Monthly Archives: September 2020
WOW, USA AND ITS DONALD
BBC FACT CHECKERS https://www.bbc.com/news/election-us-2020-54350219
SECTIONS SEARCH SKIP TO CONTENTSKIP TO SITE INDEXPOLITICS Account How the Internet Viewed the First Presidential Debate: Poorly “The worst debate of all time,” one journalist wrote on Twitter. The debate as seen from a bar in Las Vegas. The debate as seen from a bar in Las Vegas.Credit…Bridget Bennett for The New York Times Matt Stevens By Matt Stevens Sept. 30, 2020 Updated 8:20 a.m. ET 238 In these polarized political times, it can sometimes feel as if Americans really cannot agree on anything. But Tuesday’s highly anticipated presidential debate, the first between President Trump and Joseph R. Biden Jr., the Democratic nominee, managed to achieve rare bipartisan consensus: Just about everyone offering commentary online agreed that the debate was an unmitigated disaster. ADVERTISEMENT Continue reading the main story Subscribe to The Times to read as many articles as you like. nytimes.com/subscription Site Index Site Information Navigation © 2020 The New York Times Company NYTCoContact UsWork with usAdvertiseT Brand StudioYour Ad ChoicesPrivacyTerms of ServiceTerms of SaleSite MapHelpSubscriptions
MORNING AFTER THE D BATE
The Upshot on Today’s Polls
Analysis of the day in polls, from now until Election Day.
POLLING LEADER | IF POLLS ARE AS WRONG AS THEY WERE IN… | ||||
---|---|---|---|---|---|
2016 | 2012 | ||||
U.S. | +7 Biden | +4 | +11 | ||
Minn. | +9 Biden | +2 | +10 | ||
N.H. | +8 Biden | +5 | +11 | ||
Nev. | +7 Biden | +9 | +10 | ||
Mich. | +7 Biden | +2 | +12 | ||
Pa. | +7 Biden | +2 | +8 | ||
Wis. | +7 Biden | <1 | +9 | ||
Neb. 2* | +7 Biden | +8 | +1 | ||
Maine 2* | +3 Biden | +8 | +8 | ||
Ariz. | +3 Biden | +1 | +2 | ||
Ohio | +2 Biden | +5 | +2 | ||
N.C. | +1 Biden | +4 | +1 | ||
Iowa | <1 Biden | +7 | +4 | ||
Fla. | <1 Biden | +2 | +2 | ||
Ga. | <1 Trump | +2 | <1 | ||
Texas | +3 Trump | +5 | +4 |
Winners in these surveys typically gain in post-debate polls.
In the wake of Tuesday’s messy debate, we get a different kind of survey: a so-called instapoll. This is a survey of people who tell a pollster they will watch a debate and agree to take a poll after the debate is over. These surveys are quirky in some ways, but they do a decent enough job of helping us gauge the effect of a debate. This time, the news was generally good for Joe Biden.
Instant Poll Surveys of Who Won the Debate
Solid instant results for Biden. CNN found that Mr. Biden decisively won the debate, 60 percent to 28 percent, while CBS News and an early cut from a Data for Progress poll found far closer seven- and 12-point leads for Mr. Biden.
Why the huge spread? Instant-reaction poll samples can be pretty biased. Often, you’re looking at respondents who have negotiated three steps: They agree to take a poll before a debate; they agree to take a poll after watching a debate; they actually follow through and participate in the poll after the debate. They’re also people who have watched a debate, who aren’t representative of the overall electorate.
On top of that, pollsters have to decide whether or how to undo those biases. The CBS poll, for instance, weighted its sample to match the demographic composition of people who had said they would watch the debate. Data for Progress’s poll reflected the “2016 electorate,” which may make their numbers relatively comparable to national polls. CNN’s methodological document doesn’t make it entirely clear how their sample was adjusted to be representative of people who watched the debate. All of this can lead to a huge spread in the results.
A deeper look is still good news for Mr. Biden. With all of these factors at play, it can be hard to know what to make of the instant poll results. You might be tempted to just discount them altogether.
But a closer look in this case raises the possibility of good news for Mr. Biden. His favorability rating improved by a net four percentage points, compared with how the same respondents answered before the debate. The president’s rating declined by a net four points.
Similarly, the CBS poll found that voters said the debate made them think better of Mr. Biden by a margin of 38 percent to 32 percent, while just 24 percent said the debate made them think better of President Trump and 42 percent said the debate made them think worse of him.
The results are kind of useful. Historically, the winner of these polls tends to gain in the real polls over the next week. John Kerry, Barack Obama, Mitt Romney and Hillary Clinton all picked up a couple of points after debate performances that instant polls suggested they won.
This year, it’s hard to be sure whether to expect a similar shift. The national polls have been extremely steady compared with prior years (Mr. Biden has a consistent seven-point lead nationally in our average). And if there’s similar shift, it’s worth being cautious. It could just be the result of one candidate’s supporters being particularly energized or discouraged after a debate performance. We’ll have to wait a week or two before we know whether any post-debate shifts are durable.
