Close polls don’t mean a close election. Interestingly, Nate Silver blasts the pollsters.
Competing Positions
- Atlas Intel was the most accurate pollster in 2020. There’s no way Trump is losing Iowa if he’s ahead in all these states.
- Let’s not become complacent
There’s room for overlap but statement 2 is more accurate.
Let’s rephrase statement 1 so that it is accurate. “Atlas Intel was the most accurate pollster in 2020. There’s no way Trump is losing Iowa if he wins all these states.”
Trump’s lead in Michigan is +2, Wisconsin +1, and Pennsylvania +2. That’s assuming you believe that set of averages. So yeah, Trump could lose them all.
Silver has Trump ahead in Pennsylvania +0.4 but Harris +1.1 in Michigan and +0.9 in Wisconsin.
I think the first set of numbers is more likely, but put +- numbers on whatever set you want and nearly any combination can occur.
A Shocking Iowa Poll Means Somebody Is Going to Be Wrong
Nate Silver comments A Shocking Iowa Poll Means Somebody Is Going to Be Wrong
Yesterday, I complained about how so many pollsters are “herding” by publishing results that are almost an exact tie in a way that is incredibly statistically improbable given the unavoidable sampling error from surveying a small number of voters. I also noted a handful of prominent exceptions — rouge pollsters like the New York Times/Siena College that practically exist in an entirely different universe and imply a much bigger political realignment.
Another such maverick is Ann Selzer of Selzer & Co. (Selzer and NYT/Siena are our two highest-rated pollsters.) As my former colleague Clare Malone wrote in 2016, Selzer — like NYT/Siena — has a long history of bucking the conventional wisdom and being right. In a world where most pollsters have a lot of egg on their faces, she has near-oracular status.
In September, her survey had Kamala Harris just 4 points behind Trump — considered an outlier at the time.
Her new poll? It shows the state trending even bluer, with Harris leading in Iowa 47-44.
Releasing this poll took an incredible amount of guts because — let me state this as carefully as I can — if you had to play the odds, this time Selzer will probably be wrong. Harris’s chances of winning Iowa nearly doubled in our model from 9 percent to 17 percent tonight, which isn’t nothing. Polymarket shows a similar trend, moving from 6 percent to 18 percent after the survey. But that still places Harris’s odds at around 5:1 against.
The poll has a reasonable sample size: 808 likely voters. Still, the margin of error for the difference separating the candidates in a poll of that size is ±6.6 points. That means in theory, in 95 out of 100 cases, the “real” number should be somewhere between Trump +3.4 and Harris +9.6 if Selzer had surveyed every single Iowa voter instead of just an 808-person sample.
To give us a little more perspective, there was also a second Iowa poll out tonight from Emerson College that showed Trump leading by 9 points, close to the margin from 2020. Emerson is a firm that does a lot of herding, so you ought to account for that — they virtually never publish a survey that defies the conventional wisdom. However, for what it’s worth, their margin of error runs from Trump +15.7 to Trump +2.5.
Iowa is far redder than the other Blue Wall states, though. In our 40,000 simulations tonight, in cases where Harris won Iowa, her average margin of victory or defeat in Wisconsin was +6.7 points, in Michigan was +5.9 points and in Pennsylvania was +3.8. If the Selzer poll is right, she’d be heavily favored in all of them, in other words. Conditional on winning Iowa, her chances in Wisconsin are 98 percent, for instance.
The most important update you should probably make from the Selzer poll is that Harris might overperform her polls in the Blue Wall — especially in Wisconsin, the most correlated with Iowa of the three.
However, if she loses Pennsylvania, Harris has to win both Iowa and some other state since that’s a net loss of 13 electoral votes, adding 6 from the Hawkeye State but losing 19 in the Keystone.
Silver Blasts the Polling Industry
If you’ve been reading this newsletter closely, you’ll know I’ve become increasingly concerned about the possibility of another systematic polling error — although I don’t know whether it will favor Trump, favor Harris, or produce a map that’s just really strange. A lot of things are hard to reconcile in the polls:
- Why likely voter models help Harris in national polls but hurt her in the swing states.
- Why some high-quality pollsters like Selzer and NYT/Siena, who are notorious for not herding, are so different than the others.
- Why demographic shifts are extremely pronounced in the crosstabs, especially Harris losing ground with Black and Latino voters, but produce relatively small swings in the toplines.
- Why so few polls in states like Wisconsin — less than 10 percent! — dare to show anything outside the range of Harris +2 to Trump +2, even though this is incredibly statistically unlikely.
