A look into the final polling numbers down the stretch.
I’m Isaac Saul, and this is Tangle: an independent, nonpartisan, subscriber-supported politics newsletter that summarizes the best arguments from across the political spectrum on the news of the day — then “my take.”
Are you new here? Get free emails to your inbox daily. Would you rather listen? You can find our podcast here.
Today's read: 15 minutes.
A massive thank you.
Last week, we set a goal to drive 2,000 new subscriptions before Election Day. Over the weekend, “This American Life” did a radio piece about our work, which nearly crashed our website and helped us shatter our goal. So: Welcome to our many thousands of new readers, and thank you to all of you who helped us hit our goal heading into Election Day. For those of you who prefer listening, heads up: We also have a podcast which you can find here, and a premium ad-free podcast here.
Election interference?
Over the last few days, allegations have been flying about voter suppression and election interference in Bucks County, Pennsylvania. I’m a Philadelphia-based reporter who grew up in Bucks County, and so I got in my car, drove out to Doylestown (where the allegations started), and sat down with the chair of the Bucks County Republican Committee. Her perspective on the whole thing might surprise you. Here’s our interview:
Quick hits.
- Former President Donald Trump is holding rallies in North Carolina, Pennsylvania, and Michigan on the final day before the election. Vice President Kamala Harris is scheduled to make four appearances in Pennsylvania. (The latest)
- The unemployment rate was 4.1% in October, unchanged from the previous month, according to the latest jobs report from the Bureau of Labor Statistics. Employment increased in health care and government while decreasing in manufacturing due to worker strikes. (The report)
- Federal Communications Commission Commissioner Brendan Carr suggested that Vice President Harris’s appearance on “Saturday Night Live” over the weekend violated the agency’s “equal time” rule, which requires radio and television broadcast stations to offer comparable air time for competing political candidates. (The claim)
- The Pentagon announced the U.S. is sending additional troops to the Middle East in response to the fighting between Israel and Hezbollah in Lebanon. (The deployment) Separately, Israel said it captured a senior Hezbollah operative in an amphibious special-forces raid in northern Lebanon. (The raid)
- A federal jury convicted former Louisville police detective Brett Hankison of using excessive force in the 2020 raid that resulted in Breonna Taylor’s death, the first conviction of any Louisville officer involved in the raid. (The conviction)
Today's topic.
The final polls of 2024. Over the weekend, several respected pollsters released their final surveys of the 2024 election, offering a snapshot of the presidential race in its last days. Vice President Kamala Harris and former President Donald Trump have been locked in a close contest since Harris replaced President Joe Biden as the Democratic nominee, and the latest polls show the candidates remain neck and neck on the eve of Election Day.
A Des Moines Register/Mediacom Iowa poll conducted by Ann Selzer offered the most surprising result from the weekend. Selzer’s poll showed Harris leading Trump 47% to 44% among likely voters in Iowa, a shocking finding for a state that Trump won by 8% in 2020 and 10% in 2016. Selzer’s September poll had Trump ahead of Harris by four points, while her June poll showed Trump leading Biden by 18 points. Selzer said that women — particularly those who are older or politically independent — were responsible for the late shift towards Harris.
Although Iowa is not considered a swing state, Selzer is one of the most respected pollsters in the industry with a reputation for accurately forecasting election results missed by other polling outfits. In 2020, her final Iowa poll showed Trump leading Biden by seven points, a significantly larger margin than the FiveThirtyEight polling average for the state. Her final 2016 poll showed Trump leading Hillary Clinton by seven points, also larger than the final polling average. Since 2012, Selzer has a strong record of forecasting not only the eventual winner of the state’s presidential, senate, and gubernatorial races but also the winner’s margin of victory within one to three points (though she has also been off the mark at several times in her career).
Trump is still considered a heavy favorite to win Iowa, with other recent polls showing him with a commanding lead in the state. However, if Selzer’s forecast is accurate in predicting a closer-than-expected result, it could augur a stronger performance for Harris in other Midwestern swing states than the current numbers show. Iowa’s demographic profile is similar to Wisconsin and Michigan, and Selzer’s Iowa poll has been predictive of results in North Dakota, South Dakota, Michigan, and Wisconsin in the past five cycles.
On Saturday, the Trump campaign released a memo calling Selzer’s poll “a clear outlier,” pointing instead to a new Emerson College poll showing Trump ahead by 10 points. A Harris campaign official told POLITICO that the Des Moines Register poll’s result mirrors the campaign’s internal research that shows her gaining with women and seniors.
