Let’s discuss some new polls, pollster accuracy, polling bias, the economy, and a reader question on Nate Silver.
Silver Bulletin Comments on Three New Trump-Favorable Polls
🕒 Last update: 1:45 p.m., Sunday, October 20. The data continues to be pretty negative for Kamala Harris. There are now three recent high-quality national polls that show Donald Trump leading — a difficult circumstance for Harris, given Democrats’ Electoral College disadvantage — and her edge in our national polling average is down to 1.7 points. National polls don’t influence the model that much, and the race remains basically a toss-up, but it’s not hard to think of reasons that Trump could win.
1: Fox News Poll
Fox News reports Trump ahead of Harris by 2 points nationally
Former President Trump is ahead of Vice President Kamala Harris in the presidential contest 50%-48%, according to a new Fox News national survey. That’s a reversal from last month, when Harris had a narrow advantage.
Harris, however, is ahead by 6 points among voters from the seven key battleground states (within the margin of error for that subsample), and the candidates are tied at 49% each among voters in close counties (where the Biden-Trump 2020 margin was less than 10 points). Trump’s advantage comes from a larger share in counties he won by more than 10 points in 2020 (64%-35%) than Harris has in counties Biden won by more than 10 points (58%-39%).
That raises the question of whether the Democrat could win the Electoral College while losing the national popular vote. In 2000 and 2016, it was the GOP candidate who lost the popular vote but won the Electoral College.
That’s certainly an interesting, if not outright shocking set of polls that has Trump ahead nationally but trailing in battlegrounds. Seriously?!
2: AtlasIntel National Poll
📊 ATLAS POLL – U.S. PRESIDENTIAL ELECTIONS
🔴 Trump: 50.7% (-0.2)
🔵 Harris: 47.6% (+0.3)
⚪ Others/Don’t know: 1.7% (-0.2)
3: TIPP Tracking Poll Day 7
In the third of three high-quality polls, TIPP Insights reports Trump Surges Past Harris, Seizing 2-Point Lead
Powerful Momentum Shift Reverses the Race
Former President Donald Trump overtakes Vice President Kamala Harris and now enjoys a two-point lead, 49% to 47%, in the TIPP tracking poll. The data from the seventh release of the 23-day TIPP tracking poll shows the underlying change in the momentum. Recognized by The Washington Post in 2020 as the most accurate national presidential poll, TIPP continues to bring you the latest insights on the race.
The WaPo report did not include an evaluation of AtlasIntel. People are still debating what went wrong in 2020.
Here’s a free link to the Washington Post 2020 Accuracy Report.
2020 Accuracy
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.
Note: I no longer embed Tweets because X embeds are broken and randomly display on WordPress. This appears to be an X issue, not a WordPress issue.
Polling Error in 2020
Accuracy in 2020 Comments
The Hungarian Conservative notes: “The most accurate pollster nationally in the United States in the last two election cycles has been Atlas Intel. In 2020 the public opinion research firm overestimated Biden by 0.2 points. In 2022 they overestimated the GOP by 0.3 points in the generic Congressional ballot. Impressively, in both cycles, they got the final results within a few tenths of a point.“
AtlasIntel Comments: “AtlasIntel had the best performance across all pollsters of the 2020 US Presidential Election with an average error of 2.01p.p. Our final national-level poll showing Biden with a 4.7 p.p. advantage is likely to be the single most precise estimate of the US popular vote. AtlasIntel also conducted the most precise polls across all pollsters for the states of Michigan and North Carolina. In all states polled by AtlasIntel, results fell within the margin of error of our estimates. The AtlasIntel average error was not only lower than that of other pollsters, but also lower than that of polling averages and model-based polling aggregators.”
Reader Comments
Accuracy in 2020 does not imply accuracy in 2022 but it does explain why many of my readers distrust Nate Silver, then 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.
Silver has Trafalgar with a 2.7 percent Republican bias. What if Trafalgar is correct?
It makers sense to me to rate pollster bias and weight polls by sample size.
However there is certainly a question as to whether Silver has Pollster Bias all wrong.
Might I suggest accuracy is the best way to determine if a pollster is “biased”? After all Trafalgar was the second most accurate in 2020.
Pennsylvania Polls
AtlasIntel Influence
Note the influence Silver gave AtlasIntel. This makes sense to me. The reason, or at least one of them, is sample size.
