A (2:20)

Is it still hot in Pittsburgh?

C (2:21)

Yeah.

A (2:23)

Also here with us is another friend of the podcast, Nathaniel Rakich. Welcome to the pod.

B (2:27)

Howdy, Galen.

A (2:28)

Howdy. Okay, so let's talk about what's happening as we hit record on this podcast. So, from what I understand, I was on vacation. I was in Wisconsin. But the main takeaway from Trump's meeting with Putin is that Trump was particularly congenial towards the president of an adversary nation. As I mentioned, Trump's relationship with Russia was a big flashpoint of his first term, but it seems like at least people aren't. The media isn't focusing as much on it in his second term, so maybe let's bypass all of that. Mary, how is the public reacting to all of this and Trump's relationship with Putin and the like?

C (3:08)

I don't think the public has a lot of confidence in Trump's dealing with Putin in general. So in a Fox survey from July, the pollsters asked who respondents thought had the upper hand in the situation in Ukraine. 58% of voters thought Putin has the upper hand, and 35% thought it was Trump given those two options. Which does not look super good for Donald Trump. Right. Like, if almost 60% of Americans. Well, voters in this case say that Putin has the upper hand in this. I think that shows a lot of distrust, perhaps in Trump's ability to negotiate with Putin and not just take his side. And I think that's what we saw coming out on Friday. It felt like Trump was taking Putin's side. It didn't feel like he had really made any headway on the priorities of the Ukrainians or the priorities of the Europeans that had been discussed prior to the meeting. So maybe voters. I have a pretty good read on this situation.

B (4:12)

Yeah. Just to add to that, there was also a Pew Research poll that came out earlier, just before the summit, and According to that one, 33% of Americans thought that Donald Trump would favor Russia too much, and only 6% thought that he would favor Ukraine too much. 28% did say that he would strike about the right balance, and 32% weren't sure, which, of course, is kind of a constant with all of these foreign policy things. A lot of people aren't paying close attention or just don't feel like they are experts in the field, which I think is very understandable. And so you do have to bear in mind with public opinion on foreign policy that it can be pretty fluid and uncertain.

C (4:49)

Yeah. I would also highlight in that same Pew survey that only 40% said they were confident in Trump's decision making in this matter. 59% said they were not confident. So on that, you have a lot more people with an opinion and it does not seem to favor them. The president.

A (5:06)

Yeah. I mean, this doesn't necessarily have data associated with it, but what was particularly striking about the meeting was just like, they seem to be really good friends. Like, I don't know. I'm not an Expert in US Russia relations. I don't know very much about Putin, I'll just say all of that. But, like, I interact with people a lot. I know what it's like to be around people that I like and people that I don't like and like, they kind of seem like friends. And if this had happened in the first year of Trump's first term, we would be talking about this, I think, for like three weeks or so. I mean, not that any news story lasts three weeks on the Trump era, but it just seems like it was kind of crazy. They hung out in Alaska. They seem to be buddies. You know, Trump kind of deferred to Putin to speak first, and they didn't take questions from the press. And, you know, all of the tidbits that I'm sure people have read about already, kind of seems like we're just moving on.

C (6:01)

While it is an important issue and it matters a lot to the people directly involved, I don't think it's a top of mind issue for American voters. I mean, you're not seeing foreign policy or this war or any particular conflict come up in surveys as things people really want our government to be spending a lot of time focused on. People are more interested in economic issues.

A (6:24)

Yeah. I mean, to that end, Trump is meeting with Zelensky as we're speaking right now. That meeting will have happened by the time that people are listening to this podcast. But how have Americans views of the conflict evolved since Russia invaded Ukraine in 2022, when it might have been more top of mind for people?

C (6:43)

There was a survey that came out from the Chicago Council on Global affairs that was conducted at the end of July after this meeting was announced, before it actually happened, obviously. And you actually see some really interesting stuff in there. So noted this on Twitter. You can watch Republicans feelings about this respond in real time to changes in Trump's attitude. It's incredible. So, for example, at the beginning of the war, pretty much across the board, it's like 80% across the political spectrum support for sending US military aid to Ukraine. And you can see it sort of diminishes a little bit among all political groups, but it remains in the 70s for Democrats. For Republicans, it is sort of a steady decline. And when you get to March, which is right after that infamous meeting with Zelensky in the White House, Republican support is down to 30% for sending US military aid to Ukraine. And then recently, as Trump sort of seems to have been turning around and getting annoyed at Putin for continuing this conflict, it bounced back up to the 50s in the most recent survey. So that I think that's one way you can think about public opinion on this is that it's not deeply held beliefs.

