The Upshot on Today’s Polls
Analysis of the day in polls, from now until Election Day.
| Day | Place | Result |
|---|---|---|
| Tuesday | U.S. | Biden +9 |
| Tuesday | Georgia | Even |
| Wednesday | Iowa | Biden +3 |
| Thursday | Kansas | Trump +7 |
| Fri. 1 p.m. ET | Montana | – |
| Next | Texas | – |
| Polling leader | If polls are as wrong as they were in… | ||||
|---|---|---|---|---|---|
| 2016 | 2012 | ||||
| U.S. | +9 Biden | +6 | +12 | ||
| N.H. | +11 Biden | +7 | +14 | ||
| Mich. | +8 Biden | +4 | +13 | ||
| Minn. | +8 Biden | +1 | +10 | ||
| Nev. | +7 Biden | +8 | +9 | ||
| Neb. 2* | +6 Biden | +8 | +1 | ||
| Wis. | +6 Biden | +1 | +9 | ||
| Pa. | +6 Biden | <1 | +7 | ||
| Ariz. | +4 Biden | +2 | +3 | ||
| Maine 2* | +4 Biden | +7 | +9 | ||
| Fla. | +3 Biden | <1 | +4 | ||
| Iowa | +3 Biden | +5 | +6 | ||
| N.C. | +2 Biden | +3 | +2 | ||
| Ga. | <1 Biden | +2 | <1 | ||
| Texas | +2 Trump | +4 | +3 | ||
| Ohio | +2 Trump | +9 | +2 | ||
Trump faces daunting deficit in campaign’s final 11 days.
The final presidential debate was President Trump’s last, best chance to reshape the race in his favor. The post-debate instant-reaction polls suggest that he didn’t get what he needed.
Instant Poll Surveys of Who Won the Debate
What’s an instant poll? As we discussed after the first debate, instant-reaction or instapolls are a little quirky, but they’re useful. They usually represent only people who watched the debate, who aren’t the same as the overall electorate. The people who participate probably aren’t even very representative of people who did watch the debate, since they agreed to participate in a survey multiple times — and followed through. Pollsters try to deal with these potential biases in different ways, potentially yielding very different results. Nonetheless, a candidate who decisively wins in the instant polls usually gains in the traditional survey results that follow.
Double-digit Biden win in instant polls. This time, the instant polls had a consistent and clear story: Mr. Biden won the debate by a double-digit margin.
If the results in the chart look familiar, that’s because they’re reminiscent of many national survey results we’ve seen over the last month. You could interpret that to mean that the debate was more or less a draw on the merits, with most Biden supporters concluding that Mr. Biden won, and vice versa for the Trump supporters. Indeed, the CNN/SSRS poll didn’t find that either candidate improved his favorability rating compared with interviews conducted before the debate.
Not the huge win Trump needed. For Mr. Trump, even a draw would have been a blow to his chances. He trails by nearly 10 percentage points in national surveys, and although he still has a chance in the most crucial battleground states, like Florida and Pennsylvania, his path to an Electoral College victory remains narrow. With just 11 days to go, there aren’t many obvious opportunities remaining for him to change the attitudes of voters.
Maybe the post-debate coverage will focus on something that could help the president in certain battleground areas, like Mr. Biden’s comments about transitioning away from the oil industry. But the president had his own potentially damaging comments, such as his reaction to separating children from their parents at the border. Maybe there will be another big news event over the final stretch. (It was at this point in the 2016 campaign — with 11 days to go — that the F.B.I. director James Comey sent a letter to Congress about new evidence in the investigation into Hillary Clinton’s emails.) Or maybe things will just naturally tighten on their own, which happened after the third debate in 2016, even before the Comey letter and even though Mrs. Clinton won the post-debate instant polls after the final debate.
But if the polls do not tighten significantly over the final stretch, Mr. Trump will be left in an unenviable position. He will once again be left to hope for a large, systematic error in the polling, this time dwarfing the one that barely got him to victory four years ago in a much closer race.
Oct. 23
A running table of all the presidential polls released today, with newest entries first:
| Pollster | Margin | Diff. from ’16 result | |
|---|---|---|---|
| Fla. |
Rasmussen Reports/
Pulse Opinion Research 10:16 AM ET
Oct. 20-21, 800 L.V.
|
Trump +4
46-50
|
+3R |
| Mich. |
Zia Poll
9:39 AM ET
Oct. 11-18, 2,851 L.V.
|
Trump +4
45-49
|
+4R |
| U.S. |
Data for Progress (Dem. pollster)
9:13 AM ET
Oct. 20-20, 811 L.V.
|
Biden +10
54-44
|
+8D |
| Utah |
RMG Research
9:13 AM ET
Oct. 12-17, 1,000 L.V.
