What A Day - Are The Polls Wrong...Again?

Episode Date: October 26, 2024

Trump is winning the presidential race according to some polls. But others say Harris is ahead. What’s the point of following the polls if they contradict each other and, at times, seem outright broken? On this week’s “How We Got Here,” Max and Erin talk with Crooked’s Dan Pfeiffer to explain how Trump, the pandemic, iPhones and more messed with the reliability of presidential election polls.

So, Max, I was looking at the polls earlier, and I just don't know what to make of them anymore. Oh, tell me about it. One poll has Kamala up in Pennsylvania, but down in Arizona. Another has the reverse. Some have her up in both. And Mercury is going to be in retrograde. I mean, even the polling nerds are losing it. Here's CNN's election guy, Harry Enten, a few days ago. We talk about a historically close race. Oftentimes, we talk about the national polling. But I wanted to try and dig into the state numbers to get an understanding of how close this race is. Holy cow. Harry, tell me about it. This whole thing has got me pulling my hair out.

Me too, which just has me checking the polls even more. To try to get some clarity. Yes, I'm checking polls at breakfast, sneaking polls all day at work, refreshing the forecaster models right before bed. That sounds like bad sleep hygiene, Max. It's bad. I'm not sleeping. All those polls and forecasts, and none of it ever seems to bring us any closer to knowing who is going to win this election. Which, when you think about it, is kind of weird, right? I'm Erin Ryan.

I'm Max Fisher. And this is How We Got Here, a series where we explore a big question behind the week's headlines and tell a story that answers that question. Our question this week, is polling broken? You hate to beat up more on election forecasters who took such a lashing in the last two presidential elections. Well, that's my point. The forecast models got it wrong the last two times. And this time around, they can't even tell us where the race stands. Why is that? Well, it's another way of asking how we should be reading the polls in these last couple of weeks to account for all that weirdness in polling.

Like, should I be following some polls but ignoring others? Should I be tracking the early vote returns, the approval gap? Should I be flipping over some tarot cards? I mean, I'd like to know if only for peace of mind. Well, the bad news is that we can't tell you who is going to win the election. But the good news is that we can answer what is going on in polling these last few elections. Thanks to help from a special guest who's joining us this week, Crooked Media's own Dan Pfeiffer. Oh, I've heard of him. That sounds familiar. Dan is a host on Pod Save America and the subscriber only podcast Polar Coaster, which you guessed it, digs into the science of forecasting elections. Could not be a more perfect person to demystify all this. I asked Dan to start by explaining what the polls got wrong in 2016,

since one big fear is that this election could be a repeat of that one in terms of the polling error. I will never forget Election Day in 2016. We all thought Hillary Clinton had had it sewn up. I was assigned pre-writing a piece to be posted that night celebrating her victory. Oh, no. And one of my colleagues was assigned the Trump victory piece. And at 2 p.m., she and I went into the conference room to eat our, like, sad desk salads. And she had only written, against all the odds. That was all she'd written for the piece. That's it. Just that phrase. Yes, against all the odds. And that was it.

Well, she was right about that because she was reading the polls. And like, I'm pretty sure even Donald Trump thought that he was going to lose. Well, it's what the polls said, right? But then the results came in and whoopsie daisy, Trump is the president. I asked Dan whether the 2016 polls were really as wrong as we all remember, and here's what he said. Well, they were off, but they weren't off by as much as we thought. We're constantly confusing our quantitative analysis of the polls with what we think is

going to happen based on history, the political environment, et cetera. And the race of 2016 was much closer than we thought. Now, one of the challenges back then, and Clinton was ahead by a couple points the whole way, but back then there was very little public, high-quality state polling. So what people were blown away by is there just wasn't enough information. She was up by a few points nationally, but what we didn't know is that she had a soft underbelly in Pennsylvania, Michigan, and Wisconsin. And her campaign didn't relate to that either because she barely campaigned in any of those states

and did not campaign at all, somewhat infamously, in Wisconsin. So the race was closer than we remember it, but the polls were off by a decent amount. And they were off in one very specific way, which is they underestimated Trump's support and Trump's turnout in those blue wall states. Oh, God. It's like I'm remembering that feeling and just— I hate to relive it. I hate it. I don't like it at all. Let's not do it again.

