Weighting For Election Day
35:11 Share

Fall is planting season. It's true, many plants actually do better when planted this time of year, but it can be hard to know where to start. That's why over 2 million Americans trust Fast Growing Trees, the world's largest online nursery and the easiest way to get the yard you've always wanted. With more than 10,000 plant varieties, including fruit trees, evergreens, and houseplants, they have everything you need, no matter your yard or experience level.

Their plant experts provide best-in-class customer support and are available for free consultations via phone, chat, or email, giving you the confidence you need to grow. On top of it all, each purchase is backed with a 365-day Alive and Thrive guarantee.

That's the end of the show.

That's an additional 15% off at FastGrowingTrees.com. Use the code FALL15 at checkout.

I was just thinking back on voting during hurricanes. My polling place got moved in 2012 because my polling place was flooded due to Hurricane Sandy. Oh, yeah. I ended up voting in an elementary school or something like that in line behind Louis C.K. and his two daughters. Nice. I have relatives that work at that elementary school. Really? I'm not kidding. It's such a small world. We're all coming together. Such a small world of Northeastern media elites. Yeah, wow. ♪

Hello and welcome to the FiveThirtyEight Politics Podcast. I'm Galen Druk, and we are three weeks away from Election Day, and the race could hardly be closer, according to the polls. According to FiveThirtyEight's averages, every single battleground state is separated by less than a point, except for Arizona, where Trump leads by two points. If you are one of the millions of people who looks to the polls for a sense of who will win the presidential election, well...

you're looking at one big shruggy emoji. And that's where we're going to kick things off today, which is with the current state of the race. And we've also got a good or bad use of polling for you. It's a nerdy but important one. This cycle, pollsters have increasingly been using respondents individually.

past vote data to get a read on the electorate. In other words, they ask voters who they cast a ballot for in the last election and weight the sample that they get to make sure they have enough Trump and Biden voters from 2020. That may sound like a good way to avoid undercounting Trump voters yet again, but it comes with its own perils.

And we're going to look at the political reaction to back-to-back hurricanes in the Southeast. And we'll also ask what it means that Trump is hitting significantly more campaign stops than Harris. Here with me to discuss it all is senior elections analyst Nathaniel Rakich. Welcome to the podcast, Nathaniel. Hey, Galen. Also here with us is New York Times polling editor Ruth Egelnik. Ruth, you've been churning out those polls lately. How's it going?

Been a little busy. I'm a little exhausted, but I'm here. Exhausted, but here is actually our election slogan here at FiveThirtyEight. So like I said, the polls could hardly be closer. All the forecasts out there show basically a 50-50 proposition. The Scottish teens over at the betting markets see the same thing. And we have had very close elections this century. In fact, very close elections.

But when that happened, the polls were not actually this close in the lead up to Election Day. So, I mean, for example, in 2000, the polls suggested a four-point lead for Al Gore. In the last election, it looked like Biden was running away with it, at least when it came to the national popular vote before Election Day. Of course, 2016, you all know the story. So just given what the polling picture looks like today, it's

Ruth, how are you processing this election data environment? It's highly possible that, I mean, as you said, the outcomes were very close in these past elections, that we're just dealing with a more accurate polling that is similar to those previous elections. They were that close. We just didn't see it with the polling. I mean, there's no question that this is insanely, insanely close.

But I do feel like, you know, I mean, we have more high quality state level polling than we've ever had in the past, which also might be why we're seeing it so close, because we're actually able to get a lot of reads at these like very small sort of idiosyncratic places, which makes me feel good about, you know, sort of the state of what polling is telling us that it is a close race. There still could and might be polling misses, but it feels like we're getting a pretty good read.

So it's not a situation where everyone is looking at national polls and there's just going to be some really big surprises at the state level. It seems like everyone's gotten the message that the Electoral College is what decides who becomes president and everyone is pulling the seven battleground states.