B (7:56)

Right.

C (7:57)

People are responding to their trusted sources of news and information that may include the president pretty easily on this because they don't have, like, a deep, emotional, moral, whatever connection to whether we should send US Military aid to Ukraine. They're following the lead of their political leaders, and you can see that sort of across the board, a slight diminishment in support for Ukraine and US Intervention and things like that among Democrats and among independents and then really wonky lines for Republicans, depending on, like, Trump's position at that moment.

B (8:32)

Yeah, I mean, I definitely think that public opinion on foreign policy is quite malleable and susceptible to kind of elite cues, as you were saying, Mary, and as we've talked about in the past on this podcast about other foreign policy issues, I do wonder if there is kind of like a reversion to the mean thing going on, though. Like, I. So, you know, I have been unemployed since FiveThirtyEight went under back in March, and that has given me the luxury, quite frankly, of like, ingesting the news like a normal person would, which is paying, like, some attention to it on the side, but, like, not like, really mainlining it all the time and paying close attention to everything that Donald Trump says on Russia. I was not aware until I was reading some of the coverage of this summit that Trump had been starting to get annoyed with Putin and had been kind of taking a harsher line. And so I wonder if this is kind of just a small quibble with what you said, with which I largely agree, but I wonder if the movement among Republicans was just basically a reaction to the Zelensky meeting, in which famously, you know, they kind of got into a shouting match almost in the Oval Office and then just kind of reverting to the mean once that episode faded into people's memories, rather than people responding specifically to Trump taking kind of a harder line. You know, it could be a little.

C (9:48)

Bit of both, right?

B (9:49)

Yeah, absolutely. But, yeah, otherwise, I. I think that's absolutely right. It will be interesting to see what comes out of this meeting with Zelensky. You know, I. Again, by the time people are listening to this, maybe there'll be another shouting match or something like that, which to, to Galen's point, I think would contrast quite sharply with the kind of body language and the attitude that Trump took with Putin over the weekend.

A (10:09)

Yeah, this brings up one other thing in Preparing to talk about this, I was wondering sort of where foreign conflicts or foreign policy rank in terms of what's on Americans radar or what's their most important priority. And Gallup does regular polling asking Americans what the most important problem facing the country is. Unsurprisingly, foreign conflicts were the most important problem according to only 2% of Americans. But what stuck out to me actually more than that was the percentage of Americans who said that poor leadership is currently the biggest problem facing the country. So for much of the time that we have referred to this polling in the past, I don't know, probably five years, the economy has basically eclipsed everything else. And in this poll, the economy is still the most important problem, but by literally only 1 percentage point. So 27% of Americans said the economy is the most important problem facing the country. 26% said that it's poor leadership. And so, Mary, you said something like, you know, Americans are more focused on the economy than they are about Russia or Ukraine or whatever. And I think that's totally true. But it also now seems like they're equally concerned about just, I don't know, is that 26% of Americans saying that the biggest problem facing the country is Trump?

C (11:24)

Uh, there are many leaders in Washington, Galen.

A (11:26)

Yeah, yeah, I guess that's true.

C (11:30)

I mean, that's the thing is we don't know what people mean by that. Right. Are they saying the problem is the Senate? Maybe. I don't know.

A (11:37)

Yeah, they could also be talking about their state legislature. The problem is, you know, gerrymandering on the state level.

B (11:43)

Yeah, I don't kind of.

C (11:44)

I don't like my mayor. What? I don't know.

A (11:48)

Sure, right. It's, it's vague. The biggest problem facing the country is Eric Adams. No, I'm kidding.

C (11:53)

There might be people that agree with that.

B (11:55)

I mean, I think that it is like a stand in for obviously, general frustration with the ability or inability of government to respond to people's problems, which can include inflation and immigration and like a lot of different things. I do think a lot of it is probably just kind of Donald Trump and his temperament and, you know, probably some of his kind of more anti Democratic tendencies fall into this bucket too. But I think that, I'm not sure that it's quite an apples to oranges thing to include that in that list of kind of alongside the economy and foreign policy and immigration and health care and all that stuff.