- And why national polls have swung quite a bit toward Trump over the past several weeks but battleground state polls have moved only slightly.
Silver Cheerleading for Harris
Silver repeatedly praises NYT/Siena and Selzer as the top pollsters. He is definitely cheerleading for Harris.
Why is there no such mention of accuracy of AtlasIntel other than a footnote of them being “high quality”.
AtlasIntel
Nate Silver commented on Pollster Accuracy on X on March 21, 2021.
Finally, here’s how the most prolific pollsters fared in the general election. The best-performing pollster was AtlasIntel! Second was Trafalgar!
Yeah, they incorrectly had Trump winning a few states, but they were close on the margins, and that’s the better metric.
Not once this year has Silver noted that AtlasIntel was the most accurate pollster in 2020, a year in which pollsters were horrid nearly everywhere else.
Polling Error in 2020
AtlasIntel was off by an average of 2.01 percentage points.
The beloved NYT/Siena was off by an average of 5.26 percentage points, outside there stated margins of error.
YouGov was off by 5.27 PP, Monmouth University by 5.41 PP, WaPo by 6.30, CNN by 7.05 PP, and the blue ribbon for inaccuracy goes to Quinnipiac University, off 8.45 percentage points.
AtlasIntel Final Polls
Please consider the AtlatIntel Final Polls.
- Arizona: Trump 51.9 to Harris 45.1 Percent, Sample Size 967
- Georgia: Trump 48.6 to Harris 46.8 Percent, Sample Size 1,174
- Michigan: Trump 49.3 to Harris 47.8 Percent, Sample Size 1,198
- Nevada: Trump 51.4 to Harris 45.9 Percent Sample Size 782
- North Carolina: Trump 50.4 to Harris 46.8 Percent Sample Size 1,310
- Pennsylvania: Trump 49.6 to Harris 47.8 Percent Sample Size 2,049
- Wisconsin: Trump 49.3 to Harris 48.0 Percent Sample Size 728
The margins of error in those polls are +- 3.0 percentage point for Arizona, Georgia, Michigan, and North Carolina.
For Nevada and Wisconsin, the margins of error are +- 4.0 percentage points due to smaller sample size.
For Pennsylvania, the margin of error is +- 2.0 percentage points due to larger sample size.
All but Arizona are within the margin of error.
About AtlasIntel
- AtlasIntel was the most accurate pollster of the 2020 presidential election and of the 2022 midterms in the United States.
- In 2020, Atlas polls had an average error of only two points and anticipated election results within the margin of error in all swing states during the worst cycle for the US polling industry in four decades.
- AtlasIntel was the most accurate pollster of the 2022 presidential election in Brazil.
- AtlasIntel was the most accurate pollster of the last three election cycles in Argentina.
- AtlasIntel was the most accurate pollster of the 2022 Constitutional Plebiscite in Chile
- AtlasIntel polls anticipated the major swings of the 2022 presidential race in Colombia, during the primaries and the campaigning for the 1st round of voting and then the runoff.
- AtlasIntel was the only pollster to correctly anticipate the deadlock resulting from the 2023 legislative election in Spain
The Most Accurate Pollster in 2020 Has Blockbuster Polls for Trump
On October 20, I noted The Most Accurate Pollster in 2020 Has Blockbuster Polls for Trump
Accuracy in 2020 does not imply accuracy in 2024 but it does explain why many of my readers distrust Nate Silver, then at 538 and now the Silver Bulletin.
One reader asked “Why do you keep referring to Nate Silver?”
The key reason I refer to Silver is that he explains in detail what he is doing and why. Everyone is free to make their own adjustments and I do.
A Bit of Silver Bias
Nate’s bias is clearly showing with this constant praise of NYT and Selzer. Yeah, I get it. Selzer doesn’t herd. Clearly, AtlastIntel doesn’t either.
So, do you want to place faith in one poll or seven from the most accurate pollster in the US and likely the world?
Silver could have mentioned this, but didn’t. This does not make Silver’s methodology wrong, but he does have a bit of a biased slant in explaining what his polls say.
Close polls don’t mean a close election as determined by the electoral college.
We might easily see a blowout for someone, accompanied by screams of unfairness or cheating no matter who wins.
I don’t know who will win, and you don’t either. But I’ll take AtlasIntel over Selzer or NYT/Siena.
There is a reasonable chance the pollsters are close this time but uniformly off by 1-2 PP in a consistent direction. That could lead to a “very close blowout”.
Finally, and adding this as an afterthought, I just can’t shake my feeling the tipping state for Trump will be Wisconsin.