That's right. I think we've all trained our readers and listeners so well that I literally have people coming up to me and saying, why are you even conducting national polls? They don't matter. I think they do matter, but... Yeah, I was going to say, almost given that, do we risk having too narrow of a scope in terms of what

we're looking at? I mean, should we also be looking at polls of the broader Midwest, including Minnesota or New Hampshire, say Florida and Texas? Because a lot can get left out if you're only focused on seven battleground states. Yeah. I mean, it's also, there's a question of trade-offs. There's finite resources. You know, let's say we have X number of dollars or hours or whatever the limit is to poll.

Yes, we should be looking at those other places. And yes, we do national reads to get a sense of the mood and feeling and look at subgroups. But in the end, we also need to read these very close places that are going to determine the election. So you guys all see what we do at The Times. We try to balance some of those things with more of a focus on the specific battleground states. But looking at things like Florida, where we think there's been this sort of interesting movement separation from like the Electoral College elections.

And the popular vote, you know, they're all trade-offs. Yeah, I mean, you all found a 14-point lead for Trump in Florida. I think the general vibe, at least, has been that Trump has been gaining ground over the past couple weeks. Is that true in the mix of polling that we're seeing? Is that more based on vibes than actual data? Because if you look at our national average,

I think a month ago, Harris led by 2.6 points. And today she leads by 2.5 points. Um,

So it's hard to see a lot of movement over the past month. Yeah, I'd call it less movement and more tightening. This race was always going to get tighter. Harris had a little bit of a bump after the announcement, walls, the convention, the debate. And this is a little bit of tightening and kind of coming back down to earth. We can talk about this later. But like when we look at our polls over time at specific states or nationally, they're pretty stable.

The distinction here is like if the question is, has the race tightened? Has Trump gained a little bit of ground? The answer is yes, but that's not how I would overall characterize the race. I would characterize it first and foremost as stable. If you look at the national polls, as well as some of the swing state averages that we have at 538, you're looking at maybe a Trump gain of like 0.5 points over the last month or so. That is not something that is that should change your conception of the race.

These polling averages themselves, you know, obviously, like we average the polls, so we think they're more accurate than any individual poll. But they also come with these confidence intervals, like basically a margin of error. And so people should not be sweating the difference, even if it goes from like a Trump 0.2 lead to a Harris 0.2 lead in a given state. That's still basically indistinguishable from a total tie. When we're talking about an environment where there are polling error

errors routinely of three or four points is just not something that's worth sweating over. So I think that generally speaking, the race has been a toss over the last several months. I don't think Democrats need to panic about the fact that Trump has maybe gained 0.5 in the 538 polling average. Yeah, we're going to talk all

also about the different kinds of polls that are going into these averages in a second. But I do want to talk specifically about the New York Times polls that y'all put out, which backed up actually something that we discussed on last week's podcast. And that's that looking at the averages, although this is a stable race, and although the top lines look similar to 2020, you are seeing some shifting happening, which is, of course, the

Voters of color shifting right, white voters shifting a bit left, a little bit of similar sort of cross pressures for younger and older voters than we usually expect. And as a result, we're seeing better numbers for Harris in the Rust Belt, better numbers for Trump in the Sun Belt. I think that narrative has...

has increasing purchase amongst analysts. I think people are more and more bought into this, and that as a result, the electoral college gap between the popular vote and the electoral college may be lessened. But I do want to ask, as this gains purchase with elections watchers, is there any reason to be skeptical that this is happening?

Yeah, I think with polls, there's always a good reason to be skeptical. These are imprecise measurements that do very well, all things considered, but there's margin of error. And as people know, that's only the margin of sampling error. There are lots of other errors that go into these polls. Having said that,

These are also things that we've been tracking over time and lots of pollsters have been tracking over time. And to my earlier point, we also have a lot more high quality data digging into these subgroups. A lot more people are doing oversamples. A lot more people are looking regionally at some of these interesting groups. And so there's reason to believe that this is actually happening, not just because sometimes these narratives kind of

spin up themselves and there's confirmation bias. I think this is one where we actually have a sort of increasing volume of data that backs it up. And of course, there's always reason to be skeptical of any of these pre-election narratives because, you know. People got to vote. Right. But I think in this case, there's a lot of reason to believe that it is to some extent happening.