The big picture: It’s hard to say anyone clearly won the debate last night, and that’s a win for Mr. Biden. He was the front-runner heading into the debate, and it was the president who needed a win to try to narrow the race. None of the instant poll results, whatever their merit, do much to dispel that view.
New Pennsylvania polls give Biden a significant lead.
Exploring Electoral College outcomes
For the first time since we started our poll tracker several weeks ago, Joe Biden leads by enough to withstand a repeat of the polling error in 2016.
If the polls were exactly as wrong as they were four years ago, Mr. Biden would hold the Clinton states and flip Michigan, Pennsylvania, Arizona and Nebraska’s Second District, giving him more than the 270 electoral votes needed to win.
The big shift: Pennsylvania. An ABC/Post poll overnight and a Times/Siena poll Monday showed Mr. Biden up by nine points in the state, giving him a far more comfortable lead there than Hillary Clinton held over the final stretch four years ago. It’s worth waiting for a few more high-quality polls to confirm if Mr. Biden has a big edge in Pennsylvania (it’s hard to overstate the electoral consequences if he does).
Of course, the polls won’t be exactly as wrong as they were in 2016. They could understate support for Donald J. Trump by an even greater degree. But there are good reasons to believe that many causes of 2016’s poll misfire are less likely this time around:
More pollsters now weight by education, which ensures that voters without a college degree make up an appropriate share of the sample. Many state polls had far too many college graduates four years ago, leading them to underestimate Mr. Trump.
There are far fewer undecided voters, who ultimately broke toward Mr. Trump across the Rust Belt. A similar late break among undecided voters this time would be less consequential.
A more stable race. Our estimate of the polling error in 2016 is a little unfair to pollsters: We’re looking at the average error over the final three weeks of the race. That includes polls taken after the “Access Hollywood” tape and before the Comey letter. Those polls showed Mrs. Clinton with a commanding lead. The final polls were more accurate, if still biased toward Mrs. Clinton.
This time, it’s probably less likely that the polls will swing greatly over the last few weeks of the race. After all, they’ve been far more stable over the last few months than they were at any point in 2016.
That doesn’t mean that we’re immune to another misfire. Indeed, most polling misses happen precisely because the causes can’t be anticipated. That’s why our tables on this page also include the polling miss from 2012, when surveys underestimated President Obama. But it also means there’s no reason the polls couldn’t be off by even more than they were four years ago.
A race teetering between a landslide and a tight race.
The first presidential debate is Tuesday, marking the beginning of the final stretch of the 2020 campaign. It’s a fine time to take a step back and look at the big picture.
The average of polls shows Joe Biden with a seven-point lead nationwide, giving him the largest lead of any candidate at this stage since Bill Clinton in 1996. He also has a consistent, if usually more modest, advantage across the battleground states.
If polling averages today translated perfectly to election results on Nov. 3 — and they won’t, it’s worth repeating — Mr. Biden would win nearly 360 electoral votes. If he outperforms the polls, he could win more than 400 electoral votes in the largest electoral landslide since 1988. That’s a real possibility.
But so many battleground states are close enough that the president’s re-election hopes are still alive. If he outperforms the polls as he did four years ago, he could certainly win.
It might seem odd for a landslide and a competitive race to seem so realistic. But consider this: Mr. Biden is closer in our poll average to winning Texas, which would get him over 400 electoral votes, than President Trump is to winning in traditional battleground states like Pennsylvania and Nevada.
State polls
POLLSTER | MARGIN | DIFF. FROM ’16 RESULT | |
---|---|---|---|
Alaska |
Harstad Strategic Research (Dem. pollster)
Sept. 20-23, 602 L.V.
|
Trump +1
46-47
|
+14D |
N.C. |
Meredith College
Sept. 18-22, 705 R.V.
|
Biden +1
46-45
|
+5D |
Neb. 2* |
New York Times/Siena College
Sept. 25-27, 420 L.V.
|
Biden +7
48-41
|
+9D |
Nev. |
University of Nevada, Las Vegas
Sept. 10-25, 641 L.V.
|
Biden +5
46-41
|
+3D |
Ohio |
SurveyMonkey
Sept. 1-25, 3,092 R.V.
|
Trump +2
48-50
|
+6D |
Pa. |
New York Times/Siena College
Sept. 25-27, 711 L.V.
|
Biden +9
49-40
|
+10D |
Pa. |
IBD/TIPP
Sept. 24-26, 774 L.V.
|
Biden +5
50-45
|
+6D |
Texas |
P.P.P. (Dem. pollster)
Sept. 25-26, 612 L.V.
|
Even
48-48
|
+9D |
National polls
U.S. |
Monmouth University
Sept. 24-27, 809 L.V.
|
Biden +5
50-45
|
+3D |
U.S. |
USC Dornsife/Los Angeles Times
Sept. 14-27, 5,253 L.V.
|
Biden +10
52-42
|
+8D |
U.S. |
Echelon Insights (Rep. Pollster)
Sept. 19-25, 1,018 L.V.
|
Biden +9
51-42
|
+7D |
U.S. |
Harris Insights & Analytics
Sept. 22-25, 2,768 R.V.
|
Biden +5
45-40
|
+3D |
Monday’s polls didn’t change the fundamental state of the race. The teetering between a possible landslide and a competitive race continues.