I just finished my “Handy Dandy Election Night Pollster Scorecard”
I will post this on Tuesday – Please pass to your fiends
I am hopeful that this AtlasIntel polling data is accurate. Thanks for posting!
Crossing my fingers they are dead wrong.
Arnold Friend.
An old fiend.
Yes, we will pass to our “fiends”!!
Sentient: “Anne Seltzer has a long history of being laughably wrong. She had Hillary way up (9?) in Iowa and Trump won it handily.”
Let me straighten the memory of Sentient who was to inconvenienced to look this up.
Selzer accurately predicted Trump’s 2020 win over now-President Joe Biden in Iowa. In a poll conducted in late October of that year, Selzer showed Trump with a 7-point lead over Biden (48 to 41 percent). Trump won Iowa days later by 8.2 points (54.2 to 45 percent).
Selzer also predicted that Trump would win his first presidential election in 2016 against former U.S. Secretary of State Hillary Clinton by 7 points.
https://www.newsweek.com/who-ann-selzer-iowa-poll-kamala-harris-1979294
If you are going to post bullshit that contradicts me, Might I suggest doing a bit of research first.
Fair enough. I stand corrected.
Fair enough here too – removed from spotlight
Her poll is a nonsensical outlier
Her explanation was insufficient: https://x.com/RedWave_Press/status/1853254106023944612
I stated that it was over immediately following the Fox “Baer Interview” with a 100% Chance Trump Wins! I stood by it afterwards on two more occasions, and I will again with a Final call now.
– 100% Trump Wins
– Called before the night is over by some “Major Podcast” (Rogan & Bongino perhaps) And “X” for certain!
– Riots by the Opposition well into the morning.
– Sanctuary Cities well into the next evening perhaps?
– Within 3-5 Days Things will have calmed down, or will be forced too, via curfews etc.
– Noticeable “Law & Order” difference around the Country within a week.
– A National Calming effect arrives by Christmas.
Kudos to Selzer for not “herding.” However, the fact that “Future of democracy” and “abortion” combine for ¾ of the main concerns on these Iowans’ minds makes me question the sample. It happens. Sometimes a pollster simply draws a crappy sample.
“…rouge pollsters like the New York Times/Siena College…”
Clearly, rouge pollsters are biased toward red states.
Lol
Kamala Harris emulsified with SNL, giggling all day. Under Trump the reps party is a working class party and minorities, not the delicate elite. Redneck women are tougher than men.
Kamala was drunk all day Monday people she acted like a giddy fool !
So, everything is so close and within the margin of error. I guess that means that if Trump ends up winning, the pollsters were more or less correct. And, if Harris ends up winning, it means that is was a rigged election!
We should know on Wednesday, …, or Thursday, or Friday, or next week, haha!
Re “ do you want to place faith in one poll or seven”
Faith is not what’s needed. Faith, like hope, is not a plan.
Patience is certainly needed, with a touch of stoic endurance.
Garbage people doing garbage things: https://www.tcpalm.com/story/news/local/martin-county/2024/11/02/stuart-police-charge-teen-with-punching-harris-supporter-70/76014623007/?utm_source=fark&utm_medium=website&utm_content=link&ICID=ref_fark
I’ve been trained to accept the absurd reasoning that there is no such thing as massive voter fraud and that the FBI investigates and prosecutes people for massive fraud playing the McDonald Monopoly game:
https://www.msn.com/en-us/money/companies/mcdonalds-monopoly-canceled-in-us-after-fbi-scandal-dubbed-it-danger-to-public-health/ar-AA1pZGNU
“polls are garbage, but let’s obsess over them anyways… REEEEEEE !!!!”
A “tie” is where democraps slide the extra 5k, 9k ballots in.
I remember some info about Selzer poll’s internals. She seems to think that a disproportionate chunk of the voters is college-educated, older, women living in the suburbs!
I think Selzer is retiring from the polling business and is aspiring for some kind of polling ANALyst job or whatever at DNCNN/MSDNC, and this is how she made sure she will be picked for that job.
That is what I thought also, like she is going to be a sidekick to Orange Lady (Psaki) or the reliable “expert guest” on MSNBC.
If Harris wins the presidency, but republicans take the senate and house, what policies will she be able to continue. Unchecked immigration, student loan bailouts.
80′ election was suppose to be a nail biter. Who knows but this feels similar to that situation. She obviously will win more than one state unlike Mondale but she has no gas or fumes. I have no idea what she’s running on. Aborted fetuses? No idea. We will find out together.
*1984
In case anyone missed it you can not pay your Bills with Bull Shit.
People are hurting and they will be voting their wallets not what the progressive campuses deem appropriate.