|
Trump +12
38-50
|
+6D |
| U.S. |
IBD/TIPP
9:02 AM ET
Oct. 18-22, 972 L.V.
|
Biden +4
50-46
|
+2D |
| Fla. |
St. Pete Polls
8:38 AM ET
Oct. 21-22, 2,527 L.V.
|
Biden +2
49-47
|
+3D |
| Pa. |
Muhlenberg College
7:51 AM ET
Oct. 13-20, 416 L.V.
|
Biden +7
51-44
|
+8D |
| U.S. |
U.S.C. Dornsife
7:51 AM ET
Oct. 9-22, 5,278 L.V.
|
Biden +12
54-42
|
+10D |
| Mich. |
EPIC-MRA
7:50 AM ET
Oct. 15-19, 600 L.V.
|
Biden +9
48-39
|
+9D |
| U.S. |
GSG/
GBAO/ Navigator 7:50 AM ET
Oct. 15-19, 1,000 R.V.
|
Biden +10
53-43
|
+8D |
Advertisement
Trump ahead by single digits in a reliably red state; Republican holds narrow edge in Senate contest.
My colleagues Trip Gabriel and Isabella Grullón Paz have the story.
New York Times/Siena College poll of likely voters in Kansas
Oct. 22
A running table of all the presidential polls released today, with newest entries first:
A clear picture emerges.
There were a lot of high-quality polls today. They didn’t always agree, but together they tell a very clear story about the state of the race heading into the second and final presidential debate Thursday: Joe Biden still holds a significant lead, even if it has come down a bit from his post-debate peak.
State polls
National polls
| U.S. |
YouGov
Oct. 18-20, 1,344 L.V.
|
Biden +9
52-43
|
+7D |
| U.S. |
IBD/TIPP
Oct. 16-20, 1,016 L.V.
|
Biden +3
49-46
|
<1D |
| U.S. |
Ipsos
Oct. 16-20, 949 L.V.
|
Biden +9
51-42
|
+7D |
| U.S. |
Rasmussen Reports/
Pulse Opinion Research Oct. 14-20, 2,500 L.V.
|
Biden +3
49-46
|
<1D |
| U.S. |
U.S.C. Dornsife
Oct. 7-20, 5,406 L.V.
|
Biden +12
54-42
|
+10D |
| U.S. |
SurveyUSA
Oct. 16-19, 1,136 L.V.
|
Biden +10
53-43
|
+8D |
| U.S. |
Qriously
Oct. 15-19, 2,731 L.V.
|
Biden +12
51-39
|
+10D |
| U.S. |
Research Co.
Oct. 16-18, 1,035 L.V.
|
Biden +8
50-42
|
+6D |
| U.S. |
PureSpectrum
Aug. 07-Sept. 27, 26,838 L.V.
|
Biden +10
50-40
|
+8D |
| U.S. |
Harris Insights & Analytics
Oct. 13-15, 1,897 R.V.
|
Biden +4
46-42
|
+2D |
The big picture. Mr. Biden enters the debate up by nine percentage points nationwide, including a lead of five points or more in states worth more than the 270 electoral votes needed to win. He leads, even if only narrowly in some cases, in states worth more than 350 electoral votes, and he has a chance to win a 400-plus electoral vote landslide, including Texas (a Quinnipiac poll showed Mr. Biden tied with President Trump in the Lone Star State).
A silver lining for Trump. As we said yesterday, Mr. Biden’s lead has seemed to decline a bit over the last week or so, perhaps because the news of Mr. Trump’s conduct at the first debate and his hospitalization with coronavirus has faded. The national polls certainly look better for Mr. Trump than they did in the last few weeks, when it sometimes seemed as if we could go days without any survey showing Mr. Biden with less than a nine-point lead.
An unclear picture in Pennsylvania. Pennsylvania is probably the most crucial battleground state, and on Wednesday we got four polls showing Mr. Biden ahead there by an average of seven points. But if you look more carefully, there are at least a few reasons these polls weren’t quite as great for him as the topline numbers suggest, and they’re consistent with the broader evidence of modest tightening in the race.
For one, Fox, Quinnipiac and CNN/SSRS are usually more favorable to Mr. Biden than average. And Mr. Biden’s lead was three and five points smaller than earlier Fox and Quinnipiac surveys in the state, conducted in September and early October. Our poll averages care a lot about these factors — change since a prior poll, and a pollster’s lean with respect to other surveys — and Mr. Biden’s lead in the state took a bit of a dive as a result.