Exactly what Ruth said. I personally find the theory plausible, but I'm also not sweating it too much. Like, I'm not thinking about it too much because it's not really going to change my fundamental understanding of the race. Our colleague Mary Gradcliffe did a story a week or two ago and found that if you assume that there would be an environment where there is this partial racial depolarization and one where there isn't, it only makes a difference in Nevada.

There is such a thing as reading too much into this information. I think it's really interesting if a racial depolarization is happening. But I think that's the kind of thing that we can wait until after the election because then we will know and then we can do all sorts of analyses about what does this mean for kind of American politics in general. But I'm not sure we need to be spending too much time right now thinking or debating about it.

Yeah. And I will say, I mean, one of the points we make over and over, it's that this is very gradual. Like what we're seeing in the data is very gradual. And as you said, it's unlikely to have major electoral implications right now. I think there's a broader kind of interesting story about the Democratic Party sort of relying more on white voters and less on this diverse coalition that they sort of imagined would be the demographic destiny. But

It's a slow moving situation. We don't fully understand it yet. We have these kind of interesting inklings so far. And yeah, it's not necessarily like fully an election story. I think it's interesting in the election context, but I think it's the kind of thing that is worth digging into later. And it's unlikely to have major, major implications on the results.

Right. I've said this before on the podcast, but in some ways it's a shame that once the election happens, that's when everyone kind of starts tuning out to the data because once the election happens, we have so much new data that we can actually answer all of these questions with. For now, we're pretty limited in.

in terms of our tools. You know, we're using the best tools we have, but we still don't have a full data set. We're going to stick with our conversation about the polls and move on to our good or bad use of polling. But first, a break. Did you know there's a nutrition plan that could not only help you lose weight, but also may lower your biological age score? It's called Prolon by El Nutra. And here's the craziest part. It's a fast that includes food.

Introducing Prolon, a revolutionary plant-based nutrition program that nourishes the body while making cells believe they're fasting. Researched and developed for decades at the University of Southern California Longevity Institute and backed by leading U.S. medical centers, Prolon helps promote healthy blood sugar, support cardiovascular health, and reduce abdominal fat.

But Prolon isn't a diet. Prolon is science. Science based on Nobel Prize winning discoveries in medicine. And this all starts with Prolon's five-day program that includes snacks, soups, and beverages. It's unlike anything you've ever experienced. It's no wonder why thousands of doctors now recommend Prolon to support healthy blood sugar and cardiovascular health. Right

Right now, Prolon is offering 538 Politics listeners 15% off their five-day nutrition program. Go to ProlonLife.com slash 538. That's P-R-O-L-O-N Life.com slash 538 for this special offer. These statements have not been evaluated by the Food and Drug Administration. These products are not intended to diagnose, treat, cure, or prevent any disease. See site for details.

Fall is planting season. It's true, many plants actually do better when planted this time of year, but it can be hard to know where to start. That's why over 2 million Americans trust Fast Growing Trees, the world's largest online nursery and the easiest way to get the yard you've always wanted. With more than 10,000 plant varieties, including fruit trees, evergreens, and houseplants, they have everything you need, no matter your yard or experience level.

Their plant experts provide best-in-class customer support and are available for free consultations via phone, chat, or email, giving you the confidence you need to grow. On top of it all, each purchase is backed with a 365-day Alive and Thrive guarantee.

Save 90% on landscaping costs when you do it yourself and get plants delivered directly from fastgrowingtrees.com to your door. Right now, they have the best deals online, up to half off select plants. And as a listener to our show, you get an additional 15% off your first purchase when you use the code FALL15 at checkout. That's FALL15.

That's an additional 15% off at FastGrowingTrees.com. Use the code FALL15 at checkout.

There are plenty of hot button issues this election season, but in our nerdy corner of the world, there's one issue that has come to the fore. And that is the decision that some pollsters have made to wait by recalled vote. That's when a pollster asks a respondent who they voted for in the last election. So in this case, that would be either Trump or Biden. Well, of course, they could have voted third party as well, but not very many people did. Then you wait the results to match the results from the last election.