No, I don't think we should do it at all. It's wild to me that they were just, like, not polling the upper Midwest. Today we get a new poll of Wisconsin every six hours. The thing to remember is that presidential elections looked very different before 2016 for one simple reason. Whoever won the popular vote almost always won the presidency. But what about 2000 when George W. Bush lost the popular vote but won the electoral college? Thanks again, Florida. Yeah, that's an important case.

So that was the first time that that had happened since 1888. So it felt at the time like a total fluke. And everyone assumed that if Clinton was ahead by three or four points nationally in 2016, then she was going to win. What no one had realized, what had changed was that Republican support had moved around geographically in ways that gave the GOP a path to winning the electoral college, even with only 46 or so percent support nationally. And that path runs through a handful of states like Pennsylvania and Michigan, which is why we now pull the hell

out of them. So you're saying that this cycle we live in now, where we assume that Democrats will probably win the popular vote every time, but Republicans might still win the electoral college, that only started in 2016? Yeah, and seemingly because of Trump changing who votes Republican now. Here's Dan again. Trump changed the makeup of the Republican Party. And so a bunch of people who had voted for Obama voted for Trump in those states, rural voters. And so the electoral college just became more Republican than the nation as a whole. At the same time, Trump was polarizing places like California and New York to make them more Democratic. And so you got this split that people did not see coming. Oh, wow. It's almost like the system that was designed almost 250 years ago was not completely flawless. It did not actually make a ton of sense.

No, and it's actually kind of starting to become a little bit unraveled. But going back to the polls here, the thing that pushed Trump over the line in 2016 was that he got way more votes than people expected him to get from one group in particular, which is white people without a college degree. Right. Also known as working class whites. When you hear pollsters say working class, in that context, it just means that person did not go to college. It's a really, really big demographic. 44% of the electorate as of 2016. And in that election, Trump got two thirds of the vote from that group. It reminds me of a line from Blazing Saddles. No, don't do it. I'm not going to do it.

Don't do it. I know that in retrospect, it's easy to look back and say we should know Trump would win working class whites in a landslide. But remember, that's not quite what the polls were telling us. Yeah, this more than anything else was the big polling miss from 2016. So what happened? So it's not like pollsters weren't surveying working class white voters. They were. But that did not get translated into the election forecast for two reasons. And first was turnout. To forecast an election, pollsters have to predict who's

actually going to show up on Election Day. And Americans are famously lazy about voting. Only 60 percent of eligible voters bothered to do so in 2016. So it really matters which 60% show up. Pollsters knew working class whites were tilting Trump. But what none of us knew was that Trump was going to get that group to turn out at much higher rates than it had in past elections. And that is what scrambled the predictive models. So what was the second reason that the polls got 2016 so wrong? Response rates. People stopped picking up the phone in 2016, and that skewed the polls in a way that made Clinton look stronger than she was. Here's Dan again. Response rates have gone down exponentially since 2008. In the Obama 2008 campaign,

the 2012 campaign, they were cut in half. And that's in part because at some point in 2008, the iPhone was invented and it became easier to screen calls. And then it's gotten worse and even particularly worse as various cell phones and others have become very good at screening out spam calls. And so people became, like our modes of communication shifted to the point where we are more text and email than we were, than we are actually speaking on the phone. That's particularly true for younger voters. And in particular, it gets harder to reach young voters and working class voters in part, because if you're someone who's working, uh, working a job and maybe you're working a night

shift or you're, uh, at a factory or waiting tables or whatever. You can't take 45 minutes or 20 minutes, whatever, and answer a phone. Yeah, I never answer my phone. I don't either. I find it rude if a person calls me. Leave me alone. Yeah, if you want to know who I'm voting for the election, just read my damn Twitter feed. Don't call me. Yeah, exactly.