The biggest news came from Pennsylvania. This is arguably the single most important state keeping the president highly competitive, and a Times/Siena College poll on Monday showed Mr. Biden up nine points among likely voters. This was a result in the landslide category.
Now, it’s just one poll. It’s a good poll, if we may say so ourselves, but it’s still one poll. Indeed, there was an online poll today that showed Mr. Biden up five points there. So it’s not as if we can say that he has now taken a commanding lead in the state.
But it’s hard to overstate the electoral consequences if the Times/Siena poll is even in the ballpark. With Michigan and perhaps even Wisconsin slipping toward Mr. Biden, the president simply can’t afford for Pennsylvania to drift into a deficit in the upper-single digits as well.
Time will tell whether Mr. Biden has really taken a significant lead. Fortunately, we won’t have to wait long. ABC News/Washington Post will have a poll late Monday night, and Monmouth will poll Pennsylvania after the debate. There will be a Times/Siena poll after the debate as well.
If the balance of new high-quality polls confirms such a wide lead in the state, it will be devastating to the president’s re-election chances.
Biden has a backup plan. If Mr. Biden can’t pull off Pennsylvania, he has a backup plan: flip Michigan, Wisconsin, Arizona; and win Nebraska’s Second Congressional District. It adds up to exactly 270 electoral votes, and a Times/Siena poll in Nebraska’s Second showed that Mr. Biden also had a comfortable lead there (seven points).
More tantalizing clues in Alaska. Alaska is a severely underpolled state, but a lot of people think the race there is close. Today, we got another clue of a competitive race: A Democratic firm, Harstad Strategic Research, found Mr. Trump leading by just one point. This is also in the landslide category. But it’s what we expect from polls by Democratic firms, which tend to release relatively favorable results for their preferred candidate, so don’t take this too far. We’re mainly flagging it as an opportunity to explore a state that we haven’t gotten to talk about.
Alaska hasn’t voted for a Democrat for president since 1964, but there are reasons to think it could be competitive. Mr. Trump won only 51 percent of the vote there in 2016, putting it right next to some highly competitive states, like Iowa and Ohio. The president won comfortably thanks to a large minor-party vote, but polls show Mr. Biden excelling among those who supported Gary Johnson and others in 2016. If Mr. Biden is faring well in Alaska, it will certainly fit a pattern. He has been doing really well in Northern rural states with an independent streak, like Maine and Montana.
As tantalizing as it may be, the state probably wouldn’t turn blue unless there was a blowout. A blue Alaska could have some important consequences for the U.S. Senate, but we don’t anticipate it deciding the presidential election. (I could go on about the extremely remote possibility that the Alaska House race decides an Electoral College tie in Mr. Biden’s favor, even if the state voted for Mr. Trump, but these posts are supposed to be short — maybe some other time.)
Elsewhere. Most of the polling today was of … varying quality. But it was nonetheless consistent with what we already thought. I would highlight Mr. Biden’s five-point lead in the Monmouth national poll. That’s one of Mr. Trump’s best results in a while, and it was by far his best news of the day. And many of the state polls in the Sun Belt, like in Nevada or North Carolina, were consistent with a competitive race, as they have been for most of the month.
State of the race: Mr. Biden maintains his advantage heading into the debates.
Joe Biden is up nine points in a crucial battleground.
New York Times/Siena College poll of likely voters in Pennsylvania
My colleagues Trip Gabriel and Isabella Grullón Paz have the story. Some highlights:
-
A majority of likely voters in Pennsylvania — 51 percent — said they trusted Mr. Biden more to pick the next Supreme Court justice, whereas 44 percent said that about President Trump.
-
Nearly half of all voters, 47 percent, “strongly disapprove” of how Mr. Trump handles his job as president, evidence of entrenched opposition with five weeks until Election Day and voting already underway in Pennsylvania. A smaller share, 32 percent, strongly disapproved of Mr. Biden.
An advantage in a small but potentially crucial contest.
New York Times/Siena College poll of likely voters in Nebraska’s 2nd Congressional District
In a close race, the presidential election could be decided by an unlikely spot: Nebraska’s Second Congressional District, including Omaha and most of its suburbs.
If the race is decided there, Joe Biden appears to have the advantage, according to a New York Times/Siena College poll. He leads President Trump by seven points, 48 percent to 41 percent, among likely voters in the first nonpartisan survey of the district so far this year, according to FiveThirtyEight.
The district is traditionally Republican, but Mr. Trump carried it by only two percentage points in 2016. It was even closer than familiar battleground states like Arizona or North Carolina.
The district’s demographics have made it an even more plausible pickup opportunity for Mr. Biden. It’s relatively white, metropolitan and well educated, and national polls routinely show Mr. Biden running ahead of Hillary Clinton’s 2016 performance among all three groups.
More details over at The Upshot.
A quiet weekend of polling, and it was consistent with what we already knew.