Does It Matter?
20 years ago I would’ve said the presentation of poll results could have an impact, in various ways, on the election results.
Now? I have no idea. I actually have no idea how anybody gets any of their information anymore, or whether there is some definable “campaign narrative.”
If 1% of Trump supporter are playing games with pollsters, then Harris’ campaign is in jeopardy.
Trump Media Outsourced Jobs to Mexico Even as Trump Pushes “America First”
Trump says he doesn’t mind someone shooting at journalists at rally
No, I think that was me.
Don’t feed the troll…
In Kansas, the Kansas Speak poll by the Docking Institute at Fort Hays State University found Trump winning the state by a meager 5-point margin, 48-43. Trump won the state in 2020 by 15. If the poll is right, it would mark a 10-point shift, similar to Iowa.
For reference, in 2020 the poll predicted a 14.4% Trump victory. He won the state by 14.6%.
Many of the polls have overcorrected for Trump and under for Harris. This is 1980 all over again. R voters don’t want to admit publicly, to anyone, that they have (or will) vote for Harris.
Lmaooooo
Correct.
“So, do you want to place faith in one poll or seven from the most accurate pollster in the US and likely the world?”
I want to place my faith in no mail in ballots & voter ID requirements.
Voting laws in half of the country have changed so much in the last 8 years that the doors for malfeasance have been swung wide ass open. One can only hope that we get lucky this time and “if there’s a silver bullet or two to be found” that we find it.
IF there is cheating going on, the ONLY it will be uncovered is for someone like Trump or DeSantis to be elected.
May God bless America or what’s left of it.
“It’s hard to make predictions. Especially about the future.”
— Yogi Berra
I miss Yogi.
He was one-of-a-kind.
If u drink yogi tea u will know who will win
Like multi sigma misses on econ numbers. Sometimes looking up to see which way the wind is blowing works the best. The country is not as dumb as the Dems would hope.
AtlasIntel’s cross tabs show 25% of blacks voting for Trump. Does anyone really believe that? They may have gotten lucky back in 2020 which was a bit of a fluke, but they will be way off this election.
Possible – I have been discussing this all year
Not buying that at all. Harris will get north of 80% of black votes, with less than 15% going for Trump. And the story on Tuesday night will be Republican women crossing over for Harris due to abortion.
If you look at the cross table data for various minorities across the swing states you’ll see that the percentage data varies by a huge amount. So I really don’t know how to evaluate the data.
Based on prior election data for Mr Trump that figure is not unreasonable and given the change in the country over the past four years I would also guess that it might actually be understated.
IMO the polls are setting up a situation so that unless it is a landslide for Mr Trump it will give the democrats and excuse to challenge the election.
After all they are the biggest hypocrites in the world.
Bush/Gore 2000 was settled peacefully. In Nov 2008 Obama won. GW Bush transferred power peacefully. In Nov 2016 Trump won, but Obama, Nancy and the Clintons didn’t accept the results. In Nov 2020 Biden won, but Trump didn’t accept the results. Neither the dems nor the reps accept results when they lose. The losers terrorised the winners. If that cont our democracy is gone. So the loser matters most.
When Bush was inaugurated in Clinton stood on Fifth Ave and 58th st trying to get people’s attention. After GW Bush landed on an aircraft carrier the bitch was furious. The dems community entities lent money to unworthy borrowers. The dems hyenas attacked the WMD cause after 3,000 American lost their lives in WTC. Saddam killed thousands Kurds and Iranian soldiers during Iran/Iraq war with chemical weapons. GW didn’t care.The Dow and the housing market were rising and Ben Bernanke’s Regulatory Relief Act of 2006 protected him.
Hi, Mish, even thought I do not vote, I get a kick out of your coverage and keep reading your Poll summaries. Thus, I am saying that I will also watch intently and hope for the best outcome for America…based on the last 4 years.
I also pass them along to EVERYONE that we know who DO vote and I have gotten some very angry emails from both sides. I have to remind them that your work is inherently unbiased, at least in my mind it is and that seems to work. Some people get angry over bad weather reports in my clan. They cannot help their biases.
Thanks – Much appreciated
I just finished my “Handy Dandy Election Night Pollster Scorecard”
I will post this on Tuesday – Please pass to your fiends
AtlasIntel just might be herding (as Silver accuses). There should be a certain amount of sampling variation that is not there. This is similar to Gregor Mendel’s data bean counts — it’s been thought his data might have been fabricated because experiment fit theory unusually closely.
You don’t understand what heading means.
A symptom of herding is the average of what everyone else is saying
One thing is for certain.