There are other reasons to be cautious. The Suffolk poll isn’t weighted by self-reported education; all else being equal, that tends to inflate Mr. Biden’s standing. And the CNN/SSRS poll showed Mr. Biden up by only five points among registered voters, a far tighter race than his 10-point lead among likely voters. Now, likely voters are what count — both in our averages and ultimately on Election Day, should they become real voters. But a five-point gap between the preference of registered and likely voters is extremely hard to imagine in a high-turnout election. There are some exceptions — like a poll showing a notably pro-G.O.P. advantage among likely voters in a heavily Latino state (Latino voters tend to have lower turnout). But for the most part, the registered voter number represents a fairer read of a poll result if the gap between registered and likely voters gets so large.
Despite all these caveats, Mr. Biden still ends the day up by six points in our average in Pennsylvania. It may not be the big lead he seemed to hold around the first debate, when Quinnipiac showed him up by 13 points, Monmouth by 11 points, ABC/Post and NYT/Siena by nine points, and so on. But it’s a strong position in the most important state.
A simpler story in Iowa. In contrast with the murky picture in Pennsylvania, Iowa offered only positive news for Mr. Biden. Three polls showed a close race or a Biden lead, including a three-point lead in our Times/Siena poll and a four-point lead in Monmouth’s poll (averaging their two likely-voter scenarios).
Mr. Trump won the state by more than nine points in 2016. If Mr. Biden is ahead here, he’s on the way to a decisive national victory.
Splitting the difference everywhere else. You might think there’s some tension between the polling out of Iowa and Pennsylvania, and I think that’s fair. But it’s important to look at the balance of the evidence.
Besides the Iowa polls, there were other great results for Mr. Biden that would have fit in quite well with the last two weeks of data, like the Quinnipiac poll showing him tied in Texas or the Fox poll in Michigan (Biden plus-12). On the other hand, there were other polls that fit well with the Pennsylvania data, like the Fox results in Ohio (Trump plus-3) and Wisconsin (Biden plus-5).
So yes, some polls were consistent with a landslide, while others showed Mr. Biden with a clear but not overwhelming lead. That’s what you’ll get if Mr. Biden’s up by around nine points, as our national polling average suggests.
Thursday, 1 p.m. Eastern NYT/Siena in Kansas. (I know we announced Kansas would be released today. Sorry about the mix-up. I can assure you that you’ll see it tomorrow. As they say in Kansas, “To the Stars Through Difficulties.”)
Thursday night. The final debate. If today’s polls are any indication, Mr. Trump needs something big. We’ll get some of the so-called instant reaction polls to help us see whether he got that breakthrough.
Friday, 1 p.m. Eastern. NYT/Siena in Montana.
State of the race: Not great for Mr. Trump with 13 days before the election.
Advertisement
Biden holds a narrow lead, while the high-stakes Senate contest is essentially tied.
My colleague Jonathan Martin has the story.
New York Times/Siena College poll of likely voters in Iowa
Oct. 21
A running table of all the presidential polls released today, with newest entries first:
But ‘best’ still doesn’t mean ‘good.’
President Trump has had a terrible few weeks of polls. On Tuesday, he probably had his best day of polling since the first presidential debate in late September.
State polls
| Pollster | Margin | Diff. from ’16 result | |
|---|---|---|---|
| Ariz. |
RMG Research
Oct. 14-19, 800 L.V.
|
Biden +1
47-46
|
+5D |
| Ariz. |
Data Orbital
Oct. 16-18, 550 L.V.
|
Biden +5
47-42
|
+9D |
| Colo. |
RMG Research
Oct. 9-15, 800 L.V.
|
Biden +8
51-43
|
+3D |
| Fla. |
University of North Florida
Oct. 12-16, 863 L.V.
|
Biden +1
48-47
|
+2D |
| Ga. |
New York Times/
Siena College Oct. 13-19, 759 L.V.
|
Even
45-45
|
+5D |
| Iowa |
Opinion Savvy/
InsiderAdvantage Oct. 18-19, 400 L.V.
|
Even
45-45
|
+9D |
| Mich. |
Ipsos
Oct. 14-20, 686 L.V.
|
Biden +8
52-44
|
+8D |
| Minn. |
Change Research
Oct. 12-15, 1,021 L.V.
|
Biden +5
49-44
|
+3D |
| N.C. |
Ipsos
Oct. 14-20, 660 L.V.
|
Biden +2
49-47
|
+6D |
| N.C. |
East Carolina University
Oct. 15-18, 1,155 L.V.
|
Biden +4
51-47
|
+8D |
| N.C. |
ABC News/
The Washington Post Oct. 12-17, 646 L.V.
|
Biden +2
50-48
|
+6D |
| Ohio |
Rasmussen Reports/
Pulse Opinion Research Oct. 18-19, 800 L.V.
|
Biden +1
48-47
|
+9D |
| Pa. |
Rasmussen Reports/
Pulse Opinion Research Oct. 18-19, 800 L.V.