C (12:33)

Methodologically, Nathaniel, I believe this is an open end.

B (12:36)

Oh, really? Okay. Interesting. Hmm. Okay. Though then actually, maybe it Is, I mean, maybe it is, maybe it is Trump then. I mean, I probably, probably, I think it's probably safe to say you're right, it is broad and I guess I'd be curious to see the like, specific answers. But I would imagine that a fair chunk of that is Trump. And I do think that that is interesting because we are kind of in this, like, we've talked about this in general with, you know, like people's calculus on how to vote isn't necessarily going to be like, oh, what's my most important issue? And which party is better on that issue? And that's what I'm going to vote on it. We're in kind of this post issue political landscape where people vote on their general vibes toward a candidate like Galen looks like he disagrees. But like, I mean, I think like.

A (13:23)

Obviously, like, I think it's like a poor Candle and Stowe situation. I think the issues absolutely do matter. But I also think that sort of like, will this person fight for me kind of of, you know.

B (13:32)

Yeah, yeah, no, that's fair. But like, I think basically what I'm trying to say is that we are in this era of like very personality driven politics, especially Donald Trump being kind of like number one on that list. And a lot of people just like their motivating factor to vote I think is just like hatred for Donald Trump or love for Donald Trump. And sure issues absolutely feed into that, but like it has become harder to, to it's been even more so than it is in the past. It is no longer kind of a calculus of preferred issue areas and it is more about the, the attitude toward the person. And that is, I think, a notable change.

A (14:14)

Well, Nathaniel, you are teeing us up perfectly for the next segment because on the topic of politics of personality, let's talk about Gavin Newsom. Well, let's talk about gerrymandering. And in talking about gerrymandering, we will get to Gavin Newsom. So as I mentioned, Texas Democrats were return to the state means that Republican can move forward with their gerrymandering plans in Texas. And we've talked about that map previously on this podcast. But California is also moving forward with their plans. So California's redraw is a bit harder because the state has an independent redistricting commission. But they're trying to work their way around that. So first 2/3 of the state legislature needs to approve the new maps that Gavin Newsom presented last Friday. And the legislature is meeting this week to do that. And then California voters will consider the maps. In a special election on November 4th of this year, the maps would go into effect for 2026, 2028 and 2030. And then the state would return to an independent redistricting commission. And the law also stipulates that the maps would only go into effect if Texas or another Republican state redistricts first. So a little bit convoluted here, but I'm going to start with our baseline expectations. Nathaniel, you have been pretty skeptical about this whole process. When you first even talked about California's potential for retaliation, you were skeptical. Now they are attempting to retaliate. Now that we have the maps and the roadmap is pretty clear which path they're going to try to take to gerrymander. What is the likelihood you think that this becomes law?

B (15:48)

That's a good question. I would say it is a. There's a decent possibility, but I'd still probably bet against it. But, you know. Yeah, I mean, I think that.

A (16:01)

And is that because you don't think that this plan is going to work out and that it will be struck. The referendum won't reach the public. The Californians won't end up voting on this on November 4th or because Californians are opposed to it?

B (16:15)

Yeah, I think it's a combination of. There, there's very, there's real doubt about whether this will pass, this referendum will pass in November. And then there are also the questions of legal challenges and to some extent the logistical challenges of holding elections. So the first things first, I think so the secretary of state has said that basically they need to know by this Friday, August 22, whether there is going to be a special election this November because there was not supposed to be an election. Right. Like California. Some California cities have municipal elections in November of odd years, but it is not a statewide election. And the amount of effort that election officials are going to have to go through in some of these counties that weren't planning on having an election is significant. And it's to be a very tight timetable. And basically election officials have said like you need to tell us by Friday or else this isn't going to happen. So the legislature does have a week to pass this map. It seems like they're probably going to do that. They've already released the map. Right. They, they didn't bother to wait to draw the map until they came into session today on Monday. And they do have the votes necessary to pass it with a two thirds majority, as you said. At that point, there is a question of a legal challenges, I'm sure that both this and the Texas map will be challenged in. There's a lot of uncertainty, I think, about whether the maneuver, particularly in California, is legal. The idea is that they are going to pass this measure, presumably as a constitutional amendment to kind of ensure that it would be. But I'm sure that Republicans will try to take it down via every means they have at their disposal. Same with Democrats in Texas. And then this has to pass at the ballot box, which I think is going to be a hard sell, because I think that a lot of voter, like gerrymandering is not popular. Right. This is one of the few things that everybody across the political spectrum, at least in a vacuum, can agree on. They want. They don't want politicians to pick their voters. They want competitive elections. They want to say and who represents them. And basically Gavin Newsom is asking people to go against that for partisan reasons. And obviously there are a lot of Democrats in California and a lot of people who would probably be responsive to the message that Donald Trump is trying to rig your democracy. Vote against that or vote against being.