One of my readers commented that Silver should not weight the polls but average them. I disagree. In general, the larger the sample, the more accurate the poll.
AtlasIntel had a huge sample size of 2,048 likely voters. The NYT/Siena poll only had 857 likely voters.
Morning Consult had 1,395 likely voters. Silver rates Morning Consult as having a Democrat bias of 1.2 percentage points. That means to Silver that Trump is ahead in Pennsylvania.
Returning to TIPP Insights
A growing sense of economic dissatisfaction looms large. The TIPP Tracking Poll shows that 53% of Americans feel worse off now compared to their situation pre-COVID. Democrats are the most optimistic, with 51% saying they’re better off, but 68% of Republicans report being worse off. Independents largely mirror the national mood, with 55% feeling the pain. This dissatisfaction is a headwind for Harris, which Trump could exploit.
Growing Economic Dissatisfaction
Does “growing economic dissatisfaction” sound familiar? It should. It has been my repeated theme since February, weighted to renters (young voters and blacks).
Nate Silver Seriously Lags
On October 12, I asked How Much of a Political Lagging Indicator is Nate Silver?
The answer, as you can see, is a lot.
Trump’s Polling Momentum Is Real
On October 18, I commented Trump’s Polling Momentum Is Real With Under Three Weeks to Go
Let’s discuss momentum. Nate Silver mostly dismisses it, but I don’t. And the election is in 17 days.
Silver had momentum totally wrong. Momentum was there, and obvious. He chose not to see it.
Why the Momentum Now?
For 67 percent of the nation (the asset holders and the home owners), the economy (the stock market) has been good to very good, offsetting inflation.
For the renters and struggling homeowners fed up with property insurance, food, and other inflation, the economy borders on miserable.
Who is that? Young voters, blacks, and those without a college education.
That is what I said in my post on momentum. The undecided voters are breaking as I suggested in February.
Renters (Young Voters and Blacks) Will Decide the Election
I made that comment in February.
I repeated that message in April with People Who Rent Will Decide the 2024 Presidential Election
I repeated the message a second time on June 19 in Why Angry Renters Will Decide the Election, Take II
Economists say wages are now rising faster than the CPI. That’s not true for those who rent or wish to buy a house.
Trump fell way behind after a disastrous debate with Harris. That momentum ended a month ago.
People are again asking “Am I better off than 4 years ago?”
If you own stocks or refinanced your house at 3.0 percent, you likely are better off. And that’s a big majority. But that’s not who will decide the election. And it is a big part of why Harris is struggling.
Trump, not Harris is the candidate of change.
The economy is not as good as economists claim. People are suffering, especially Black and young voters, and those traditional Democrat votes are breaking hard for Trump.
That is what the Fox News, TIPP Insights, and AtlasIntel polls all suggest. And it’s what I have been saying all year.
I find your accuracy and table a useful tool when analyzing poll results on realclearpolitics.com. What amuses me the most is that you could have a similar poll with the same MOE 1 day apart and the difference in results match your statistical bias. That tells me that reality is probably smack dab between AtlasIntel and Trafalgar. Also, Trafalgar is bashed in the media (because they try to break through the fear of trump supporters responding) but I have always found them to be pretty accurate.
There are a number of things I’d argue with here, but most egregious is “In general, the larger the sample, the more accurate the poll.”
Sorry, but no! The math here is pretty clear: you get diminishing returns, and a sample of 2000 doesn’t give you much better margin of error than a sample of 1000.
Much much much more important are things like: method of contacting voters, phrasing of questions, weighting demographics properly, etc. Those have a huge impact on the accuracy of the poll. Generally 800-1000 is good enough to get a MOE of 3%, increase that to 2000 and maybe you shave it down to 2.5% MOE. Not that significant.
So, you want to base “pollster bias” 100% on the outcome of one election. That’s 1 data point. It’s not exactly a trustworthy sample size, as that result could have (And there is a very good case it was indeed) been a freak outlier as opposed to the norm. Seriously? “Well, they got it spot on at a time when polling had a systemwide failure, so we’re going to assume they have no bias” is a ridiculously unscientific take.
Apparently, over 100% of Michigan eligible voters are registered to vote, according to an article on Zerohedge. What is much more troubling is that historically, around 70% of eligible voters actually register to vote. I’m fairly certain it’s not a sudden gush of civic duty that has recently pushed the number up from 70% to over 100%.