Historically, this has not been a gold standard practice because some respondents have trouble recalling who they voted for or if they voted at all. And of course, new people join the electorate all the time. Nonetheless, Nate Cohn, Ruth, your colleague at the New York Times, wrote recently that about two-thirds of recent polls used waiting by recalled vote. Among other things, it may be a way to ensure Trump voters are not undercounted again.

So, Nathaniel, let's start with you. Is this a good or bad use of polling? I think it's a bad use of polling. It's just too blunt of an instrument. If you're asking, could it make polls more accurate? Sure. Could it make sure you're kind of

not discounting Trump voters the way that polls have done in 2020 and 2016, sure. But you're basically just making an assumption that the election is going to look like it did in 2020, which is probably true because we have a very stable environment and states are very close and everything is very polarized. But at that point, like, what is even the point of doing polls, right? I could tell you right now that Wisconsin is going to be close.

I could have told you that in 2021 because of the result of the 2020 election. And that's kind of what waiting by recalled vote does is it just forces the states into a range that is pretty close to how they voted in 2020. You're just making assumptions about the underlying electorate didn't change and things like that. I would prefer to see pollsters vote.

stick to their guns and their methods and try to find figure out frankly how to find Trump voters in other ways which granted is a very difficult proposition that people are still struggling with and I very much have sympathy for that but I don't think that you can kind of take the shortcut way out of just rounding to the last election sounds like a bad use of polling from Nathaniel Ruth I agree that it is not a good use of polling

I'm really interested in the fact that it's sort of on the rise among some of these higher quality pollsters. In the past, it was really limited to lower quality pollsters who were kind of trying to slam their data into place. And in this case, it kind of feels like people are sort of fighting the last war in an attempt to get the right number of Trump voters. And the thing I worry about there is you might get the right number of Trump voters, but you get the wrong type of Trump voters. Or I

I mean, as we know, this electorate isn't the same as that electorate. And like you said, Nathaniel, it's probably pretty close because we're in insanely polarized, tight eras. But those little differences make a huge difference.

On the whole, I don't think it's good, but I am really interested in it as sort of a natural experiment right now to see. I mean, we did our own kind of experiment where we waited our polls, previous polls, to recalled vote in the past cycles to see, and it made them less accurate for our time Siena polls, which is really interesting. And I know not everyone has experienced that.

The one case where I think there is like an interesting argument for waiting by recalled vote are panels where people have your contemporaneous recalled vote at the time. Because you can actually check. Because you can actually check. They said at the time how they voted. I still would argue that I wouldn't want to wait on recalled vote there, but I think it's kind of an interesting exception where you could argue that the flaw of recalled vote that comes with people not being able to accurately recall votes

Their vote sort of goes away. But you still have all of the other issues, which is just kind of like slamming your data into place without getting the right number of Trump voters, the right type of Trump voters, to the extent that we're trying to worry about underestimating Trump's support.

The bias that people have is that they'll say that they voted and they'll say that they voted for the winner. And so if you get too many people saying that they voted for Biden in your sample this year, that means you're going to put more emphasis on Trump voters, which is to say that the pollsters who are doing that could actually be underestimating Harris. I don't know that that's happening. No one knows that that's happening. Otherwise, they would fix it. But that's the risk that you run. But

But the other risk that pollsters run into, of course, is coming up with results like Biden being up 16 or 17 points in Wisconsin in 2020. And so I think pollsters are looking at

sort of the way that they, excuse me, shit the bed in 2020 and are saying, reputationally, we can't do that again. We got to get something closer to the actual results. And this, like you said, is just forcing the data into place in that way. Yeah, and they're stuck between a rock and a hard place. And like, I get it. I sympathize with that. I wouldn't want to have a big polling miss like 17 points in Wisconsin again either. But-

I do think that the best practice is to try to innovate and find Trump voters in other ways. And again, I recognize that that is hard and I get how strong the urge must be when you have this lever over here that you can pull that's like, here's a lever that'll just automatically get you within five points of the result. Very probably just because we already knew what Wisconsin was going to be within five points of three years ago.

But I don't think it's a sound practice methodologically, kind of scientifically, if that makes sense. So, Ruth, what are some other methodological ways that we can try to make sure that we're not undercounting Trump voters again?

