or head-to-head standing with Biden. The Atlantic article yesterday (not the one referenced above) reporting on the shocking insults this president has made about fallen veterans is not going away – which is why the White House and Fox News are focusing on Nancy Pelosi’s hair appointment. But, this post is about the polling. Perhaps we will see some impacts on the polling from this story in next week’s Polling Week in Review.

According to First Read, in the past two weeks “[p]oll after poll showed Biden ahead nationally and in the key battlegrounds, all conducted after the Republican convention and the GOP’s “law and order” focus. Poll after poll found Trump losing the message battle on law/order/safety. A new ABC News/Ipsos poll finds a significant majority of Americans believe President Trump’s rhetoric on the unrest in parts of the country is making matters worse — with even members of his own party divided on his approach. “Over half of the country — 55% — in the new poll said they think Trump is aggravating the situation, while just over one in 10 Americans, 13%, said they think he is making it better.” Biden is even winning among active military personnel.

What we saw this week is that Biden continues to have a strong and stable lead in the national polls, and that he might have even received a bump from the convention (his approval rating definitely went up) while Trump may have lost support from his convention. However, we saw a lot of reports and social media complaints that the race is tightening and that means that Biden is going to blow it. The Biden team now has a name for Democrats who are constantly worrying that the sky is falling no matter how good the polling data is and continue to make completely uninformed comparisons to 2016: bedwetters. I am a little surprised that they let that get out (in fact, it was intentionally tweeted out by Biden’s press secretary), but the worriers are tedious to deal with. One thing I hope to accomplish with this blog is to get people to either focus on understanding the data or just ignore it altogether (as long as they are volunteering on campaigns). I don’t care if people ignore the polling if it helps them stay motivated, but at this point the worriers are probably more frustrating to me than they are to Team Biden.

Cook Political Report’s Amy Walter reinforces Biden’s press secretary’s concerns. “At the same time, high-quality national polls are more consistently fielded than high-quality state polls. And, for all the criticism of polling in the 2016 election, the national polls were also the most accurate.”

“As such, the best way to think of the national polls is that they can help get us in the right neighborhood, but not necessarily to the correct address. The bigger the margin for Biden, the less important that we get the exact address. For example, a seven or eight-point margin may not be enough to put Georgia or Texas into Biden’s column. But, it’s enough to get him the states he needs to hit 270. The closer the margin is to four points (like we saw back in March), the less helpful the national polls become. And, the more we should just assume a close contest for the Electoral College.”

Nate Silver of FiveThirtyEight seems to agree with Walter. He has calculated the probability of Biden winning the Electoral College by popular vote margin. While the former vice president has only a six percent chance of winning the EC if the margin is between zero and one point and the race could go either way if the margin is between two and four points, a margin bigger than four points is significantly in Biden’s favor. Between four and five points, Biden’s chances of winning the EC are 89%. Between five and six points, the probability is 98%. Over six points and the odds of Biden winning the EC is 99%.

The current FiveThirtyEight polling average does show a narrowing of the race from over 8 points last week to 7.4 advantage for Biden this morning. A simple polling average of all of the national polls over the past week show Biden+8.4. There were at least 49 national polls over the past week. Let’s take a look at what the best pollster and worst pollster found.

• Selzer & Co. is the best pollster that released results over the past week. Ann Selzer is great pollster and most people unfortunately have never heard of her. She gained some fame by correctly calling the 2008 Democratic caucuses in Iowa. Selzer & Co. is one of six A+ rated pollsters by FiveThirtyEight and finds this week that Biden is up by eight points.

• Zogby is the worst pollster who released survey results this week. Zogby doesn’t even have a pollster rating, and that is probably because FiveThirtyEight only gives F ratings for demonstrable fraud. Even so, Zogby released two survey results this week. One was Biden +3 and the other was Biden +6.

While CNN/SSRS reported national results this week that are more in line with what we are seeing from other pollsters, they are still using its completely unhelpful 15-state “Battleground Sample,” Fox News is polling individual swing states the way it’s supposed to be done. And the results are once again not good for Trump.

• Arizona: Biden 49%, Trump 40%; Biden +9

• North Carolina: Biden 50%, Trump 46%; Biden +4

• Wisconsin: Biden 50%, Trump 42%; Biden +8

Election “Day” is about to begin. According to the AP, “Mail balloting in the presidential election is set to begin Friday as North Carolina starts sending out more than 600,000 ballots to voters — responding to a massive spike in requests that has played out across the country as voters look for a safer way to cast ballots during the pandemic.” Many states will follow suit afterwards and in October early voting will begin. It’s not too early any more; the election is upon us.

According to Silver, he knows of no high quality polls in the field right now. But, we should expect to see them again after Labor Day. That could mean that we will not see any really good pollsters with results until after next week’s Polling Week in Review. And what could that mean? Well, based on the past few months, when we don’t see high quality pollsters reporting results the polling average tightens. So, keep that in mind next week if we end up seeing pollsters who are no better than B/C rated reporting results.

And a final thought from The Economist about the likelihood of the election being stolen or rigged: “Admittedly… China and Iran have joined Russia in seeking to influence this election through covert means, presumably emboldened by Russia having paid little price for having done so last time. This is a shocking development. But even if they were all pushing in the same direction—which is unlikely—there is no reason to think that they could decisively tip the result.”

“America’s electoral system is sufficiently decentralized for attempts to rig the vote on a large scale to be incredibly hard. And though voter fraud occasionally takes place, both in-person and by means of absentee ballots, it is harshly punished and very rare; various studies have found the rate to be well below one in a million votes cast.”