|
Biden +3
50-47
|
+4D |
National polls
| U.S. |
IBD/TIPP
Oct. 15-19, 1,046 L.V.
|
Biden +3
49-46
|
<1D |
| U.S. |
U.S.C. Dornsife
Oct. 6-19, 5,488 L.V.
|
Biden +12
54-42
|
+10D |
| U.S. |
Redfield & Wilton Strategies
Oct. 17-18, 2,915 L.V.
|
Biden +11
51-40
|
+9D |
| U.S. |
Léger
Oct. 16-18, 821 L.V.
|
Biden +9
50-41
|
+7D |
| U.S. |
New York Times/
Siena College Oct. 15-18, 987 L.V.
|
Biden +9
50-41
|
+7D |
| U.S. |
Ipsos
Sept. 22-Oct. 8, 2,004 adults
|
Biden +7
46-39
|
+5D |
Trump’s not-so-bad day. The president might not have led in any poll today, but almost every survey was better than the polling average for him. Three stand out: the ABC/Post poll showing Joe Biden up in North Carolina by one or two points, depending on the inclusion of minor-party candidates; the UNF poll, which showed Mr. Biden up by just a point in Florida; and The New York Times/Siena poll, which showed Mr. Biden up by nine points nationwide.
None of these results are good for the president. But they are all better for him than the average, and they came from high-quality, live-interview telephone polls. In general, these types of surveys have been Mr. Biden’s best. In fact, the Times/Siena poll is the first national live-interview survey to show a single-digit race since the first debate.
‘Best’ is relative. My best game as a quarterback in middle school football was getting shut out, 7-0. True story, and I was pretty proud of it. Here again, Trump fans have something to be pleased about, but they didn’t put any points on the board. He didn’t lead in any battleground or national survey conducted today.
And while most of the results were better than average for the president, there were still plenty that were flat-out poor, like trailing by three points in North Carolina, per Ipsos, or trailing in a Rasmussen poll of Ohio. He was even tied in a Republican poll of Iowa.
Is it a real change? You can go nuts trying to answer this question based on one day of polling, but I would note that there haven’t been many of those eye-popping big leads for Mr. Biden lately. And some recent results have seemed pretty underwhelming for Mr. Biden compared with the average, like the CBS/YouGov polls of Wisconsin and Arizona on Sunday. As a result, our poll averages have started to move a bit back toward Mr. Trump. They haven’t moved much; they’ll need to see more data before they’re convinced the race has tightened by a point or two.
If the race does move a point or two back toward the president, I don’t think we’d be short of an explanation. It’s a natural moment for some of the president’s more reluctant supporters to finally get back on board. The brutal run of bad news for Mr. Trump at the start of the month — the debate, a positive test for the coronavirus — has seemed to come to an end. Over the last week, the news was arguably favorable to him. The hearings for his Supreme Court nominee, Amy Coney Barrett, put her on course to a confirmation, with majority public support. The New York Post story about Hunter Biden, although questioned within the paper’s own newsroom, has seemed to energize Mr. Trump’s base. Together, it might be just enough to reset the environment and allow some truly hesitant Trump supporters to return to his side, where they were before the debate. There’s a long history of voters returning to their preferred party as the election nears.
Poll watch. On Wednesday, we get a Monmouth University poll in Iowa and a Suffolk University poll in Pennsylvania. And at some point over the coming days, we’ll have Times/Siena polls in Iowa, Kansas and Montana.
State of the race: It’s still Mr. Biden’s to lose.
Advertisement
Biden and Trump tied as suburbs shift toward Democrats up and down the ballot.
We have the story over at The Upshot.
New York Times/Siena College poll of likely voters in Georgia
Oct. 20
A running table of all the presidential polls released today, with newest entries first:
Biden leads by nine points.
My colleagues Alexander Burns and Jonathan Martin have the story.
New York Times/Siena College poll of likely voters in U.S.
Advertisement
A quiet but not empty day.
State polls
| Pollster | Margin | Diff. from ’16 result | |
|---|---|---|---|
| Colo. |
YouGov
Oct. 5-9, 800 L.V.
|
Biden +9
47-38
|
+4D |
| Mich. |
Mitchell Research & Communications
Oct. 18-18, 900 L.V.
|
Biden +10
51-41
|
+10D |
| Pa. |
Ipsos
Oct. 13-19, 653 L.V.
|
Biden +4
49-45
|
+5D |
| Pa. |
Trafalgar Group
Oct. 13-15, 1,041 L.V.
|
Biden +2
48-46
|
+3D |
| Va. |
Cygnal
Oct. 9-11, 607 L.V.
|
Biden +9
51-42
|
+4D |
| Wis. |
Ipsos
Oct. 13-19, 663 L.V.