State polls
POLLSTER | MARGIN | DIFF. FROM ’16 RESULT | |
---|---|---|---|
Ga. |
YouGov
Sept. 22-25, 1,164 L.V.
|
Trump +1
46-47
|
+4D |
Mich. |
Marist College
Sept. 19-23, 799 L.V.
|
Biden +8
52-44
|
+8D |
Minn. |
Mason-Dixon Polling & Strategy
Sept. 21-23, 800 L.V.
|
Biden +6
48-42
|
+4D |
N.C. |
YouGov
Sept. 22-25, 1,213 L.V.
|
Biden +2
48-46
|
+6D |
S.C. |
YouGov
Sept. 22-25, 1,080 L.V.
|
Trump +10
42-52
|
+4D |
Wis. |
Marist College
Sept. 20-24, 727 L.V.
|
Biden +10
54-44
|
+11D |
National polls
U.S. |
USC Dornsife/Los Angeles Times
Sept. 13-26, 5,257 L.V.
|
Biden +10
52-42
|
+8D |
U.S. |
New York Times/Siena College
Sept. 22-24, 950 L.V.
|
Biden +8
49-41
|
+6D |
U.S. |
ABC News/The Washington Post
Sept. 21-24, 739 L.V.
|
Biden +8
52-44
|
+6D |
Two polls, two substantial Biden leads. Joe Biden has been leading by about seven points nationally for as long as this page has been on the internet, and there’s no indication that’s changing. A new Times/Siena poll Sunday found him ahead by eight points, and a new ABC News/Washington Post poll had similar findings.
One twist: The ABC/Post poll released two sets of results for the presidential race. Mr. Biden led by 10 points in a one-on-one matchup against the president, but he led by six points in a four-way race with named minor-party candidates. Most polls asking both versions of the question haven’t shown anywhere near so large a gap, but Mr. Biden has tended to do about a point worse when the minor-party candidates are listed.
As an aside, our tables will average multiple versions of a poll, which is why they show Mr. Biden with an eight-point lead in the ABC/Post poll. At least in this case, that may be closer to the truth: Naming the minor-party candidates can often inflate their support, but it’s not a true one-on-one race either.
Be careful with new polls in Wisconsin and Michigan. There has been no shortage of good news for Mr. Biden in the Upper Midwest, and this weekend was no exception. NBC/Marist polls showed him with a significant lead in Michigan and Wisconsin. But this time, it’s worth taking those results with a grain of salt: They’re not weighted by education.
(Weighting by education is critical for understanding modern polling and what went wrong in 2016.) In practice, it means the NBC/Marist polls probably have too many college graduates represented and are probably too good for Mr. Biden.
Strong Democratic results in the Senate. CBS/YouGov joined a long list of firms to show very close races in Georgia and North Carolina. It also offered some promising news for Democrats in their pursuit of the Senate. It found Cal Cunningham up by 10 points against the Republican incumbent, Thom Tillis, in North Carolina, and even found Senator Lindsey Graham locked in a tight race with Jaime Harrison in South Carolina. Another poll, conducted by a Democratic firm, even found Mr. Harrison narrowly ahead.
State of the race at the end of the day: The same as it was on Friday: A solid lead for Joe Biden.
Biden leads by eight points nationally.
NYT/Siena poll of likely voters in U.S.
Some highlights:
-
A clear majority of voters believe the winner of the presidential election should decide who will fill the Supreme Court seat left open by the death of Justice Ruth Bader Ginsburg. About 56 percent of likely voters said they preferred to have the election act as a sort of referendum on the vacancy. Only 41 percent said they wanted President Trump to appoint a justice before November.
-
The coronavirus pandemic, economic distress and racial justice protests have done little to reshape a presidential campaign that polls show has been remarkably stable. Joe Biden’s eight-point lead nationally is propelled by a wide advantage among women and voters of color. Mr. Biden has made gains among constituencies that strongly favored the president in 2016, including men and older voters.
My colleagues Jonathan Martin and Alexander Burns have the story.
And Joe Biden’s problem in a nutshell.
There weren’t many polls today, which is expected. It’s a Friday, and many pollsters will try to produce a final poll just before the debate Tuesday. That means they’re probably in the field or just wrapping up right now.
The polls we did get largely reinforced the big picture: a solid Biden lead, but with President Trump keeping it somewhat closer in the battleground states likeliest to decide the election.
To feel comfortable, Mr. Biden has one goal: claim a solid lead in just one more state. He has a fairly consistent lead in the states won by Hillary Clinton, plus Wisconsin and Michigan. From there, any other flipped battleground state would get him over the top. But that final piece — probably Pennsylvania, Florida or Arizona — hasn’t always given Mr. Biden a consistent or comfortable lead. Sometimes the polls are good for Mr. Biden. Sometimes they’re good for the president. The same can be said of today’s polls.