Meanwhile, the final New York Times/Siena College poll shows Trump and Harris deadlocked in most swing states. The poll found Harris leading by three points in Nevada, two points in North Carolina and Wisconsin, and one point in Georgia; meanwhile, Trump is up by four points in Arizona and the candidates are tied in Pennsylvania and Michigan.
Conversely, a new survey of likely voters from AtlasIntel found Trump leading Harris in every swing state. AtlasIntel was the best-performing polling outfit in the 2020 presidential race, having predicted every swing state’s results within the margin of error.
Today, we’ll share perspectives from right and left about the latest polls. Then, my take.
What the right is saying.
- The right feels confident in Trump’s standing, suggesting his level of support isn’t fully captured in the polls.
- Some consider the election a toss up but say polling results like Selzer’s make no sense.
- Others say the polls hold little value in predicting the outcome of the race.
In The Washington Examiner, Elizabeth Stauffer asked “is the presidential race actually as close as current polls suggest?”
“It’s appropriate to describe the current presidential race as neck and neck. Right? If polls are the only indicator you’re looking at, that might be a fair statement. However, there are other ways to gauge where the candidates stand with the electorate,” Stauffer said. “Trump has momentum in this race, and it has come at the best time possible. He is riding high after a string of public relations wins. Trump’s brief stint as a fry cook at a Pennsylvania McDonald’s was a stroke of genius. His three-hour interview with top podcast host Joe Rogan was viewed by tens of millions of voters. He made liberal heads explode with his tour de force at Madison Square Garden last Sunday.”
“Conversely, Harris has stumbled repeatedly. When her campaign’s media-created momentum stalled in early October, she embarked upon a media blitz. Her performances ranged from lackluster to disastrous,” Stauffer wrote. Further, “the most important economic indicator of all, the right track/wrong track poll, favors a Trump victory. According to CNN senior data reporter Harry Enten, ‘There isn’t a single time in which 28% of the American public thinks the country is going on the right track in which the incumbent party actually won.’”
In RedState, Bonchie said “we need to talk about that Selzer poll.”
“I don't think Selzer's final offering in Iowa is anywhere close to reality, and there's empirical data to support that viewpoint. For example, the poll has Harris leading with seniors by 19 points. Trump won seniors there by nine points in 2020. The idea that Trump has lost 28 points among seniors in a relatively red state just doesn't compute,” Bonchie wrote. “The overall composition of the electorate in Selzer's poll would be Democrat +3. Again, this is a state that Republicans [won by] +8 in 2020. Nothing else in the crosstabs makes any sense either.”
“How did we end up here? Did Ann Selzer release this poll to juice Harris given her long history of being allied with figures like Hillary Clinton, Claire McCaskill, and J.B. Pritzker? I don't know, and I'm not going to go that far. It's possible she genuinely ended up with an outlier and had the guts to go ahead and release it anyway instead of massaging it like some other pollsters would have done,” Bonchie said. “With all that said, I have no idea who is going to win this election. In fact, with just a couple of days left of voting, I'm more convinced than ever that this is a pure toss-up race. What I'm sure of, though, is that this Selzer poll is not close to what the real result will be in Iowa, and those using it to predict a Democratic landslide nationally are fooling themselves.”
In PJ Media, Matt Margolis wrote “maybe just forget the polls.”
“As the election nears, the polls tell us it’s anyone’s game, but seasoned observers and political instincts suggest otherwise. After a significant momentum surge for Donald Trump, the mainstream narrative is attempting to suggest that his lead has mysteriously evaporated, with Kamala Harris suddenly performing better in critical states. But for many who’ve followed these races for years, the reality may not be as close as polls are letting on,” Margolis said. “Before polling became the election-season industry with an endless supply of polls to sift through, reporters once had to size up a candidate’s standing by observing their rallies, organization, and crowd energy.”
“Perhaps the bigger picture is going to tell us a lot more about how this election is going to turn out. Are the polls blinding us to the bigger picture, especially when the broader indicators of enthusiasm and voter discontent weigh heavily against Kamala? Maybe,” Margolis wrote. “At a time when polling is less reliable because pollsters are either trying to influence voters or are too scared to make a prediction, perhaps it is time to just go old school.”
What the left is saying.
- The left still views the race as neck and neck, despite recent polls in Harris’s favor.
- Some say the Selzer poll is ominous for Trump.
- Others argue polls have lost their value and become a political tool for campaigns.
In The New York Times, Nate Cohn said the race is “still a deadlock.”