I mean, that's the problem is like there's no silver bullet. And that's why people are trying these kind of unconventional, typically considered questionable things. I mean, you know, one thing that's happened in general over the last 10 years is that pollsters have really diversified their methods. So some people are trying to

text messages or text that pushes to a web survey. People are trying IVR, which is like an auto dialer that calls, you know, certain types of phones. Like everybody's trying to, and obviously people are trying non-probability polls, right? And everybody's trying all of these different levers to see if some combination of them is going to get, reach these Trump voters that aren't here. And there's, there's no good answer. I think the, probably the best answer

is the hardest for political pollsters to use, which is mail surveys. When you send out mail surveys, you get a higher response rate. We have done experiments on this in the past at Time Siena, and you see that you get these more, or less, I should say, politically engaged, more moderate people who were sort of missing. And those are the kind of people maybe you want. At the same time, those surveys take

weeks sometimes to conduct. And in an environment where, you know, you want a quick read in the polls, it may be the best way to answer some of these questions, but it's not actually realistic with our current political environment.

We also don't know for sure that there is a systematic problem reaching Trump voters because it's a sample size of only two elections. One was a pandemic. Yeah, exactly. And maybe probably different causes for the misses in 2016 versus in 2020. And so I don't think that pollsters should just be lying down and being like, oh, everything is fine. This was a fluke.

I think it's good, and I can attest to having talked to pollsters and having gone to the conference of pollsters. People are thinking about this a lot. It's keeping people up at night. But at the same time, it is possible that the data people are getting back, the data good pollsters are getting back without waiting by recalled vote, is already accurate. And I would like to see that data reflected.

without being kind of polluted by the recalled vote situation and be able to compare that because that will actually tell us a lot as well. Compare that to the final election results, I should say.

I think the reality is one way or another, one group of pollsters will probably be quite off, right? Because right now the averages are taking folks who are waiting by past vote and folks who aren't and averaging them together. And so we're taking two pretty different pictures of the electorate and in some ways smushing them together. And so maybe...

Like the averaging system will work and that will just get us closer to the result, but also maybe one methodology gets us closer than the other. And so one group of pollsters will be off significantly and one set of pollsters will be dead on. And that is also something that we will only be able to answer with time.

And this is what I appreciate what Nate is doing now, which is he's already starting that separation to tell us what those two visions might look like. And even though it's not showing up, we're not doing that in our averages. You're not doing that in your averages as he's writing these stories. It's like kind of a reminder to people and to analysts that, you know, there is, there are these kinds of two paths that we're following that we're interweaving and they really do act separately. And so explaining to those, those differences publicly is a really nice thing.

I would also say – and this is going to sound perhaps bad – but I would say that even if after the election, the weighted by recalled vote polls get closer, say they nail the result, right?

I'm still not sure that validates the method, right? Because if that's the case, then it'll probably be the case that the election was very close to the 2020 results, which is just like a thing that might happen in a very polarized society. And I worry about if everybody suddenly starts waiting by recalled vote because like, oh, this was the solution and it's great.

Four, eight, 12 years down the line, when we do start to see a significant shift in the electoral coalition and everybody's waiting by recalled vote, pollsters are going to get it wrong again because, oh, we didn't predict that Florida was going to go blue this year because of changing patterns or whatever, because everybody's just assuming that the election is going to be the same as the election four years prior. So it's just a very tricky problem. And I have all the sympathy in the world for pollsters who are trying to figure it out. But yeah.

For people to think of this in maybe more concrete terms, the wait by recall vote is showing something that looks like 2020. The not wait by recall vote is showing something that looks more like 2022, which on one hand,

It's the closer election. You know, things changed. The pandemic receded. People moved around the country. Closer in terms of time. Closer in terms of time. It's also post-Roe. So maybe it makes sense that things would look more like 2022. But the problem here is that that was a midterm election. And the electorate is so much smaller and so much more engaged politically. And so you do worry that if the...

electorate that we're seeing in the polls looks more like 2022, it's because you're getting too many politically engaged voters because you're not getting the marginal, say, Trump voter or the person who pays very little attention to politics who will end up showing up on election day or the days before. Which are sort of, by definition, the swing voters this time around. They're low engagement voters. And so you're effectively missing them in 2022.