State polls
POLLSTER | MARGIN | DIFF. FROM ’16 RESULT | |
---|---|---|---|
Ariz. |
Data for Progress
Sept. 15-22, 481 L.V.
|
Trump +1
45-46
|
+3D |
Fla. |
Data for Progress
Sept. 15-22, 620 L.V.
|
Biden +3
46-43
|
+4D |
Maine |
Colby College
Sept. 17-23, 847 L.V.
|
Biden +11
50-39
|
+8D |
Maine 1* |
Colby College
Sept. 17-23, 416 L.V.
|
Biden +18
54-36
|
+3D |
Maine 2* |
Colby College
Sept. 17-23, 425 L.V.
|
Biden +3
46-43
|
+13D |
Minn. |
Suffolk University
Sept. 20-24, 500 L.V.
|
Biden +7
47-40
|
+5D |
Nev. |
ALG Research (Dem. pollster)
Sept. 15-21, 800 L.V.
|
Biden +4
47-43
|
+2D |
Texas |
Data for Progress
Sept. 15-22, 726 L.V.
|
Trump +1
45-46
|
+8D |
National polls
U.S. |
USC Dornsife/Los Angeles Times
Sept. 11-24, 5,377 L.V.
|
Biden +10
52-42
|
+8D |
Data for Progress, and Biden’s discomfort. Taken together, the Data for Progress results could be considered fine for Mr. Biden. The Texas result (Trump only +1) again raises the possibility that Mr. Biden could win in a landslide with more than 400 electoral votes.
But if the Data for Progress results are viewed through the lens of Mr. Biden’s quest for one last state with a comfortable lead, then none of the results are particularly good for him. They’re all fairly close, and Mr. Biden’s worst result — trailing by one in Arizona — comes in the state where it seemed most likely that Mr. Biden had a good shot of claiming some security.
Here’s a different way of looking at it: Mr. Trump can keep his hopes alive as long as he stays close in Pennsylvania, Florida and Arizona, but he’d be in serious jeopardy if any one of them slipped out of play. Today’s polls kept them in play.
A Democratic poll of Nevada that’s not so great for Biden? There’s also the poll of Nevada from ALG Research, a Democratic pollster. It’s a somewhat curious release. As a rule, the quality of partisan polls is as high as that of any reputable public pollster, but most of them are used only internally. Which means if you’re seeing a partisan poll released in the wild, it’s because someone wanted you to see it.
In this case, it’s not a particularly great result for Mr. Biden. He’s up four points in Nevada, a state Mrs. Clinton won in 2016, and it comes just after a Fox News poll showed Mr. Biden up by double digits there. Usually, partisan polls are released because they show good news for the party’s candidate. So why release bad news? Perhaps if you’re legitimately worried about a state, and want to make sure people take it seriously.
An early Supreme Court tidbit. ABC/Washington Post had a good set of numbers for Democrats on the coming Supreme Court battle, suggesting that more people trusted Mr. Biden to choose a Supreme Court justice, or would prefer that the winner of the election get to choose. The poll was taken before the nomination of Amy Coney Barrett, which undoubtedly could have a big effect on public opinion. But the results suggest that Democrats start on firm ground at the outset of what’s sure to be a bitter fight.
More national data to come. We’ll get a lot more data this weekend, including a new Times/Siena national poll, the rest of the ABC/Washington Post poll, and almost certainly more.
Historically, a lot of the major national pollsters like to conduct a poll just ahead of the debates, so they can come back and see how things change afterward. This year, the Supreme Court vacancy gives the big pollsters even more reason to weigh in.
The polls just ahead of the first debate tend to be pretty accurate. Here’s one that might surprise you: Four years ago, Mrs. Clinton led by just 1.2 points in national polling heading into the first debate, according to FiveThirtyEight. She wound up winning the national vote by 2.1 points — the pre-debate polling was even closer to the election result than the (already fairly accurate) final national poll numbers.
Today, Mr. Biden is up by just over seven points in our average.
With one exception: three eye-opening state polls from Fox News.
Today was a pretty typical day of polling for this point in the race. There were a lot of polls. Some were good and some were bad — whether in terms of quality or what it meant for your preferred candidate. But if you took all the polls together Thursday, you wound up in about the same spot you started at the beginning of the day, though Fox News did try to shake things up at the very end.
State polls
POLLSTER | MARGIN | DIFF. FROM ’16 RESULT | |
---|---|---|---|
Ariz. |
Data Orbital
Sept. 14-17, 550 L.V.
|
Biden +2
49-47
|
+6D |
Ga. |
New York Times/Siena College
Sept. 16-21, 523 L.V.
|
Even
45-45
|
+5D |
Ga. |
Data for Progress
Sept. 14-19, 800 L.V.
|
Even
46-46
|
+5D |
Iowa |
Monmouth University
Sept. 18-22, 402 L.V.
|
Trump +3
46-49
|
+6D |
Iowa |
New York Times/Siena College
Sept. 16-22, 501 L.V.
|
Biden +3
45-42
|
+12D |
Mich. |
YouGov
Sept. 10-21, 641 L.V.
|
Biden +6
51-45
|
+6D |
Mich. |
Data for Progress
Sept. 14-19, 455 L.V.
|
Biden +6
49-43
|
+6D |
Ohio |
Quinnipiac University
Sept. 17-21, 1,085 L.V.
|
Biden +1
48-47
|
+9D |
Pa. |
YouGov
Sept. 10-21, 642 L.V.
|
Biden +4
49-45
|
+5D |
Pa. |
Franklin & Marshall College
Sept. 14-20, 480 L.V.
|
Biden +6
48-42
|
+7D |
Texas |
New York Times/Siena College
Sept. 16-22, 653 L.V.