It should be pretty obvious. So between response rates, not seeing how Trump had scrambled the electorate, and understating working class white support, the polls made a few big mistakes that everyone is worried they could make again this election. Max, do you remember how angry everyone was with pollsters? Like every news network hauled out their head polling guy to yell at him on air, like practically tarring and feathering. Here was CBS News the day after the election. So how did the polls lead us astray, including our CBS News poll,

which is considered to be one of the best in the industry? Anthony Salvato is our director of elections. In other words, he is our polling expert. So the rest of the interview was just that guy getting yelled at. It was really rough. But after all this, you know, news networks, pollsters, election forecasters were all absolutely committed to learning from their mistakes

and getting it right next time. If only to save themselves from another round of national humiliation, a.k.a. eating shit. So then in 2020, pollsters applied what they'd learned. They rejiggered the models to account for this weird new electoral reality.

They ran dozens and dozens of polls in every swing state. And then they got it even more wrong than they had in 2016. Yeah, they sure did. Here's Dan and what happened. 2020 is the biggest polling error in history. And it was four points. And it was much bigger in the swing states. The polls were off by eight points in history. And it was four points and it was much bigger in the swing states. The polls were off by eight points in Wisconsin. And it just, and this is what is alarming about the whole thing is there was a lot of high quality polling. All the best A plus rated pollsters

started doing state polls. New York Times, Wall Street Journal, the Quinnipiac, all these other people who are quite good at what they do did did state polls, and they all still got it very wrong. And it was so wrong that the entire industry, the trade group of public opinion research professionals, put together a working group, and they tried to figure out what went wrong, and no one could figure it out. Really? They never got to an answer? They never got to an answer. Wow. There are so many things that happened.

It really was a black swan event in the sense that you had a pandemic. There are theories that came out of this. One theory is that during the pandemic, liberals were at home, therefore more likely and available to respond to polling. And Republicans and more conservative people went back to work or were out in the world. And particularly if you lived in a Southern state where we're less restrictive on movement during the pandemic. And we once again, pretty dramatically undercounted Trump's turnout.

It was, I mean, just a massive fail across the board. Okay, well, you know, in that answer, Dan said all these people who are quite good at what they do. At what point... No, you're not quite good at what you do if you've just spent eight years fucking up. I believe that they tried hard.

Okay. So I would try hard playing for the New York Yankees. That doesn't mean that you're quite good at what you do. And I would strike out every time. And you can't say that I'm the Quinnipiac of the Yankees. That doesn't mean that you're quite good at what you do. And I would strike out every time. And you can't say that I'm the Quinnipiac of the Yankees. Or maybe you could, but I wouldn't make the team.

We don't call me a good player. Like, can we? This is the thing that really bothers me talking about polls, because I did pretend to like them at the beginning because that's what was scripted for me. But, like, why are we allowing them to drive the conversation around how the campaign is going? Yeah. What else do we have? What else do we have to look at? 300 million people. What are they going to do? We have to

look at something. I don't know. I guess the campaign events. So the polls were world historically wrong in 2020, but wrong how? Like, what did they miss? So just like in 2016, they underestimated turnout and Trump support among white voters without college degrees. Again? Same mistake they made in 2016? Well, the same mistake, but more so. And in fairness, those numbers were at least off for different reasons. So it's not like they made literally the exact same mistake.

Rather, it was a new set of mistakes that just happened to have the same result. Okay. So it's like if I got a car and I just crashed it into a river, and then I got a new car, and I crashed this one into the side of a building. So the pollsters are winning you over, is what you're saying? I just think, I just think if this were a female-dominated profession, this would be something that would be dismissed as pseudoscience instead of treated like, oh, these geniuses are just working. They're very artists.