“Usually, the final polls point toward a relatively clear favorite, even if that candidate doesn’t go on to win. This will not be one of those elections,” Cohn wrote. “While the overall poll result is largely unchanged since our previous wave of battleground polls, there were some notable shifts. Surprisingly, the longstanding gap between the Northern and Sun Belt battlegrounds narrowed considerably, with Ms. Harris faring better than before among young, Black and Hispanic voters, while Mr. Trump gained among white voters without a degree.”
“The overall effect of these swings is somewhat contradictory. On average, Ms. Harris fared modestly better than our last round of surveys of the same states, but her gains were concentrated in states where she was previously struggling. Meanwhile, the so-called Blue Wall (Michigan, Wisconsin and Pennsylvania) does not look quite as formidable of an obstacle to Mr. Trump as it once did. As a result, Ms. Harris’s position in the Electoral College isn’t necessarily improved.”
In Talking Points Memo, Josh Marshall assessed “the Selzer poll.”
“I think the Selzer worship is a bit over the top. But she has a very good record. What makes this pretty hard to make sense of is that if Harris is really positioned to win Iowa or even come close that would suggest we’ve pretty dramatically underestimated Harris’s electoral power. Like REALLY underestimated her electoral power. As you know, I’ve long believed there’s a good chance that pollsters are doing just that. But this would be at a more dramatic level,” Marshall wrote. “Even for a pollster with a great record, it’s just one poll. Polls can be off. It’s also the case that Iowa is a state with a lot of white, college-educated voters. Democrats do pretty well with those voters right now. It’s also not that long since it was a swing state.”
“Probably the best way to interpret this is to see it as a bit bad for Trump in directional terms and not get too hung up on the specific numbers. But even that may not really add up since the actual result is so hard to believe that I’m not sure it makes sense to pick and choose — accepting that it must be good news for Harris in directional terms while dismissing the actual results as just not credible. One way or another, it’s an ominous sign for Trump. How ominous? How important? I really don’t know. But not good for him.”
In The Nation, Chris Lehmann suggested “no more polls.”
“For all the election-season lamentations over AI mischief, deep fakes, and dis- and misinformation, there’s a central source of toxic data derangement hiding in plain sight: the erratically useful, wildly misleading, and ideologically disfigured polling industry,” Lehmann said. “The ways that hotly touted polling findings distort and disfigure our basic understanding of what’s going on are teeth-gnashingly familiar by now. Just over the past month, we’ve had news that the Rasmussen Group, the long-standing right-leaning polling operation, has been sharing advance findings with the Trump campaign.”
“In our own age of counter-majoritarian, blinkered negative polarization, polling increasingly functions to reinforce the narratives campaigns put forward about surging popular support and finely crafted appeals to undecided swing state voters. As the Rasmussen episode shows, some pollsters even appear to be distorting their own research,” Lehmann wrote. “So here’s a modest proposal: Let’s block poll findings out of the final stage of our presidential elections. In the hectic last lunge toward Election Day, the defects of polling become greatly magnified, as voters are prone to use the alleged shifts in the mass electorate’s mood as rationales for casting ballots in a fog of empirically dubious pseudo-pragmatism—or else to refrain from voting at all.”
My take.
Reminder: "My take" is a section where I give myself space to share my own personal opinion. If you have feedback, criticism or compliments, don't unsubscribe. Write in by replying to this email, or leave a comment.
- Selzer is putting her reputation on the line here, and her poll has me reimagining the state of the race as we enter Election Day.
- The poll could indicate a number of things — from a possible Harris blowout to nothing at all — but it definitely isn’t a ruse from Democrats.
- I wouldn’t read too much into any poll right now.
The Tangle team has all assembled in Philadelphia for election week, and when the Iowa poll dropped it immediately caused a ruckus. My first thought was: Can I take my prediction back?
I'm kidding — mostly.
You can interpret Selzer's poll in a few ways. Before I break those ways down, though, I want to counter a bit of nonsense I'm seeing about it online. A few prominent talking heads have suggested the poll is a "psyop" or “gaslighting” or some attempt to "suppress" the Republican vote. That read strikes me as very, very silly. For starters, Selzer isn't going to torch her sterling reputation in an attempt to play politics — she's been doing this a long time and is widely respected across the political spectrum. Let's not forget Trump was, himself, complimenting the quality of Selzer's polls just last year (when they showed him with a big lead in Iowa).
Second, I don't know a single Trump voter anywhere in the U.S. who would decide not to vote because of one poll in Iowa showing Harris winning. It makes zero sense as some kind of suppression tactic or Democratic plot.