|
Trump +3
43-46
|
+6D |
Texas |
Quinnipiac University
Sept. 17-21, 1,078 L.V.
|
Trump +5
45-50
|
+4D |
Wis. |
YouGov
Sept. 10-21, 664 L.V.
|
Biden +4
50-46
|
+5D |
National polls
U.S. |
YouGov
Sept. 21-23, 1,125 L.V.
|
Biden +5
49-44
|
+3D |
U.S. |
USC Dornsife/Los Angeles Times
Sept. 10-23, 5,369 L.V.
|
Biden +10
52-42
|
+8D |
U.S. |
Harris Insights & Analytics
Sept. 19-21, 2,803 R.V.
|
Biden +5
45-40
|
+3D |
U.S. |
Global Strategy Group/GBAO/Navigator Research
Sept. 17-21, 1,230 R.V.
|
Biden +11
53-42
|
+9D |
If you’re a Biden supporter, you might have taken a measure of happiness from encouraging results in Ohio, Georgia and Iowa, including our own Times/Siena polls. You might also have fretted about your lowest national number from YouGov in a long time, or been disappointed that your dream of a Blue Texas was a little farther (down by three in our poll) from reality than you hoped. But the polls were almost exactly in line with the average in the most important battleground states, like Pennsylvania, Wisconsin and Arizona.
If you’re a Trump supporter, you may have found encouragement in YouGov numbers showing a mere five-point race nationwide and a tighter contest in Wisconsin and Michigan than other pollsters have found. But there was nothing that would have changed the fact that the president was behind in polling in states worth upward of 350 electoral votes, including Ohio and Iowa, which Mr. Trump carried easily four years ago.
Now about those Fox polls … The well-regarded Fox News polling team released three state surveys, in Nevada, Ohio and Pennsylvania. They were awfully good for Mr. Biden.
POLLSTER | MARGIN | DIFF. FROM ’16 RESULT | |
---|---|---|---|
Nev. |
Fox News/Beacon Research/Shaw & Co. Research
Sept. 20-23, 810 L.V.
|
Biden +11
52-41
|
+9D |
Ohio |
Fox News/Beacon Research/Shaw & Co. Research
Sept. 20-23, 830 L.V.
|
Biden +5
50-45
|
+13D |
Pa. |
Fox News/Beacon Research/Shaw & Co. Research
Sept. 20-23, 856 L.V.
|
Biden +7
51-44
|
+8D |
Fox has him up 11 in Nevada, a figure that evokes the 👀 emoji. Fox also has Mr. Biden up seven in Pennsylvania, the kind of result he had been looking for in recent days. And the poll showed Mr. Biden up five points in Ohio, a state Mr. Trump carried by eight points in 2016.
It’s great news for Mr. Biden, but it’s worth taking these results with a grain of salt. Fox News state polls have consistently had fantastic results for Mr. Biden, including near double-digit leads for him in Arizona and Wisconsin immediately after the Republican convention. And it’s also worth mentioning that the Fox polls tilted pretty far to the left in key 2018 Senate races.
Ohio, really? The biggest eye-popper is Ohio. Fox wasn’t alone in showing Mr. Biden ahead. Quinnipiac had him up by a point. Even if you assume these pollsters tend to produce results that tilt to the left, they still support the idea of a highly competitive race.
It’s been a bit hard to know what to expect from Ohio. There haven’t been many high-quality polls there in 2020, and the polls severely underestimated Republicans in 2016 and 2018. The state is traditionally highly competitive, but it swung so hard to Mr. Trump in 2016 that few would have put it toward center of the electoral map.
If you’ve been following along here, you’ll have noticed that many of the states with the biggest shifts toward Mr. Biden have been relatively white states in the North or Upper Midwest, like Maine, Minnesota and Wisconsin. Ohio isn’t exactly like those states, but the northwestern part of the state is somewhat similar to the rest of the Farm Belt, and the Trump campaign has aired plenty of advertisements in the state. Taken together, it seems we should consider pretty seriously that it’s a close race there at this point.
State of the race at the end of the day: Still solid for Biden, and if the Ohio polls are any indication, a little better for him than it was yesterday.
The rest of the country stays strong for Biden.
Over the last few weeks, Joe Biden has often seemed to do well in the state polls but not necessarily in the national polls. Today, it’s the opposite, and that makes it a fine day to roll out a new chart of the difference between our poll average nationwide and in the so-called tipping-point state:
Measuring Trump’s Electoral College edge
What’s the tipping-point state? It’s the one that would get either candidate his 270th electoral vote if he won every other state that’s more favorable to him. In 2016, it was Wisconsin: If Hillary Clinton had won every state where she was stronger than Wisconsin (the states she carried, plus Pennsylvania and Michigan), then it all would have come down to Wisconsin, the tipping point.
Back in 2016, Donald J. Trump won Wisconsin by 0.8 percentage points, while he lost the national vote by 2.1 points. That three-point difference is a measurement of the president’s relative edge in the Electoral College, and that’s about what our poll averages say today.