And like, we would run the polling section next to the funny papers, the horoscopes. You know what? If you had a doctor that was as exact as pollsters, they would operate. They tried to do heart surgery on your arm. They would be cutting through the wrong tendons. But these are the doctors we have. I'm just not going to go to the doctor. I'll let my body heal itself.

It's funny because pollsters objectively blew 2020 way worse than they had in 2016, but people were not mad in the same way because the pollsters projected that Biden would win in 2020. And at that point we were like, he won. It's fine. It's fine.

Water under the bridge. So Dan brought this point up too, actually. Everything It's fine. It's fine. Water under the bridge. So Dan brought this point up too, actually. Everything was off in a pretty dramatic way. And we just don't remember it that way because Biden won. And the poll said Biden was going to win and Biden won. He just won by so much less people thought. People thought this election would be called by 11 p.m. on election night. And obviously it took several days and Biden took 40,000 votes over a small handful of states to make him the president. It's one of the closest elections in history. I just got to say, we're talking about the news, right? Literally, astrology has been more accurate.

And I'm not being completely hyperbolic. I don't know if that's true. Come on. Yes, it is. Joe Biden decided he wasn't running for re-election on the last day of the Capricorn full moon this July. And, and... So you're saying you saw it coming? I'm saying a lot of astrologers saw it coming. Did they really?

Yes. I didn't know that. Yeah. I mean, the pollsters didn't. Did they? Here's the thing. For as wrong as election forecasters were in 2016 and 2020, they were actually pretty dead on in 2018 and 2022.

Both midterms in which Democrats performed much better than expected. Well, better than expected, according to talking heads and newspaper columnists who all predicted a red wave. But if you read the polls, then you knew to expect that Democrats would do well. Except nobody believed the polls after they blew it in the presidentials. It is a pattern, yes. Here's Dan on the polls nailing the midterms. It is easier to poll the midterm, a much smaller population of people. And more of that, the voters who are least likely to respond to polls are least likely to make it through a voter screen in a midterm election because they're unlikely to vote in a midterm. And so they were right in 18, they were right in 2022. Also, no Trump on the ballot. That's the other factor that is out there. So that sort of surge in turnout that was not caught in the polls in 16 and 20 did not exist in 2022. So it seems like a big lesson here is that Donald

Trump is to elections what a magnet is to an old-fashioned TV. Like he just scrambles voter behavior in a way that is so outside the norm that pollsters haven't really figured out how to capture it. I just realized some of our listeners probably don't know that if you put a magnet on an old TV. I guess that's true. Yeah. It wasn't good. It wasn't good. Is that not true with the new TV? I've never, don't do it, kids. Don't put a magnet on the TV. If you do it, write into us and let us know what happens. Yes, please. Because pollsters had not figured out how to capture the Trump effect as of 2020 for sure.

But I asked Dan whether they figured it out now. And of course, the short answer is we'll find out in November. But the longer answer is that some pollsters are testing a new metric to help them finally capture or at least try to capture the Trump effect. Here's Dan. The industry has taken a whole bunch of steps and campaign pollsters have taken more steps because they have more money and more time. But generally in the industry, one thing that most pollsters are doing to try to account for Trump turnout is what they're doing is in polling, it is typical to ask people who they voted for in the last presidential election. They then wait those polls to ensure that the makeup of the race includes the right number of people voted for

Trump, the right number of people who voted for Biden, which is about a four-point Biden popular vote lead. And the theory there is we're assuming the 2024 election will look like the 2020 election. And therefore, one way to ensure the polls reflect the Trump turnout is to weight it to a known election when there was Trump turnout. So there is a name for this technique. It's called weighting on recalled vote. The New York Times ran a whole column about it because their poll is one of the few that does not use this, which is why their projections are so different from everybody else's. So this is why that one Times poll had Kamala up four in Pennsylvania, but down like six in Georgia.