|
Biden +7
51-44
|
+8D |
| Wis. |
Trafalgar Group
Oct. 14-16, 1,051 L.V.
|
Biden +2
48-46
|
+3D |
National polls
| U.S. |
YouGov
Oct. 16-18, 400 L.V.
|
Biden +11
51-40
|
+9D |
| U.S. |
Morning Consult
Oct. 16-18, 14,994 L.V.
|
Biden +9
52-43
|
+7D |
| U.S. |
IBD/TIPP
Oct. 14-18, 949 L.V.
|
Biden +6
50-44
|
+4D |
| U.S. |
U.S.C. Dornsife
Oct. 5-18, 5,557 L.V.
|
Biden +13
54-41
|
+11D |
| U.S. |
RMG Research
Oct. 15-17, 1,265 L.V.
|
Biden +8
51-43
|
+6D |
| U.S. |
Public Religion Research Institute
Oct. 9-12, 591 L.V.
|
Biden +16
55-39
|
+14D |
It was another fairly quiet day of polling. We did get a handful of polls in the Northern battlegrounds, where Joe Biden still leads in states that President Trump won four years ago.
The polls Monday didn’t necessarily change the big picture, but they were a bit better for Mr. Trump than the average. Mostly these came from two firms — Reuters/Ipsos and Trafalgar Group.
Ipsos polls in Wisconsin and Pennsylvania. Let’s start with the results from Reuters/Ipsos, which had Mr. Biden ahead by four points in Pennsylvania and by six or eight points in Wisconsin, depending on the inclusion of minor-party candidates (and therefore seven, on average, in our chart). On balance, that’s a bit better for Mr. Trump than average, with the Wisconsin result right on the average and the Pennsylvania result looking a lot better for Mr. Trump. I don’t recommend thinking too hard about any single result. For now, I’d be inclined to take them together rather than dwell too much on either individually.
Quick sidebar about Pennsylvania. The strong Ipsos result for the president in Pennsylvania reminds me that we’re in a bit of a polling drought there. You might recall that there were a lot of polls of the state before and after the first debate, but not many since. I think there’s a pretty good reason for that: Most pollsters will want to conduct surveys in Pennsylvania, perhaps the most important state, right before the election. That means they’re less likely to poll Pennsylvania during mid-October. We’re probably just going to have to wait to the end to get a wave of results and a good read.
What about Trafalgar? Trafalgar released results showing a closer-than-average race in Wisconsin and Pennsylvania, but in this case it’s just the norm: Trafalgar usually shows Mr. Trump doing about six points better than average, so Mr. Biden’s two-point leads here are consistent with his leads in our averages.
It’s worth pointing out that Trafalgar doesn’t really meet any of the basic conditions for transparency in public opinion research: They don’t explain the methodology; they don’t provide full questionnaires; there are no crosstabs; there’s only a very limited look at the composition of the electorate, and so on. All we really have are bits and pieces.
The Hill reported that the poll weights to account for “social desirability bias.” In other words, they’re looking for “Shy Trump” voters, but it’s not clear what that means in practice.
Another tidbit about Trafalgar’s methodology from 2016: It described its published ballot test as a “combination of survey responses to both a standard ballot test and a ballot test gauging where respondents’ neighbors stand,” which was supposed to address the “bias of traditional polling, wherein respondents are not wholly truthful about their position regarding highly controversial candidates.” As far as I’m concerned, that wouldn’t even count as a poll — even if it turned out to still be useful or predictive.
With that said, I’d draw your attention to a recent interview in National Review with Robert Cahaly, who runs Trafalgar’s polling. There’s no mention of utilizing how a respondent’s neighbors will vote or weighting to avoid social desirability bias. Instead there’s quite a bit that’s consistent with rigorous survey research: the importance of reaching low-turnout voters, the advantage of a shorter questionnaire, the problems with the exit polls, the advantages of the voter file, and the benefit of relying a little more on a well-balanced sample than on a lot of weighting. All of this makes sense, but none of it explains Mr. Trump’s persistent strength in Trafalgar’s polls, and neither does its decision to eschew live-interview telephone surveys. There are lots of pollsters that employ some or all of these methods and that still show Mr. Biden ahead.
In the end, I’m still left with very little idea of what’s going on here that might explain its results — other than its reference to a “proprietary digital technology that we don’t explain.”
Maybe Trafalgar — and Trafalgar alone — has figured it all out. The burden of proof is on them, though. I’m open-minded! But it’s hard to take someone terribly seriously if they tell you they’ve pulled off a feat no one else has achieved — say, running a 90-minute marathon — and then won’t tell you how it was done or show you any serious evidence.
Nationally on Monday. Nothing out of the ordinary. Today’s polls average to a 10.5-point national lead for Mr. Biden, more or less where he was when the day started.