The polls used to be very close to the election results before the voter ID verification requirement was dropped. So math works on polling data with flying colors until voter ID is dropped? Then all of a sudden it magically doesn’t? And there is no fraud? Just saying, math isn’t the thing that changed.
Poll stuffing in the last few weeks is well documented as is the appearance of pollsters who did nothing much before the past month.
This is 1980 but Trump is Carter.
I will say the stuff on the economy is spot on. The historic Democratic base of younger city dwelling younger folks are getting clobbered by the Democratic controlled economy with inflation and the flat out lack of lower income working class housing especially in the large blue states. It’s simple, people mostly vote their pocketbooks..
Inflation is the soul-killer.
I’ve read the comments here and, as usual, everyone has an opinion. But, one thing I haven’t seen is anyone really dissing the polls, as they forecast a red wave in the midterms and it didn’t appear. Conventional wisdom is that Trumpers don’t respond to polls, thus his underpolling historically. Might the same happen here, but, in the opposite direction? For example a lot of women, who don’t really poll, but come out to vote for Harris because they just cant stand what a blowhard Trump is? Only time will tell.
The polls may not be in error, but can still be misleading.
If one side is significantly more motivated to vote, that alone will decide the election.
The tilting of the independents may only be a secondary factor.
And of course the issues of cheating…
Polls are unreliable. They are polling less than 5,000 people and are biased. I think Trump will win with over 300 electoral college votes.
“The Most Accurate Pollster in 2020” is like saying the most profitable stock picker in 2020, just because they were right or lucky then does not mean they are right now or in the future. I see no more justification for any emotional attachment to anything these pollsters put forth than one would attach to the stock picker of four years ago as we cannot know who is more accurate or lucky for this cycle until the election is over.
Absolutely bonkers that folks prefer the unserious would be strongman who mocks our nation.
I think people fully understand Harris is an empty vessel and will not risk four years more of that.
It’s a choice between a D%ck and a C&nt. It’s all just Sh$t either way.
I’ve seen a few Harris/Walz ads in the NYC area where I live.
Very unusual as dem prez candidates dont usually see the need to commit resources to NYS, especially in the downstate Blue area.
Funny you should mention that.
Where I am on North Shore of Suffolk County noticed that after Bret Baier interview, yard signs for Harris Walz started disappearing and leaving up just down ticket candidates.
Interesting. I travel around the Westchester area a lot and the Harris/Walz Joy signs, In this House signs are highly correlated with the upper income areas like Bedford, Chappaqua, Mt. Kisco etc
My theory is that the pollsters dont personally know many Working Class ppl and that this is a blind spot for them.
Its the same is Upper Income County, NJ. Bunch of rich morons supporting feel good candidiates. Nobody below the average income and supporting Trump responding to polls. Pollsters get the overpaid overwealthy work at homes and stay at homes. Easy to juice your poll if you ask the right people.
As mentioned by Ben Hunt, polls are nothing but turning politics into football. None of matters, and none of it portends anything until the final whistle blows.
I know the area . I grew up on the poor side of westchester but I know that area very well . And isn’t that additional proof the the Democratic Party is the part of the wealthy??? I only see Hollywood idiots and millionaire rock stars and people from westchester voting for democrats
If it’s close, I think Harris will win. Harris will get the mail-in ballots she needs, exactly where she needs them.
No postage necessary!
No voter necessary!
For Trump the greatest danger seems to be he will open his mouth before election day and utter something impossibly stupid. If he could hold his breath for two weeks . . .
He utters something incredebly moronic daily. That said, the right wing news ecosystem is doing a great job of not reporting any of it. Considering Fox News is the only news outlet many independents watch, he’s got good cover there. More helpful still is that Elon is doing catch an kill for Trump on X/Twitter now (like the Inquierer used to), so, unluess Trump decides to do another debate, there’s almost 0 chance his insanity will be remotely exposed to the general electorate again. That doesn’t mean he’ll win, but he’s certainly avoiding the catastrophic campaign failure his daily behavior would have caused any other candidate not named Trump – just calling the left wing “the enemy within” and saying he’ll send the military after them would have destroyed the campaign of any other candidate ever.
My gut tells me Trump will win simply because Harris is not remotely qualified to be president.
“My gut tells me Trump will win because (Proceeds to include copy pasted Fox News talking point)” – Cool. Good thing your gut is irrelevant.