That being said, it's worth pausing here to make the case that we should take Selzer seriously. In 2020, a lot of people thought Biden was going to blow Trump out, with some polls showing the candidates tied in Iowa — then Selzer released a poll showing Trump up by seven in the Hawkeye State, which gave the impression the national race would be much tighter. Biden ended up winning by a whisker across the swing states, largely vindicating her.
Here’s what Selzer’s recent record looks like:
Does she miss? Yes, she does. Would I bet against her? No, I would not.
The most interesting aspect of this poll is that it’s out there at all. Sometimes, when pollsters get an outlier like this, they might spike the poll. Selzer hasn’t. She clearly feels some level of confidence in her results, and she certainly understands that people are going to remember it — and judge her accordingly. So, here are those different ways of interpreting this poll that I mentioned at the start:
1) Selzer, as she has in the past, is outperforming her peers. If this poll is close to accurate, Harris will win in an electoral blowout, and likely take all the battlegrounds (with the possible exceptions of Arizona and Georgia). Remember: Iowa is important because it has a lot of white and white working class voters, which makes it a good barometer for other states like Wisconsin and Michigan, with some signals about the others thrown in, too.
2) Selzer is off, but directionally correct. Maybe Trump will win Iowa by four or five points instead of eight or nine, and Selzer is capturing some late-break toward Harris. Even in that situation, the blue wall is almost certainly a lock, securing Harris’s road to the White House.
3) The poll is a complete outlier. Sean Trende, the expert who runs RealClearPolitics, said about one in 20 polls will be "flat-out wrong" — legitimate outliers. Perhaps a well respected pollster just happened to drop an outlier poll at an incredible time.
4) Maybe, just maybe, the poll is both right and Trump could still be in a good position to win the swing states. I don’t find this likely, but it is possible. Iowa is not being treated like a swing state this year, but maybe — given that neither candidate is spending much time there, and the state isn't being blanketed in ads — it’s just moving differently than the midwest swing states. That outcome would obviously be something pollsters end up analyzing for years, but it doesn't seem out of the realm of possibility to me.
If Selzer is capturing some genuine break with the polling average, it obviously begs the question: What happened? One answer is that Republicans in Iowa passed a statewide, six-week abortion ban, despite it being a relatively pro-choice state. Another is what the New York Times poll also picked up on, which is late-breaking independents going toward Harris. We also have a lot of data that Harris is outperforming Trump among women and voters older than 65, offsetting any potential erosion in black or Latino support. That, too, is consistent with the crosstabs in Selzer's poll.
Or, perhaps, it is what Nate Cohn — the chief pollster at The New York Times — said this week: Pollsters have been so concerned about underestimating Trump again that a lot of them have been spiking or under-weighting polls that look really good for Democrats. We have some statistical evidence for that, too. Nate Silver recently explained how it is a statistical improbability that so many polls are showing a coin-flip race, which is evidence of "herding" — pollsters all publishing neck-and-neck races to avoid being wrong.
Or, maybe, it’s the exact opposite. Cohn wrote that some evidence indicates Trump's support could once again be undercounted. Cohn said this about the non-response rate in The New York Times' final poll:
Across these final polls, white Democrats were 16 percent likelier to respond than white Republicans. That’s a larger disparity than our earlier polls this year, and it’s not much better than our final polls in 2020 — even with the pandemic over. It raises the possibility that the polls could underestimate Mr. Trump yet again.
The truth is, once again, that all the early voting, polls, and late shifts can only tell us so much. Based on information I had two weeks ago, I predicted that Trump would win the election but lose Pennsylvania. For me, predictions are a good way to test my hypotheses, and if I'm being totally honest (which I always promise to be) I think if I were making my prediction today, with the benefit of the information we've gotten in the last two weeks, I'd probably pick Harris. But I'm also not feeling so moved as to abandon my initial call. I just think Trump has more paths to the presidency, and so I'm hanging tight for the final 48 hours.
My expectation is a very tight race in all the battlegrounds, but one that could easily end up in an Electoral College blowout if there is a polling error in one direction. We'll see. In the end, as is true in every presidential election, the only result that matters is the actual vote.
Take the survey: Who are you voting for in the 2024 presidential election? Let us know!
Disagree? That's okay. My opinion is just one of many. Write in and let us know why, and we'll consider publishing your feedback.
Help share Tangle.
I'm a firm believer that our politics would be a little bit better if everyone were reading balanced news that allows room for debate, disagreement, and multiple perspectives. If you can take 15 seconds to share Tangle with a few friends I'd really appreciate it — just click the button below and pick some people to email it to!
- Email Tangle to a friend by clicking here.
- Share Tangle on X/Twitter by clicking here.
- Share Tangle on Facebook by clicking here.
Your questions, answered.