Yeah. Although the models that use this technique and the models that don't use this technique both still predict basically a dead heat toss-up race, just a slightly different toss-up. So maybe it's a distinction without a difference. Well, the point here is that all of these elaborate modeling techniques and weighting methods are all meant to try to capture the Trump effect on the race because a lot of swing Trump voters behave in ways that are so unusual and so outside of the norm that it makes them very hard to predict. It's like if the entire bajillion dollar polling industry was built around getting inside the head of the weirdest 4% of the country. Yeah, and the thing that makes this all excruciatingly difficult is that because the race is so, so, so close, it's not just that pollsters have to design a technique to predict Trump voter behavior. In order to be precise enough to tell us how the race is going to turn out, it has to predict Trump voter behavior with something like 90% accuracy. Is that even possible?

It is not. No. So doesn't that mean that the polls election forecast modeling is like meaningless. Like, like read my palm. Tell me how the election is going to go. I mean, I don't know. Like Kamala Harris is a Libra and Trump's a Gemini. What does it mean? It just seems kind of like they might as well flip a coin. Yes. For the purposes of this moment and this election, that is exactly what it means. And that's exactly what the election forecasters will tell you. Here's Dan again. Yeah, they're absolutely a poll that has Kamala Harris up to and a poll has Kamala Harris down to. Frankly, a polling average that has Kamala Harris up to and a polling average that has Kamala Harris down to tell you the exact same thing, which is that this

is an incredibly close race. The result will be within the margin of error of the polls and the four to 6% of undecided voters will make their decisions and decide it. And the polls can't tell you what's going to happen. So there could be a world where we go into election day and Donald Trump has a two point lead in Michigan, Pennsylvania and Wisconsin in the battleground states and Kamala Harris wins those states by one. The polls were not wrong. That was essentially within what the polls were telling us was a possible outcome. Okay. Here's my thing. This episode is going to break you. I am like, I am right. Hanging on by a thread. I don't, I cannot compute how Donald Trump has somehow expanded his

voter base, especially in these last weeks of the campaign where he seems to just be behaving in ways that are so deeply weird that what he's trying to do is turn out every possible person who might vote for him which has been his strategy the last two elections just turn out every like don't worry about the people who are middle of the road don't worry about like you know the liz cheney's of the world just get like look underneath every you know scrape the Liz Cheney's of the world, just get like, look underneath every, you know, scrape everywhere to get every single possible Trump voter by acting in a way that really gets them excited. I do not see the number of those people that are out there growing. If anything, I see them being more impacted by, for example, the COVID-19 pandemic. People who were strong Trump believers

were more likely to get seriously ill, be disabled, or possibly die as a result of COVID-19 and the misinformation that he spread. So, like, where are these new voters coming from? I mean, I know that he's made a play for young Black men, disenfranchised, angry dudes, but they're in a demographic that is like unlikely to vote. You know what I mean? Like, where are these people coming from? Well, here's the thing. It's not actually clear that there are more Trump voters than there were in 2020. First of all, there is some indication that he is losing ground in blue states, but also gaining ground in already red states, which because of our system, none of that actually matters, of course, if he gets more votes in Florida or loses

a bunch of votes in California, New York. The thing to remember is that 2020 was decided by like seven votes in the upper Midwest. And where the polls are means that he could actually be at significantly less support than he had then. But he could also be at enough of a higher level of support in those states to win. It's just that that's like 10 votes. Here's another thing that's different between 2020 and 2022 that Dan didn't bring up, and I'm sure he would have if we would have been able to talk to him for longer. But we have the X factor of Roe v. Wade being overturned and the Dobbs decision and the continuing charge that the issue of abortion and bodily autonomy are giving to voters who, even before Dobbs, were not necessarily charged up by abortion because they never thought that Roe v. Wade

would be overturned, or they'd never thought about it deeply enough to consider the impact that that would have on the lives of women. And it had never occurred to people how deeply weird it is to put up the bodies of half of the population to a state legislative vote. Like, that's fucking weird. It's a weird way to do things. And so I feel like there isn't an issue that people are as charged up about on the Trump side of things as there is abortion. There's just nothing to me. And I'm just maybe I just like hang out with a lot of women. But there's nothing to me that seems in any way equivalent in terms of enthusiasm. And we I don't know if like pollsters are—it's just—