Polls reporting a close race, means more people will turn out to vote!
Which means everyone is drinking the Kool aide. And when, the Victor is announced. Phase 3 of the agenda will commence.
A close race sells more of everything.
It’s not over til it’s over.
A probable scenario : neither side will accept the result.
Only one side has a history of actually doing so.
In 2016 Trump said he would accept the results “if he won”.
And he still refuses to accept his 2020 loss.
Just one side has an issue with fact checking.
Fact check: Both parties have a long history of fighting over any close election result, at any level of government, and often with strong justification.
Still recently, Hillary called Trump an illegitimate president.
Will US trading halts on 5 November extend its 5 Nov 6.5 trading hours to 6 to 8 November?
Its over… get out your hankies:
https://www.desmoinesregister.com/story/news/politics/iowa-poll/2024/11/02/iowa-poll-kamala-harris-leads-donald-trump-2024-presidential-race/75354033007/
Accuracy in 2020 doesn’t imply accuracy in 2024.
Correct – I stated that many times
As I said before the smart media owners and their MBA lackeys know the best way to keep the publics’ eyeballs on the screen/paper (watching advertising) is to keep pushing that this is a 50/50 race, whether it is or not. Theres no downside advocating a 50/50. All that matters is not what is ethical, moral or the right thing to do. All that matters is whats legal.
Does it being legal or not even matter when laws are not enforced or enforced selectively? When was the last time a mainstream tv or newspaper editor was sent to prison for breaking laws such as election interference?
I’m expecting a red wave but even that would still mean 40% plus of voters voted for someone dumber than Biden and her tampon sidekick.
Your red wave was supposed to come in 2020. Instead the party got its doors blown off and nearly lost the House. Oh, and they got Biden too.
Excuse me … 2022 elections were supposed to be red wave. Big loss.
This year: Fail Hard 2: Fail Harder
Turns out the majority of Americans doesn’t want to join the angry, deluded weirdos in their quest to make a moron king.
Thanks to the stock market my mom is richer than she’s ever been. And, a gallon of gasoline is the same price it was 10 years ago (chicagogasprices.com). There is inflation, but not in the usual places.
Scott you need to be 18 to post here
I don’t often agree with you but I do on this. Not only does a close race help, but media & marketing corps also benefit if the stakes appear to be as high as possible, even if it means soiling all their major brands with rhetorical nonsense.
MIke, here in vote by mail Oregon CNN has the state only “leaning” blue, which shocks me. The two biggest Dem hive mind counties, Multnomah and Clackamas, are running way behind the other red/purple counties in vote percentages and turnout. Perhaps even the Progs don’t like their neighborhoods being burned and infested with drug-addicted bumdom. Or, it could be the Progs really don’t think Kamala is prog enough and have little enthusiasm to vote. There may be a last minute swell of Democrat votes but that completely goes against their voting pattern. You being a financial guy, past performance is no indicator…well, you know.
Thanks for that Info. It provides a possible explanation for some truly bizarro numbers on my “Handy Dandy Election Night Pollster Scorecard”
I will post this on Tuesday
If Multnomah goes red or even pink I will have lived long enough to have seen it all. The world would surely end if Eugene also went that way. I would definitely look for the coming of the Messiah.
The women voters are pissed as I stated months ago. Not sure why ppl are doing baselines from 2020 when we had c19. Can’t ignore the outpeforming dems of 2022 thanks to Roe
Pissed about men in their locker rooms and stealing their awards in sports.
Agree
Feminists generally don’t care about the family…why would they care about perv dudes with the girls? It’s been an anti-family agenda from the start marketed under the guise of equality.
The radical anti-family types are bleeding support from the mainstream feminists who are pissed as hell about being called “birthing people” and told their daughters have to lose to “trans” Cats in sports …
I’m pissed about the economy and all the illegals.
… as you’ve been told to be, over and over and over, until you think you arrived at this conclusion yourself, and it just happens that a bunch of other weirdos have reached the same conclusion.
Stick a fork in ya… yer done.
An experienced pollster wrote here that 99% don’t answer the phone. Those who do are mentally retarded Dimocrats or elderly Republicans.
That probably maps onto most of the actual voters too.
Nobody knows. You might as well be reading entrails at this point. Polling is broken forever. We’ll know in a couple days.
Buh but how am I gonna throw a tantrum when my candidate loses if I can’t point to the poles?
Peanut did not die in vain.
Anne Seltzer has a long history of being laughably wrong. She had Hillary way up (9?) in Iowa and Trump won it handily.
You need a memory check and a fact check.