And so the reality is you could tell a coherent narrative about either vision of the electorate right now being accurate. And so it's hard. Let's move on and talk about some recent events before we wrap up. But first, a break.

Fall is planting season. It's true, many plants actually do better when planted this time of year, but it can be hard to know where to start. That's why over 2 million Americans trust Fast Growing Trees, the world's largest online nursery and the easiest way to get the yard you've always wanted. With more than 10,000 plant varieties, including fruit trees, evergreens, and houseplants, they have everything you need, no matter your yard or experience level.

Their plant experts provide best-in-class customer support and are available for free consultations via phone, chat, or email, giving you the confidence you need to grow. On top of it all, each purchase is backed with a 365-day Alive and Thrive guarantee.

Thank you.

That's an additional 15% off at FastGrowingTrees.com. Use the code FALL15 at checkout.

Last week, Hurricane Milton hit Florida's central west coast, leaving behind a path of devastation that claimed the lives of at least 16 people and left over 3 million without power. The Category 3 hurricane is estimated to have caused up to almost $50 billion in property damage. And of course, that came just after the Category 4 Hurricane Helene, which left at least 230 people dead across North Carolina, Georgia, South Carolina, Florida, Tennessee, and Virginia.

The back-to-back storms are a human tragedy and also now a political talking point. And so Nathaniel, as we've discussed, election season and hurricane season do collide. And so we have plenty of examples actually of hurricanes becoming political talking points in the past.

Is it clear how the hurricane response or the hurricanes themselves are shaping voter opinion at the moment? At the moment, no.

Because A, it's early, especially with regard to Milton, which just hit last week. And I don't believe we have any polls of Florida since then, which obviously is relevant. And there's no good way to poll Florida right now, I would imagine. With Helene, we have a little bit more data. But it is still, frankly, kind of early because these things do take time. You'll recall that the response to Hurricane Katrina took a long time. And that was a new story for a very long time for George W. Bush in 2005. But

So far, basically, the country is pretty split, even along partisan lines. Kind of sad that it's a partisan issue, but here we are. So, for example, there was a YouGov CBS News poll that asked how people think the Biden administration is handling the

aftermath of the hurricanes, 51% said approve, 49% said disapprove. You had similar polling from YouGov, The Economist, which found that when you asked about Biden's handling of Helene, it was slightly underwater. When you asked about FEMA's handling, it was slightly above water, but both of them were pretty polarized by party. So we're kind of in a situation where

Democrats and Democratic-leaning independents think that the government is doing a good job responding, and Republicans and Republican-leaning independents think that the government's doing a bad job responding. And of course, I think this is in large part, or at least partially, because of some of the misinformation that Donald Trump is spreading about FEMA being out of money and money going to undocumented immigrants instead of to disaster victims, which is not true.

It has become kind of partisan, which in terms of impacting the election, at least from a persuasion point of view, I think it's basically a wash. And we're not looking at a situation like Sandy in 2012 with Chris Christie and or Katrina in 2005 with Bush, although he wasn't running for reelection. But like there are these past examples, as you alluded to, Galen, of hurricanes either strongly helping or strongly hurting an incumbent government. And it doesn't seem like we're shaping up that way because everybody's just retreating to their partisan corners.

Yeah. And, you know, to your point earlier, to the extent that this is shaping the election narrative, it sounds like the candidates are really trying to tie this into all of the issues that they're already talking about. Right. They're not pivoting to talk about climate change or even necessarily infrastructure or disaster relief kind of thing. Trump is using it to talk about climate.