Tuesday, 5 a.m. Eastern: A national NYT/Siena poll.
Tuesday, 1 p.m. Eastern. An NYT/Siena poll of Georgia.
Later in the week: NYT/Siena in Iowa, Kansas and Montana. Then, the last debate.
Oct. 19
A running table of all the presidential polls released today, with newest entries first:
A dull weekend, but things will heat up soon.
State polls
| Pollster | Margin | Diff. from ’16 result | |
|---|---|---|---|
| Ariz. |
YouGov
Oct. 13-16, 1,074 L.V.
|
Biden +3
50-47
|
+7D |
| Wis. |
YouGov
Oct. 13-16, 1,112 L.V.
|
Biden +5
51-46
|
+6D |
National polls
| U.S. |
IBD/TIPP
Oct. 13-17, 1,021 L.V.
|
Biden +6
50-44
|
+4D |
| U.S. |
U.S.C. Dornsife
Oct. 4-17, 5,674 L.V.
|
Biden +13
54-41
|
+11D |
If you, as a poll lover, spent your weekend indoors on a beautiful fall day, working and waiting for polls — I’m speaking only hypothetically — I’m sorry to report that you were not rewarded for your patience. There weren’t many polls, and those we did receive didn’t change our understanding of the race.
Why didn’t we get anything? You can blame the second presidential debate. It was supposed to be last week — instead, we had the dueling town halls Thursday — so most pollsters would have wanted to wrap up their surveys on Wednesday or start anew on Friday. A poll that started on Friday wouldn’t be done until early this week.
What we did get. To repeat, not much. The main event: YouGov/CBS polls of Arizona and Wisconsin. These were relatively solid results for President Trump compared with our averages, with Joe Biden ahead by just three in Arizona and by five in Wisconsin.
That said, one thing to note about CBS/YouGov polls is their stability: I don’t believe they’ve ever had Mr. Biden ahead by more than three in Arizona or six in Wisconsin. The results Sunday are right about at those highs, even though they’re worse than the average of other polls. So while I don’t think they count as great results for Mr. Biden, they also don’t really count as evidence of a shift in the race.
The weeks ahead. As dull as the weekend was in the world of polling, I expect things to heat up in the days ahead. With many pollsters beginning surveys on Friday, we’ll probably see a steady increase in the number of polls early this week, peaking right ahead of the debate Thursday.
Then you can expect another lull, as we wait for pollsters to wrap up the surveys that began immediately after that debate. Then the fun begins: a big wave of final polls from just about every pollster, starting maybe a week from Tuesday right up until the election. That ought to give us a clear picture of the race heading into election night.
Oddly, the calendar (we’re 16 days away from the election) could cut against the value of this week’s surveys. If you’re a pollster, would you rather have your final Pennsylvania poll this week, or the week before the election? You’re probably not doing both.
Early this week: NYT/Siena in Georgia and the U.S. We’ll have more to say about timing soon, but probably not on Monday.
Then! Iowa, Kansas and hopefully another, all before the final debate.
Advertisement
Oct. 18
A running table of all the presidential polls released today, with newest entries first:
Oct. 17
A running table of all the presidential polls released today, with newest entries first:
Polling firms that tend to lean toward the president.
Today was a mostly quiet day in the world of polling, but with results from some pollsters we don’t usually comment on much. So what did we get?
State polls
| Pollster | Margin | Diff. from ’16 result | |
|---|---|---|---|
| Alaska |
New York Times/
Siena College Oct. 9-14, 423 L.V.
|
Trump +6
39-45
|
+9D |
| Fla. |
Harris Insights & Analytics
Oct. 12-15, 965 L.V.
|
Even
48-48
|
+1D |
| Fla. |
Mason-Dixon Polling & Strategy
Oct. 8-12, 625 L.V.
|
Biden +3
48-45
|
+4D |
| Mich. |
Harris Insights & Analytics
Oct. 12-15, 1,289 L.V.
|
Biden +11
54-43
|
+11D |
| Mich. |
Trafalgar Group
Oct. 11-14, 1,025 L.V.
|
Trump +1
46-47
|
<1R |
| Pa. |
Harris Insights & Analytics
Oct. 12-15, 992 L.V.
|
Biden +5
51-46
|
+6D |
National polls
| U.S. |
Ipsos
Oct. 13-15, 920 L.V.
|
Biden +10
51-41
|
+8D |
| U.S. |
IBD/TIPP
Oct. 11-15, 951 L.V.
|
Biden +6
49-43
|
+4D |
| U.S. |
U.S.C. Dornsife
Oct. 2-15, 5,543 L.V.
|
Biden +13
54-41
|
+11D |
| U.S. |
Kaiser Family Foundation
Oct. 7-12, 1,015 L.V.
|
Biden +11
49-38
|
+9D |
| U.S. |
GSG/
GBAO/ Navigator Oct. 8-12, 1,003 R.V.
|
Biden +10
53-43
|
+8D |
| U.S. |
AP-NORC
Oct. 8-12, 1,121 adults
|
Biden +15
51-36
|
+13D |
It was a somewhat curious batch of polls. If you didn’t look closely, you might view it as one of the president’s best polling days in weeks, at least at the state level. But it turned out that many of the state results Friday were from the pollsters who tend to give President Trump his best results.