Q: How much does the 24-hour news cycle and the late night talk shows that do nothing but openly feed us with their political bias affect the way people vote or how they feel about the candidates?
— Jennifer from Salt Lake City, Utah
Tangle: It matters a ton. Maybe this is my bias as the guy who has made media echo chambers his sworn enemy, but 24-hour news cycles, cable TV pundit shows, late night talk shows, radio talk shows, print media, digital media, social media — it all adds up and has an enormous impact on the way people think about politics (let alone the individual candidates themselves).
It happens very easily, but slowly, over time. Maybe you hear someone on TV say that a candidate you don’t know a lot about yet but are somewhat interested in will be terrible on the economy. You hear that once, and you can write that off as just someone’s opinion. You see a news article mention offhand the issues that candidate has with voters on the economy. Then you see another. Then you go online and you see posts mocking that candidate’s stance, or snide comments talking about it, and over time it all adds up to just a background assumption you have: “I like Candidate X, but I know they’ll hurt the economy.”
But how do you know? At this point, you’ve never actually read their economic stances. You didn’t read an article that convinced you, and you didn’t have a good-faith debate with someone who challenged your worldview. You only “know” it because the media you get exposed to — on TV, on the radio, in print, online — all comes from sources you already know you like, who say things you already agree with.
Over time, that shapes not only how you see political candidates but politics in general, and people who think differently from you, and the whole world around you. It’s one of the defining forces in the media ecosystem, and something that takes consistent, daily effort to push back on.
I hope deeply that I’m helping provide a way for you to do that, without requiring too much effort on your part.
Want to have a question answered in the newsletter? You can reply to this email (it goes straight to our inbox) or fill out this form.
Under the radar.
Over 75 million Americans have voted early as of Sunday, roughly 48% of the total number who voted in 2020 (154.6 million). Some of the highest levels of early voter turnout are in swing states. In Georgia, more than 4 million have voted, nearly 80% of the state's 2020 total; in North Carolina, 4.4 million have voted, also roughly 80% of 2020 turnout. Of those who have cast their ballots in states that report party registration data, Democrats have voted at a slightly higher rate than Republicans — 37.9% to 36.2%. Overall, the high levels of early and mail-in voting suggest Americans increasingly prefer these methods after the Covid-19 pandemic prompted greater adoption of early voting in 2020. Axios has the story.
Numbers.
- #12. Ann Selzer’s rank (out of 282) in FiveThirtyEight’s rating of pollsters (based on historical track record and methodological transparency).
- #1. The New York Times/Siena College’s rank in FiveThirtyEight’s rating.
- #22. AtlasIntel’s rank in FiveThirtyEight’s rating.
- +15. The increase in Kamala Harris's support in Iowa compared to Joe Biden’s in the last Des Moines Register poll conducted before he dropped out.
- 58%-42%. Harris’s lead over Trump among voters who made their voting decision “in the last few days,” according to the latest New York Times/Siena poll.
- 11%. The percentage of voters who say they remain undecided or persuadable.
- +4% and +1%. Harris’s improvement with black and Hispanic voters, respectively, since the last Times/Siena swing state polls.
- +6.5 and +5.5. Trump’s lead over Harris in Arizona and Nevada, respectively, according to AtlasIntel’s November 2024 poll.
The extras.
- One year ago today we had just written about the mass shooting in Maine.
- The most clicked link in Thursday’s newsletter was the internet’s spookiest video.
- Nothing to do with politics: The history of why we turned our clocks back this weekend.
- Thursday’s survey: 1,833 readers responded to our survey asking about Kamala Harris’s Washington, D.C., rally with 68% saying it will have no effect on how they vote. “Already voted! Too little, too late. The best thing about her short presidential run is that it proves we do not need ‘months’ of campaigning. 90 days is sufficient,” one respondent said.
Have a nice day.
Access to polling places can be an impediment to voting. A 2021 study from Boston University and Harvard University found that, in 2018, 66% of registered voters with a car voted. However, only 36% of registered voters without a car did. To address this disparity, the ride-sharing app Lyft introduced the Lyft Up Voter Access Program to provide free or discounted rides to polling places. In its decade of operation, this program has helped 3 million individuals vote, and the company aims to increase the number of individuals benefiting from it in 2024. Lyft announced the story here.
Don't forget...
🎥 Follow us on Instagram here or subscribe to our YouTube channel here
💵 If you like our newsletter, drop some love in our tip jar.
🎉 Want to reach 190,000+ people? Fill out this form to advertise with us.
📫 Forward this to a friend and tell them to subscribe (hint: it's here).