I wonder how many people here talk politics at the dinner table and who look for every chance to voice their opinions on politics? It’s an informal poll.
I forbid political talk at my dinner table, and I never engage in it when its brought up at someone else’s. For most, politics has turned into a religion and religion is best left as acknowledged but unspoken.
Everyone wants to talk about themselves. When people open their mouths, there’s no need to debate or counter argue. Simply ask enough questions and their opinions fall apart. It doesn’t matter what you think. You won’t convince them of anything. But when they see an idiot in the mirror, they start to change their tune.
Here is a Poll question that should be asked.
Do you want to live in a State that is run like California or run like Texas?
Do you want to live in a State run like New York and Illinois or run like Florida?
Extrapolate that to the National level.
In NYS there is a proposition on Ballot that will turn NYS into a sanctuary State.
Just what New York needs.
My prediction: the winner of Pennsylvania will win the election.
My prediction: Trump will win Pennsylvania comfortably, but then magically ‘lose’ while we sleep.
It’s very possible.
I worked a poll once in college for extra credit. Something like 99% of people hung up. The ones that answered were either elderly or idiots.
The polls get it ass-backwards:
It isn’t a vote for someone as much as it is…
a vote against someone else.
Dislike for Hillary translated to a win for Trump.
Dislike for Trump translated to a win for Biden.
Ask yourself who is disliked more between the two now.
What the Amish Can Teach America About Economic Mobility
One of my favorite comments from the article:
“If we’ve got plenty of jobs, I just go to my competitor and I give him a couple jobs,” said Miller, who is Amish. “I’m here so my employees can make a good, honest living…. I’m not here to collect all the wealth I can absolutely collect.”
https://archive.is/qM3oi
I will stick by my gut feeling, after what my Eyes saw and Ears heard, watching that Fox Interview. I truly thought, as I turned it off, Wow! She just Lost the Election! I see only more reason to believe that is still true today…
The interviewer should have started by telling her that she has an hour of air time, normally worth $xx,xxx,xxx if it was paid, ready, set, go! and then let her take it all away from there.
I do not think She is capable of speaking for an hour? A repeated “Word Salad” over and over again, would be too much. She could possibly have an aneurism or something. Not a good idea health wise.
My officemates are able to have open and respectful conversations about the election. We came up with this idea about polling bias: Republicans / Conservatives are less likely to participate in polls because they don’t want their opinions made in polls, but in VOTES. This is what is meant by “Silent Majority”. It explains why Trump beat Hillary in 2016. Contrast that behavior with the Democrats / Liberals who are vociferous about every topic, so much so that they will obstruct your commute to hold a protest.
I work at a public school where my coworkers frequently demonize conservatives.
My experience living in liberal areas (CA & MA) is that the conservatives are more outspoken about their views in public, but when I lived in conservative areas, the liberals were more outspoken in public and workplaces, etc.
Liberals and Lib institutions like the media have become so dogmatic and intolerant that ppl are just opting out of dealing with them.
They are like State Run media in the old USSR: ppl dont believe them and dont want to talk to them.
Mish, note another polling company that had been among the most accurate in 2020: Rasmussen.
Many of the most solid Bush republicans are now solidly behind Harris. The best of the now desolate dem party are now behind Trump. I do not find this surprising.
I hope that his second term is successful. I hope that folks can work together to keep this country great.
Actually, I’m sure they will.
Yes, both blue and red warmongers are backing Harris as the greater of two evils.
PROTECTTHEVOTE daught kom. This website seems to be dedicated to learning and teaching how to prevent election theft and fraud.
It has information for every state on how it’s votes are counted, and many links that show how YOU can get involved in making sure that ballots are counted fairly wherever you live.
Citizen involvement is crucial to having fair elections.
There has been a dearth of public interest in this priceless duty.
This is why so many are dissatisfied with the process today.
Amazingly, there is LOTS of room for folks to participate.
Most districts have been pitifully understaffed.
24 hour scrutiny by many is vital. Especially in swing states.
There are almost no obstacles to keep you from being a part of and a witness to our precious elections.
PLEASE check it out, and get involved if you can!
YOU are wanted and needed.
As the sheer horror of this installed DNC ticket sinks in deeper and deeper with just under 3 weeks to go, I see more and more folks adjusting their druthers and hoping that Trump’s second term and cabinet selections will be wiser and stronger.