As you can see, the president’s relative edge in the Electoral College appeared to grow a bit Wednesday because Mr. Biden did well in the national polls but somewhat poorly in the state polls.
A good day for Trump in state polling. If you’ve been reading along recently, you know that Mr. Biden has had a good run of polls in Michigan and Wisconsin. That’s enough for 258 electoral votes — just short of the 270 needed to win — if you also give him the states carried by Mrs. Clinton.
From there, he would need one more state: maybe Arizona (in conjunction with Nebraska’s Second District or Maine’s Second District) or Florida or Pennsylvania. And while that might seem easy enough, President Trump has stayed somewhat competitive in polls of those states. On Wednesday, he was particularly competitive in these states:
POLLSTER | MARGIN | DIFF. FROM ’16 RESULT | |
---|---|---|---|
Ariz. |
Change Research
Sept. 18-20, 262 L.V.
|
Biden +6
49-43
|
+10D |
Ariz. |
ABC News/The Washington Post
Sept. 15-20, 579 L.V.
|
Trump +1
48-49
|
+3D |
Ariz. |
Ipsos
Sept. 11-17, 565 L.V.
|
Biden +1
47-46
|
+5D |
Fla. |
St. Pete Polls
Sept. 21-22, 2,906 L.V.
|
Biden +3
50-47
|
+4D |
Fla. |
Change Research
Sept. 18-20, 702 L.V.
|
Biden +3
49-46
|
+4D |
Fla. |
ABC News/The Washington Post
Sept. 15-20, 613 L.V.
|
Trump +4
47-51
|
+3R |
Fla. |
Ipsos
Sept. 11-16, 586 L.V.
|
Even
47-47
|
+1D |
Pa. |
Baldwin Wallace University
Sept. 8-22, 1,012 L.V.
|
Biden +2
47-45
|
+3D |
Pa. |
Change Research
Sept. 18-20, 579 L.V.
|
Biden +4
49-45
|
+5D |
Pa. |
CPEC
Sept. 15-17, 820 L.V.
|
Biden +5
50-45
|
+6D |
If you scan down the list, you won’t see too many polls suggesting that Mr. Biden is up by more than seven points (as he is nationwide) in Arizona, Florida or Pennsylvania. You’ll even see something we haven’t seen much of so far this year: the color red.
Trump’s best polls of the cycle? ABC/Washington Post found the president leading in Arizona and Florida. When you factor in the firm, the polls are arguably his best of the whole cycle.
The ABC/Post poll is a high-quality survey. It’s so good that it has tended to beat the average of all other polls in a race. But it can also be a somewhat noisy poll, especially at small sample sizes like these, since it doesn’t have many ways to control the composition of the sample, like weight on party registration. (You may recall an eyepopper in the other direction from the ABC/Post poll last week, showing Mr. Biden with a 16-point lead in Minnesota. Same idea.)
Outlying results are a common — and expected — part of polling, even from the best pollsters. That’s part of why the average of all polls is helpful. The average shows Mr. Biden with a slight or modest advantage in both Arizona and Florida, and he has similar edges there in the other polls published today.
But there’s no way to confuse these polls with the kind of results we’ve seen nationwide or in Wisconsin. And until that’s true, the president’s path to re-election will be a bit easier than the national polls suggest.
A good day for Biden everywhere else. Well, everything else looked pretty good for Mr. Biden.
State polls
POLLSTER | MARGIN | DIFF. FROM ’16 RESULT | |
---|---|---|---|
Ga. |
Monmouth University
Sept. 17-21, 402 L.V.
|
Trump +3
46-49
|
+2D |
Mich. |
Baldwin Wallace University
Sept. 8-22, 1,001 L.V.
|
Biden +8
50-42
|
+8D |
Mich. |
Change Research
Sept. 18-20, 568 L.V.
|
Biden +8
51-43
|
+8D |
N.C. |
Change Research
Sept. 18-20, 579 L.V.
|
Biden +2
48-46
|
+6D |
N.C. |
Harper Polling
Sept. 17-20, 612 L.V.
|
Trump +1
44-45
|
+3D |
Ohio |
Baldwin Wallace University
Sept. 8-22, 1,011 L.V.
|
Biden +1
45-44
|
+9D |
Wis. |
Baldwin Wallace University
Sept. 8-22, 863 L.V.
|
Biden +9
50-41
|
+10D |
National polls
U.S. |
YouGov
Sept. 20-22, 1,124 L.V.
|
Biden +7
49-42
|
+5D |
U.S. |
Rasmussen Reports/Pulse Opinion Research
Sept. 16-22, 3,000 L.V.
|
Biden +1
48-47
|
+1R |
U.S. |
USC Dornsife/Los Angeles Times
Sept. 9-22, 5,440 L.V.
|
Biden +10
52-42
|
+8D |
U.S. |
Quinnipiac University
Sept. 17-21, 1,302 L.V.
|
Biden +10
52-42
|
+8D |
U.S. |
Qriously
Sept. 17-20, 2,134 L.V.
|
Biden +7
46-39
|
+5D |
U.S. |
Marquette University Law School
Sept. 8-15, 1,357 L.V.
|
Biden +10
50-40
|
+8D |
Almost all the national polls showed him ahead by at least seven points. A Monmouth poll showed the president with a modest lead in Georgia, which counts as a pretty good result for Mr. Trump at this point. We’ll get another reading on the race there — as well as in Iowa and Texas — from our Times/Siena College polls tomorrow.