A better day for President Trump? An even race in Florida? Up one in Michigan? Down just five in Pennsylvania? Those are poll results President Trump would be thrilled to have right now, as most polls show him at a significant disadvantage nationwide and in the critical battleground states. But for the most part, these aren’t a relatively favorable set of poll results for him — they’re simply results from a relatively favorable set of polling firms.
Trafalgar Group showed Mr. Trump up by one point in Michigan, but Trafalgar lands pretty consistently about six points to the right of the national average. It’s the clearest partisan tilt of any polling firm this cycle.
HarrisX doesn’t lean quite as far to the right as Trafalgar, but it has been about four points to the right of the national average since the first debate. Over that period, it has had two national polls showing Joe Biden up just five and seven points. In that context, its polls showing Mr. Biden up by 11 in Michigan, five in Pennsylvania and tied in Florida seem pretty standard — and that Michigan result would be great for Mr. Biden by any standard.
Even if they were accurate. Just pretend for a second that these better polls for the president were dead-on. Even then, Mr. Biden is ahead.
Another strong day for Biden nationwide. We’re getting pretty used to national polls showing Mr. Biden up by at least 10 points. All but one national poll fell into that category Friday. That one exception — an IBD/TIPP poll showing a six-point Biden lead — was a bit closer than we’ve seen in a while, but it’s a three-point improvement from an earlier survey by the firm.
Still, of all the polls Friday, this IBD/TIPP poll was the one that could be most easily taken as a sign of a somewhat tighter race, as the poll was also conducted relatively recently, from Oct. 11-15. But it’s still just one poll — and a fairly unusual one that has tended to be better for the president than most national polls. It’s an odd mix of telephone and online samples, and it’s not very transparent about its approach.
And there happened to be another recent national poll that was just as recent: Biden +10 from Ipsos.
A close race in Alaska. We released this at 1 p.m. sharp today; I did not have to eat my hat.
Next week. Georgia and a national poll. We’ll try to squeeze at least another survey in before the debate on Thursday, but we’ll see how productive we are over the next few days.
State of the race at the end of the day. Not much different than where we started. And the election is a day nearer.
Advertisement
Trump holds six-point lead. In the Senate contest, the Republican Dan Sullivan is up by eight.
We have the story over at The Upshot.
New York Times/Siena College poll of likely voters in Alaska
And they are not good for President Trump.
We’ve talked a lot recently about how the polls have lagged behind the news. Events were moving so fast that by the time the polls had measured the state of public opinion after something like a debate, the news had already moved on to the next big thing, like the president’s coronavirus diagnosis.
Now the news has slowed a bit and the polls are finally caught up. Just before Thursday night’s dueling town halls, we were left with an unmistakable picture: The president trails badly, even though his negative news cycles, like the first debate or his coronavirus hospitalization, are well in the rearview mirror.
National polls
| Pollster | Margin | Diff. from ’16 result | |
|---|---|---|---|
| U.S. |
Morning Consult
Oct. 12-14, 15,499 L.V.
|
Biden +9
52-43
|
+7D |
| U.S. |
IBD/TIPP
Oct. 10-14, 864 L.V.
|
Biden +8
50-42
|
+6D |
| U.S. |
U.S.C. Dornsife
Oct. 1-14, 5,433 L.V.
|
Biden +13
54-41
|
+11D |
| U.S. |
Whitman (Dem. pollster)
Oct. 8-13, 1,103 L.V.
|
Biden +12
54-42
|
+10D |
| U.S. |
Marist College
Oct. 8-13, 896 L.V.
|
Biden +11
54-43
|
+9D |
| U.S. |
University of Massachusetts Lowell
Oct. 5-12, 819 L.V.
|
Biden +11
54-43
|
+9D |
| U.S. |
NBC News/
The Wall Street Journal Oct. 9-12, 1,000 R.V.
|
Biden +11
53-42
|
+9D |
| U.S. |
Edison Research
Sept. 25-Oct. 8, 1,378 R.V.
|
Biden +13
48-35
|
+11D |
Nothing good for Trump nationwide. We got a lot of national polls on Thursday, perhaps in part because some had been scheduled to be released ahead of the now-canceled second debate. Whatever the reason, they weren’t good for the president. Not at all.