immigration. And the people who agree with him on that would say the government shouldn't have been spending money on the migrant crisis to begin with, whether or not there was a hurricane. And for Harris, who's saying, you know, this shows that Trump is heartless and he doesn't have sympathy and whatnot, the people who feel that way have felt that way about Trump for a decade. And so it truly is taking a new issue that could maybe have shaped a past election or past elections that kind of

direction that the conversation would go in. We need to talk more about how FEMA is funded or how it disperses those funds or something like that. And it's all just the comfortable talking points for both sides. Exactly. Both of those examples that I gave is how in 2005, of course, Hurricane Katrina hit and the Bush administration was seen to have responded poorly to that. And that really hurt Bush's popularity. He obviously didn't run for reelection again after that. He couldn't have, but it

would not have been pretty, I think. And then in 2012 with Hurricane Sandy, that hit New Jersey. And Chris Christie was really seen to have had a really positive response to that as governor of New Jersey. And he really cruised to reelection in 2013, of course, as a Republican in a blue state, which was pretty remarkable. But both of those elections or non-elections happened in less polarized times where, yeah, I think that people were perhaps more likely to react objectively to the facts on the ground instead of kind of the situation we have here.

One more data point before we wrap up today, which is according to VoteHub, between August 23rd and October 9th, Trump held 39 public events compared to Harris, who held 28. I've talked a little bit about the meaning of campaign events and campaign stops before on this podcast.

Does it shape voters' minds? I mean, we're just seeing so much of the final weeks is what was it? Trump was in Wisconsin four times in eight days. They're both in Pennsylvania today. I mean, campaign stops really do seem to be the driving focus of the final weeks of this campaign. So you might think, hey, Trump having, you know, 10 more campaign stops than Harris is an advantage for him, is it?

No, not really. I wrote an article on this last week, but the studies that have looked into the effect of campaign visits have generally found that they don't actually affect people's votes. So I thought this was interesting, especially in the context of Donald Trump holding a rally in California the other day, and he's holding one in Madison Square Garden in New York City. I think it's really interesting because the

Especially like lacking the debates, right, which would normally be peppering the calendar here in October. Really, those rallies are the only thing to cover in terms of the day-to-day campaign stuff. But I think they really are overrated in terms of actually moving votes. They just don't do it.

Yeah, completely agree with Nathaniel here. I don't think that they move votes at all. There's no literature that suggests that they move votes. One thing that is interesting, though, I think, and you mentioned Trump's rallies in California and New York, is I think that's less of a tool to move voters and more sort of a...

tool between the campaigns, right? Forcing the Harris campaign to think about California, forcing the Harris campaign to think about New York, to think about contesting these areas, which tracks with some of the data that we've seen lately showing New York moving a little bit more red. I think that was sort of

a little bit of a taunt maybe to the Harris campaign, suggesting they might need to contest that area. So I don't think it actually moves votes. If anything, it's sort of a reaction to potentially already moved votes. Well, especially when it comes to the House races, right? I mean, Trump may be saying, you know, I'm helping out down ballot Republicans by hosting rallies in New York and California, because between the two, there are nine competitive House seats all held by Republicans. And

The other thing is that it's the seat of national media. I mean, New York City, you may not be changing people's minds directly, but you certainly are attracting a lot of attention to what you're doing if you're hosting rallies in California and New York's midtown Manhattan. Yeah, it will get people talking and it will get people talking about, are those competitive? We're talking about it now. It worked.

Yeah. And that's I think that's the goal. I don't think that it's necessarily about moving votes because we have no reason to believe that any of these are moving votes. But it's, you know, starting a conversation and a little bit taunting. I think another potential benefit to them is fundraising as well.

There is evidence from the literature that says that campaign visits can increase donations. And of course, there are so many Trump voters in California and so many Trump voters in New York, and maybe they'll open their wallets now and Trump could use the money. However, there is also evidence in the literature that campaign visits have a backlash in terms of donations. So you get more donations, but also your opponent gets more donations. So it could be washed in that regard. But I think that Harris probably has more money than she knows what to do with. And Trump could probably use a little bit more. So maybe that's a net benefit for him.

All right. Well, we are going to leave things there for today. Thank you, Ruth and Nathaniel, for joining me. Thanks, Galen. Thank you. My name is Galen Druk. Our producers are Shane McKeon and Cameron Shortavian, and our intern is Jayla Everett. You can get in touch by emailing us at podcasts at 538.com. You can also, of course, tweet at us with questions or comments. If you're a fan of the show, leave us a rating or a review in the Apple Podcast Store or tell someone about us. Thanks for listening, and we'll see you soon.