I mean, this is what is tough about predicting voter behavior, especially because you are ultimately talking about a very, very small number of potential swing voters in a small number of states. It's like, I'm sure that there's an effect of charging up especially women voters in those states. At the same time, if you poll people and ask them, what is the issue that you care about? They think of the Dobbs decision as something that happened in the ancient past, you know, two years ago, was a lifetime ago. And what they what they care about, if you ask them, is the more recent issue that has affected them, which is inflation, which is like way, way and above the thing that at least swing voters and independent voters say is the number one motivating force for them. It's not what would be my choice, but I'm not the kind of voter who's

going to decide this election. But then there are referendums on the ballot in Arizona, I believe in Nevada and definitely in Florida that directly impact abortion rights in that state. And so that's another factor. You know, if abortion is like super important to you, even if it's the only thing that's important to you, you can actually go to the polls and vote for that one thing. And you should. And you should. And you can also vote for the candidate that is not going to have us losing any more ground when it comes to abortion rights and bodily autonomy. Makes sense to me. Well, Max, when we started this episode, you said we would not be able to answer who will win this race.

And in fact, I feel even less confident that I know what will happen than I did at the start of the show. And I'm now got to be in my bonnet about it. So it's distressing. I know we all want to feel reassured that we know how the race will turn out or we know why it is in the state that it is in. But that has given rise to this entire, frankly, snake oil trade and people who promise to have the one secret metric that will tell you how the race is going to turn out. Oh, yeah, like that octopus that would pick World Cup teams. OK, so, yeah, just in the last 24 hours, I've seen people claim to be able to predict the outcome of the race based on early turnout numbers in Pennsylvania, opinion surveys about inflation and even campaign fundraising numbers. All meaningful, but none of it enough to actually tell you how those last

couple million swing voters are going to fall. But Dan did mention one metric that he thought bears some attention, which is that very simply, Kamala has a much higher approval rating than Trump. Look, everything, you know, Trump is an unprecedented candidate, but in general, in history, the undecideds break more for the better-liked candidate. And we know that Trump just has a pretty hard ceiling at around 49% of the vote. And so she just has it. It would, on paper at least, have a greater chance of getting to 49, the 49.5% it could take to win these races. But it is as close as it could possibly be. As close as it could possibly be between a former senator, vice president, district attorney of a major, major population center and a guy who talks about golfers junk.

Yes, that is that is an accurate assessment. But the closest means that that one or two or three percent of voters who have still not made up their mind. I can't understand. I can't understand it. How do you still not know? I know it's hard to wrap your head around, but these people exist in large enough numbers that they will almost certainly decide whether Trump or Harris becomes president. And they probably haven't picked yet. Which is terrifying. It's like there was a choice between two types of ice cream. One of them was like vanilla and one of them was just something crazy that nobody likes, like worm ice cream. And you're like, ah, which one am I going to have to eat?

Anyway, it's terrifying. But it's also good news because it means that it is not too late. This election is not only still up for grabs, but it is going to come down to such an excruciatingly tiny number of votes that you as one person can actually make a difference for once. Okay. Why do I suddenly feel like I'm in an after school public service announcement? Oh, it's because you are. Okay.

One thing I actually kind of love about working here is that we're not just a media company. Crooked also has a political advocacy arm, and they've put together this whole guide for how you, as a regular person sitting at your desk, can contact, if everyone listening to this podcast did that and contacted one person, literally the race would not be close anymore. I think it's so important to reach out to people that you already know who live in swing states, because I mean, how many texts from Jon Tester and Kamala have you ignored? But if I get a text from a friend of mine, it works. I'm much more likely to listen to them. I'm much more likely to engage on the election. And I'm much more likely to make a plan.