Most obviously, the polls show Joe Biden ahead by more than 10 points — that’s about the same as our average. President Trump’s best result is, what, down by just eight points? That would have been an above-average result for Mr. Biden just a few weeks ago, and now it counts as good news for the president. It’s a stark reminder of how much the race has changed in just a short period.
Less obvious: There’s no sign that things are getting better for President Trump. There’s no hint of tightening. No way to credibly construe the numbers as much better for the president than the average, or better than the prior edition of these polls. You could note that the venerable NBC/WSJ poll shows Mr. Biden up 11 points, down from its previous survey after the first debate, when it had the lead at 14. But that previous result looked as if it was running pretty hot for Mr. Biden, and it’s a lot easier to interpret the new result as expected movement back toward the overall average, rather than as an actual shift toward Mr. Trump.
The trend line and the clock. If you’re a fan of the president, you had some reason to be cautiously optimistic that the polls might trend your way this week. It’s hard to sustain a 10- point lead in this polarized era, and the Amy Coney Barrett confirmation hearings had the potential to finally offer a positive news cycle.
Instead, we’ve gone yet another week without much evidence that the president’s standing has improved. Mr. Trump doesn’t have many opportunities left.
What happened in state polling. Again, I’d advise you to scan the far right column of the poll table, which shows how much a poll result is different from the 2016 election result. Today, you’ll see lots of +8D and +9D — consistent with Mr. Biden’s national lead.
| Pollster | Margin | Diff. from ’16 result | |
|---|---|---|---|
| Ariz. |
Monmouth University
Oct. 9-13, 502 L.V.
|
Biden +4
50-46
|
+8D |
| Ariz. |
OH Predictive Insights
Oct. 4-8, 608 L.V.
|
Biden +4
50-46
|
+8D |
| Ariz. |
Targoz Market Research
Sept. 23-Oct. 2, 1,045 L.V.
|
Trump +1
45-46
|
+3D |
| Iowa |
Data for Progress (Dem. pollster)
Oct. 8-11, 822 L.V.
|
Trump +1
47-48
|
+8D |
| Maine |
Pan Atlantic Research
Oct. 2-6, 600 L.V.
|
Biden +10
50-40
|
+7D |
| Mich. |
Civiqs
Oct. 8-11, 543 L.V.
|
Biden +9
52-43
|
+9D |
| N.C. |
Civiqs
Oct. 11-14, 1,211 L.V.
|
Biden +5
51-46
|
+9D |
| Ohio |
Civiqs
Oct. 8-11, 586 L.V.
|
Trump +3
47-50
|
+5D |
| Pa. |
Opinion Savvy/
InsiderAdvantage Oct. 12-13, 400 L.V.
|
Biden +3
46-43
|
+4D |
| Pa. |
Civiqs
Oct. 8-11, 600 L.V.
|
Biden +7
52-45
|
+8D |
| Wis. |
Trafalgar Group
Oct. 11-13, 1,043 L.V.
|
Biden +2
47-45
|
+3D |
| Wis. |
Civiqs
Oct. 8-11, 560 L.V.
|
Biden +8
53-45
|
+9D |
That said, the state polls were still a bit better for the president than his national numbers. That’s been something of a trend in the last few weeks, and it’s hard to say why. It could be something about the demographic makeup of the well-polled battleground states. It could reflect the robust campaigning in those states, including relentless television advertisements. Or it could just be because the pollsters are different.
Medium-blue Arizona. The polls in Arizona have been pretty split over the last month or so, with some showing a very competitive race, and others showing Mr. Biden far ahead. But Thursday, there was a rare consensus in the state from two pollsters who seemed to differ by a lot last month. OH Predictive Insights, which showed Mr. Biden up by 10 last month, found Mr. Biden up by four. A Monmouth poll also showed Mr. Biden up by four points (averaging their high- and low-turnout scenarios), after showing him up by two points last month. The big picture in the race remains essentially unchanged — a meaningful though not overwhelming Biden lead — but it is nice to see a little more of a consensus.
Friday, 1 p.m. Eastern: Alaska. Sadly, we never did get any responses in the arctic town of Utqiagvik, but we did get them as far west as the Yukon Delta and into the Aleutian Islands. (I assure you this poll will be on the internet at 1 p.m. or I’ll eat my hat.)
Next week: We conducted our first few interviews in Georgia last night, and we have more polls going into the field this evening. More soon.
State of the race: 19 days to go and the president is trailing badly.
President Trump and Senator Lindsey Graham have clear advantages in an increasingly competitive state.
My colleague Jonathan Martin has the story.
New York Times/Siena College poll of likely voters in South Carolina
Posts
No comments:
Post a Comment