A (18:27)

Called the Election Rigging Response act, which, by the way, means that Gavin Newsom will have to go on an ERAS tour E R R A S in order to promote this referendum.

C (18:38)

Just to. To, like, put some numbers on some of what you were saying there, Nathaniel, you. You mentioned people don't like gerrymandering. So in their most recent weekly survey, YouGov and the Economist asked about this. And like, they. People really, really don't like gerrymandering. So the question was, should states be allowed to draw congressional districts in a way that intentionally favors one party over the other? Do you guys want to guess what the margin was on that?

A (19:05)

54 points.

C (19:08)

Nathaniel?

B (19:09)

Yeah. 80, 20.

C (19:11)

It was 69 to 9. 60 point margin.

A (19:15)

Okay. I was only. I was only 6 points.

C (19:17)

Yeah, you were pretty close. You were pretty close. Galen wins by Price is Right rules. Yeah, I mean, but with. With a bunch of people sort of undecided or unsure about the question. If you forced people to give a response, I bet you would get that. 80, 20. Nathaniel, like, people really, really hate this.

B (19:33)

That was a national poll, though, right?

C (19:34)

That was a national poll. And in the same poll, 52% said they would favor an independent commission to draw the lines in their own state. 17% favor the state legislature drawing the lines. So just like blindly, without having any of the context of this, people are inclined to prefer lines drawn by the independent commission rather than the state legislature. And that is like, really, you know, the problem that you Were outlining Nathaniel in California is like, can you convince a bunch of people to set aside what they would generally prefer to have happen, the independent commission in favor of a map drawn by the state legislature that does this thing that by a 60 point margin they say they don't want to do. That's a hard sell, I think. And like the, the Democrats in California might have the, the advantage of it being a low turnout special election which like generally you see the most engaged, like highly educated voters in California that's going to favor Democrats. So I don't know, maybe, maybe that helps. But does it help enough to overcome like a 60 point margin against this sort of thing in a vacuum?

A (20:45)

This has actually been polled in California. So the Politico Citroen Center Possibility Lab survey found that just 36% of respondents in California backed returning congressional redistricting authority to state lawmakers. And by a 2 to 1 margin they preferred keeping the independent line drawing panel. However, I think that there is a problem with this. Well, first of all, this was done before Newsom presented the maps and started his camp, his ERAS tour. And I think that it is true that Americans overwhelmingly prefer an independent redistricting commission. They don't find it reasonable or logical that politicians would draw their own districts. Like we said, this is, this is a pretty overwhelming issue. But the way that this is going to get framed in California is not like should we end the redistrict independent redistricting commission and have politicians draw their own lines, it's going to be framed as Republicans are doing this thing that you don't like in other states. One of the only ways that we can retaliate is if we suspend our independent redistricting commission's maps. And temporarily, temporarily. And also, by the way, this only goes into effect if they do it first. And so I think that it's going to end up being, I don't know if this will win. I mean, it needs a simple majority In California, a 2 to 1 margin if that's the. I don't think that this poll is actually asking the question that Californians are going to get asked or feel like they're going to get asked. So I don't think it starts from like a two to one position. I would be curious for more polling as these maps get presented and as Gavin Newsom's, you know, positioning of the issue sinks in further.

C (22:27)

Yeah, well, good news on that front, Galen. UC Berkeley was in the field this week. They, they're finishing fielding today. So we'll get a California pull out of them in the coming days. And I know another, I know another firm that is about to be fielding in California. So data will be coming.

A (22:43)

I mean, this makes honestly our coverage of the off year election definitely more interesting. I mean, a statewide special election in California on an issue as nerdy as redistricting. Come on. I mean, thanks, Gavin Newsom.

B (22:56)

Well, to your. I mean, you know, you guys, you.

C (22:57)

Should be thanking Greg Abbott.

A (22:59)

Thanks, Greg Abbott.