State of the race at the end of the day Today was not a good day for Mr. Biden by the measure that matters most: whether he has a clear and comfortable lead in states worth 270 electoral votes. Even so, he still leads in our average of Florida, Arizona and Pennsylvania — and therefore in the race for the White House.
A landslide and a close race are both easy to imagine.
State polls
POLLSTER | MARGIN | DIFF. FROM ’16 RESULT | |
---|---|---|---|
Ga. |
University of Georgia
Sept. 11-20, 1,150 L.V.
|
Even
47-47
|
+5D |
Iowa |
Selzer & Co.
Sept. 14-17, 658 L.V.
|
Even
47-47
|
+9D |
Mich. |
Marketing Resource Group (MRG)
Sept. 14-19, 600 L.V.
|
Biden +5
46-41
|
+5D |
S.C. |
Morning Consult
Sept. 11-20, 764 L.V.
|
Trump +6
44-50
|
+8D |
National polls
U.S. |
USC Dornsife/Los Angeles Times
Sept. 8-21, 5,482 L.V.
|
Biden +10
52-42
|
+8D |
U.S. |
Léger
Sept. 18-20, 830 L.V.
|
Biden +7
48-41
|
+5D |
U.S. |
Morning Consult
Sept. 18-20, 12,965 L.V.
|
Biden +7
51-44
|
+5D |
We often focus on the battleground states that decided the last election and seem likeliest to decide the next one. Today, we got polls from two states that Donald J. Trump won handily in 2016, and they’re an important reminder of the wide range of possibilities in this election.
An even race in Iowa and Georgia. We haven’t had much high-quality polling in either Iowa or Georgia recently, but we got one for each state Tuesday: from Ann Selzer in Iowa and from The Atlanta Journal-Constitution in Georgia (conducted by the University of Georgia). They found the same result: a tie.
The Des Moines Register poll of Iowa, conducted by Ms. Selzer, is always highly anticipated: She is one of the most respected pollsters in the country. This time, it comes as the Biden campaign begins airing advertisements in the state. It’s not hard to see why.
What the results don’t have in common. These polls share something important: They show tied races in states that Mr. Trump won fairly comfortably in 2016 (by nine in Iowa and by five in Georgia). But what’s behind these shifts is quite different.
In Iowa, Joe Biden seems to be securing large, broad gains among white voters. He’s benefiting from a similar shift across the Northern and mostly white battleground states — think Minnesota, Wisconsin, Michigan and Maine.
In Georgia, there’s a shift among college-educated white voters, particularly in the well-educated suburbs of Atlanta. But the state’s white working-class population remains staunchly Republican. Unlike in Iowa, demographic changes are also on Mr. Biden’s side. Georgia is growing fast, and it’s increasingly diverse.
Landslide? Georgia and Iowa might be competitive, but for Mr. Biden victories there would probably merely be icing on the cake: If he has won them, he’s almost certainly already won other battleground states like Florida and Pennsylvania, on track to a total nearing 400 electoral votes.
That’s a genuine possibility. We’re all understandably focused on the states likeliest to decide the election, like Pennsylvania and Florida. Those states remain close enough that Mr. Trump remains competitive. But states like Texas, Georgia, Ohio and Iowa also remain very competitive. In fact, they’re even closer than Florida or Pennsylvania — so close that a Biden landslide is just as real a possibility as a Trump victory.
Here’s a different way to think about it: If Mr. Biden outperformed today’s polls by just two points, he would be declared the winner early on election night. Florida would be called by around 8 p.m., and Texas could be the state that makes Mr. Biden the president-elect. (Yes, Texas). He’d have a good shot at the largest electoral vote landslide since 1988.
But if Mr. Trump outperformed the polls by the same margin, suddenly we’d have an extraordinarily close race on our hands, potentially waiting days or weeks while mail-in votes were counted in Pennsylvania, Michigan and Wisconsin.
How is it possible for the race to teeter on the edge between a landslide and a close contest? A lot of it depends on the president’s apparent resilience in Pennsylvania and Florida. There aren’t a ton of polls in these states, so it’s possible that Mr. Biden has a wider lead than we think. But Mr. Biden holds just under a five-point lead in Pennsylvania and a two-point lead in Florida, according to our averages, even as he fights to a draw in places where Democrats haven’t won in decades.
If Mr. Biden has gained as much in Pennsylvania and Florida as it seems he has in Iowa and Texas, this would be a different story.
Not much new nationally. We did see a pair of national polls with Mr. Biden up by seven points, more or less consistent with our average.
State of the race at the end of the day: Mr. Biden’s leading in major battlegrounds, and even in a few states that Mr. Trump won handily four years ago. A landslide is still possible. But until Mr. Biden claims a more consistent and wider lead in a state like Pennsylvania — more like his seven-point lead nationwide — Mr. Trump will remain more competitive than you might guess from the steady stream of poll results hinting at a possible blowout.