Now, I am probably the person among my group of friends who is most likely to vote if that was an award. But I'm just saying, you know, think about the way you would feel being contacted by somebody and how much better you feel about it if it's your friend, if it's somebody that you know. Absolutely. And that is so, you're not annoying people. You're reaching out. You can start a conversation. You can rekindle a friendship.

And you can deliver all the swing states to Harris walls in the process. And you could save America, not to over you know, overhype what you can do. Hell yeah. So, Erin, after everything we've learned about the polls, how they work and don't work, how they're being modeled for this election, what do you think? 538 gives Kamala a 49% chance of winning. Nate Silver gives her 47%. Do those forecasts mean anything to you? Absolutely not. Nothing whatsoever. I feel like what those forecasts tell me is that these pollsters want to make sure that they're not wrong by claiming every possible viewpoint and every

possible scenario as a possibility. So then they can be like, I was right. See, I told you that this would happen. And it is a disease that inflicts punditry, opinion journalism. And I feel like what we as a country need to do is go to therapy and get comfortable sitting in uncertainty. Because uncertainty is deeply uncomfortable. And all we can do is all that we individually have the capacity to do. And there are certain things that we cannot understand and we cannot control. All we can control is ourselves and our reaction to things. And we just, look, the Nates would be burned as witches if this were the 1600s. Bad witches, too, that aren't even good at being witches. I really am going to just treat the election like it could go either way. And if I

don't give it my all between now and then, I'm failing my country. Well, I came out, I would say, slightly more sympathetic to the pollsters than you are. I think that it is an incredibly hard, veering on impossible task to try to predict how 170 million people are going to behave when our politics are so weird. So many voters make their decisions for such strange reasons. They make it on such short notice. And we are within such like 0.0001 percent margin of victory for all of this that I'm sympathetic. It seems hard. But you and I agree that the polls right now, they don't tell you anything. No, There's no information polls. So why are they driving the news cycle?

That is the thing that really bothers me. Well, in fairness, we are recording an episode about them, although I guess I picked the topic, so this one is on me. But this is, for me, this is a chance for me to be a hater. And we love that. That for me, because it feels cathartic, especially given the fact that I feel kind of like we're trying to figure out blindly what is about to happen. I'm like trying to think about what the next several months of my life are going to look like. And that varies just wildly based on what happens in this election. Well, the good news, as you said, is that one reason that even if you are sympathetic to the polls, one reason that they don't tell you anything is that the race is incredibly close. But the race being incredibly close means that you really can make a difference. You go canvas, you call someone.

Just it would take such a small number of people doing that to swing the race. So it's really possible. It is actually a very exciting moment where you as one person can change the course of history for once. So let's go out with this final comment from Dan, who raised what I thought was an important point about the difficulty of predicting the race when I asked him how his stomach was handling these final weeks. I've kind of come to the conclusion that no one knows anything anymore because we're just not equipped to analyze a race this close. And so every little thing people are saying that's happening could mean everything, could mean nothing, right? And there's just, there is something out there that we are not expecting in the numbers that is going to happen, right? Either this idea that Trump is making all these gains with black voters,

which would have to be totally fake, that Kamala Harris is going to win a larger share of the Republican. We just don't know what it is, but there's always one thing in the crosstabs of the final results that blow everyone away. And we just don't know what it is yet. And it could go either way. How We Got Here is written and hosted by me, Max Fisher and Aaron Ryan. Our producer is Emma Illick-Frank. Evan Sutton mixes and masters the show. Jordan Cantor sound engineers the show. Audio support from Kyle Seglin, Charlotte Landis, and Vassilis Vatopoulos. Production support from Leo Duran, Raven Yamamoto, and Adrian Hill. 다음 시간에 만나요