DJT has been ahead by 5pts. minimum since the Afghanistan debacle.
DJT has been ahead by 10pts. minimum since the truth about Hunter Biden & Ukraine became undeniable.
Anyone telling you otherwise is selling something… like clicks.
The polls, based on scientific mathematical methodologies, have a strong bias in reality but they fail to take in account people’s emotions. In this case, the emotion is to hide from the stigma against voting for Trump.
The “social rule” that has been broadcast for years is that Trump is a bad actor. I like Trump, but it’s hard to avoid the stigma that has been cast on him. When a pollster, a stranger, asks you on the phone who you’re going to vote for, and it’s Trump, you may be inclined not to tell him, or to say you’re undecided. This was shown during the 2016 elections when turnout for Trump was better than the polls indicated. And it is even more the case today when we have had almost a decade of unceasing trumpeting against Trump from the media bullhorn and from the corners of “good, polite” society that set the social rules. To read more of my writing, go to: dark.sport.blog … have a fine day.
its all a hugger mugger
I will check all this out Nov 5th
Since 1866, the London wagering market has called every election correctly but two: Truman beat Dewey in 1948, and Trump beat Hillary in 2016, and where she was up 9 points in every major poll on this date, 2016. London currently has Trump at 59% odds vs, Harris at 39%. No educated person pays any attention to polls whatsoever.
What does it mean to be the most accurate pollster in the 2020 election? If you subscribe to the theory that the 2020 election was stolen by democrats using an unprecedented amount of cheating, it means that the most accurate pollster was actually among the worst.
Anyone who says 2020 was stolen has TDS type II.
Furthermore your math skills are lacking or your reading ability is.
Most pollsters had Biden up by 5-6 points. AtlasIntel was Biden +2. In other words they had more votes for Trump than anyone else.
So, care to rethink your math?
Mish, it wasn’t “stolen” in the traditional sense. But there is no doubt it was the most heavily “influenced” election in U.S. history, by outside entities such as the blatantly complicit traditional and social media and even the FBI. You’ll be happy to know virtually every post-election poll agreed with this thesis.
It was the saddest moment in History to Witness THE BIGGEST THIEF EVER (with his grimy hands getting the loot out of Ukraine and everyone saw it – – and I know people there in Political Circles who left Ukraine in 2019 due to the Greed of PREZ Z).
yeah… suppressing the truth about Hunter’s laptop was a figment of our imagination… got it.
Suppressing the laptop was real and I don’t recall ever saying otherwise. I do recall saying the media and the DNC flat out lied.
No reason for the insults. What I meant was the list of pollsters in your post are among the worst. I’ll take responsibility for the poor wording.
Democracy Institute, the national pollster who correctly predicted BrExit and the 2016 presidential election, had Trump up by almost 2 points in Oct of 2020. So they were the most accurate, not AtlasIntel. There were many other pollsters in the same ball park, but they were not national.
Nate Silver is misleading at best. He’s a political hack, as his performance during the 2016 election shows.
You may want to rethink your sources.
No one was more accurate than AtlasIntel unless you believe WaPo. Then another one got it nearly on the nose. How can closest be the worst?
Anyway the claim the election was the “stolen” is what set me off, not the math. The combo was too much to take at that moment.
Was the election fair? No – none of them are. Was it “stolen”, no. Besides, if Trump did not campaign like a moron he would have won.
He appears to win this time despite doing the same thing. The difference is the economy and people see through Harris’ lies as being the candidate for change.
The incumbent for change. What a hoot!
Trump and Hillary have this in common: They both blame others and take no credit for their own mistakes. Neither apologizes. I will.
Apologies for the nastiness.
No worries! I’m guilty of that too.
The closest can be the worst if the accuracy of their predictions is a result of election-rigging and chance rather than a fair election and good polling. Do note that I said “IF you believe in the theory that the 2020 election was stolen…”
Based on the negative votes that I expected, a lot of people have TDS and don’t understand simple math.
How many Covid shots did you take Mish?
Yes it was stolen and what a gift it is. We get Trump back after enduring the worst Prez/VP in history. LFG!!!
The RCP final average was Biden +7.2% and Biden won by 4.5%. Biden was also up just over 10% on average just a few weeks before election day.
https://www.realclearpolling.com/polls/president/general/2020/trump-vs-biden