B (23:01)

I mean, you guys were talking about it being a low turnout special election. I think it's going to be a high turnout special election, at least like high turnout by the, by the standards of.

A (23:08)

Okay, way to be pedantic, Nathaniel.

B (23:10)

Sure. But like, it's going to be a bit like, I think this is going to be a really like, if you thought that like people were nerding out over this Wisconsin Supreme Court election, which had like midterm level turnout, I think this is going to be like, like basically, I mean, this is, basically, this is like a pre election for the midterms because it basically is going to decide five congressional districts right off the bat. Right. Like, so, like you're kind of like, it's a little, little appetizer. But anyway, I do want to, I do want to talk a bit more about the, the point of the polling on this ballot measure and really ballot measures in general. Right. Which is that it is difficult to pull ballot measures because how do you do that? Do you try to summarize the ballot measure and like, how do you do that in a fair way? Do you ask literally just the text of what is going to be on the ballot? Sometimes that could be like, misleading or like confusing or something like that. Like, there's basically no consensus in polling world about how to do this accurately. And you can get really tripped up, as we saw actually in Ohio last cycle, because they had a ballot meas on an independent to institute an independent redistricting commission. But the way that it was framed on the ballot by the Republicans who wrote the ballot summary, it basically made it sound like it was the opposite of what it actually was. Which was it? It made it sound like a kind of like biased commission of like Republicans and stuff like that. And the, a lot of the polls asked about it in kind of the straightforward way of like, do you support an independent rejection commission? And this proposal ended up failing because probably of the way that it was worded on the ballot. And I think you can probably expect some shenanigans, Democrats part as well. So we actually did get an internal poll or There was an internal poll that was shared internally with legislators as they were kind of considering this. And kc, a television station in California, was able to obtain the results of the poll. And so that poll said 52% would vote in favor of the map, 41% would vote against it. But it was again, very kind of biased wording. And the wording was the proposed ballot initiative retains California's Citizens Redistricting Commission and declares state policy support supporting use of fair nonpartisan redistricting commissions nationwide. It also adopts new California congressional districts passed by the legislature for use in congressional elections through 2030, retains California's original congressional district maps. If Texas or other states also retain their original district maps, that's obviously basically as good messaging as possible for Democrats.

A (25:29)

Wait, but Democrats are the ones writing the ballot. So aren't ballot initiatives. So aren't they going to just write a. A very kindly worded. Yes, absolutely.

B (25:39)

But also, Republicans are going to spend tens of millions of dollars saying, this is a power grab by, you know, Good point. California. Yes, absolutely. Hundreds of millions of dollars making the point that, like, Gavin Newsom is trying to rig the election for Democrats. And quite frankly, on the same side.

C (25:58)

Democrats are going to spend hundreds of millions of dollars saying, sure, no, we are trying to redress, like an unfair thing that Texas, like, you're gonna get all that messaging on all sides.

B (26:09)

You absolutely are. But I think, I guess what I'm saying is, and the reason why I, if I had to pick right now, and I think we are operating on very kind of sketchy information, but if I had to pick right now, I would say that it's gonna fail, is that this poll which tested the Democrats message well, like, if people are gonna get both these messages right, and like, there's gonna be a lot to be said about, like, who spends more money and stuff like that. But, like, if you just like, do a strai of the polls, one of these polls, which is like pretty biased toward Democrats with a plus 9 margin for the ballot measure. The other, which is, I would say, mildly biased against Democrats, which has a, what, like a negative, like minus 28 margin for the ballot measure. When people are hearing both these messages, I feel like it's going to come somewhere in the middle. And that middle point is not good for Gavin Newsom.

C (26:58)

Also on ballot measures, people have that, like, I ain't gonna read all that experience all the time. And generally what that means is they are more likely to just say no, right? Like, because they're like, I don't want to Read all that. I don't, I don't. This is too much. Just. No, don't stop.

A (27:19)

Really is a question of is the electorate in California showing up on November 4th a group of people that are, well, one more Democratic, obviously they are, but also more inclined to be following the ins and outs of this off year redistricting story. If people are paying attention enough to be like, well, Greg Abbott did this because Donald Trump said that, and Texas is trying to get five more Republican districts and California is only doing this to try to address that and be fair, then I think it wins. If it's a larger population of people who are not paying attention to redistricting stories to the degree that we are on this podcast, then it probably doesn't win. Because you're like, wait, we already have an independent redistricting commission. Why do we need all of this stuff that's in here? That seems pretty complicated. And if the electorate skews towards the, you know, the type of people who would be following the ins and outs of redistricting, then I think this definitely wins. But I think, I think that's hard to say. If this ends up being a high turnout election, I don't think we're going to have turnout any more close to the Wisconsin Supreme Court race. By the way, Nathaniel, California just doesn't vote as at nearly as high rates as Wisconsin. Wisconsinites vote like their life depends on it.

B (28:33)

And I think we might have midterm, like midterm level turnout, though.

A (28:37)

Okay, all right, we'll see. We'll see.

C (28:39)

Competitive states, competitive states always have higher turnout.

A (28:42)

Yeah.

C (28:42)

Because you can actually, like, do something.

A (28:45)

Yeah. Speaking of all of these shenanigans, do we expect other states to join the fun now? Because it started with Texas and then there were rumblings about Missouri and Ohio, which does have to redistrict anyway. Then also New York maybe was joining the chat. But now that, like this is sort of busted out into the open in this way, it seems like Indiana might also be joining the chat. Like, what's going on here, Nathaniel?

B (29:11)

Ohio was going to redistrict this year anyway. They, they have to because the map that they passed back in 2021 or 2022, whatever it was, was explicitly only supposed to be good for two cycles because it didn't pass with bipartisan support. Now they could theoretically just repass the same map that they passed before. I don't think they're going to do that because again, you know, the Republicans are in control of Ohio and they are probably going to pass a map that is better for them than the one that they got last time. And that was probably going to be true even before this kind of can of worms got opened with Texas and stuff like that. I actually have a question about this because that Californ does say this is only going to take effect if, if a Republican controlled state like redistricts. And what if Texas doesn't? It looks like Texas will almost certainly. But like Ohio is going to. So like is that also, would that also trigger California?

A (29:57)

I'm sure that would trigger it, yeah.

B (29:59)

So that's kind of sneaky if so. But anyway, in addition to that, Missouri seems like they're likely to go for it. They, it seems like they're expecting a special session in September to dismantle the Democratic district around Kansas City. So that would be plus one for Republic. Just Florida.

C (30:16)

Just a piece of news from this morning. I saw that in Missouri, like they've already spent state money contracting with a map drawing company. So like, yeah, they're moving.

B (30:26)

Missouri seems like it's going to happen. Florida has been kind of talking about it, but that seems like they're still kind of in higher level discussions about it. Obviously that's a state with a lot of congressional districts, so that would be a kind of higher impact one. And then Indiana, as you mentioned, Galen, the, the Republican leaders in Indiana seem cool to the idea, but it seems like national Republicans and Donald Trump are pushing them to dismantle the Democratic district in northwest Indiana around Gary. But a lot of these are just kind of like one offs like other than Florida, you've got, it's like one district in Missouri, one district in, in Indiana. In terms of Democrats. Yeah, again, they're just a lot more legal restrictions on what Democrats can do because a lot of those states have these kind of redistricting reforms. And if you do the math in terms of, you know, if every state basically like did what they pushed up to the limits of the law of what they could do, Republicans would be way up in front like Repub. The Republicans have much less holding them back and just more districts that are kind of at their disposal to redraw. So if this kind of accelerates into a full blown redistricting arms race, which I think we're well on the way toward and you know, eventually it'll stop. But who knows where Republicans do have the upper hand in that.

A (31:36)

Yeah. Speaking of that arms race, New York Congressman Mike Lawler wrote an op ed in the New York Post calling for Congress to ban gerrymandering and limit redistricting to once per 10 year cycle. You know, I think that Democrats and good government groups thought that the way to handle this was to pass these sort of redistricting commissions, independent commissions, bipartisan commissions on the state level and sort of expand them through voter referenda and things like that. But it might seem now like the faster way to get movement on this, I don't know because I would still be skeptical of anything happening on the federal level is just like an all out war where everyone gerrymanders as much as possible and then that push pushes federal lawmakers to act because as it currently stands, the Democratic states have limited their ability to gerrymander, but the Republican states have largely not. And so that seems like a great deal for Republicans, the status quo.

C (32:40)

Yeah. Kevin Kiley also mentioned a couple weeks ago it was a Republican out of California that he was introducing legislation that would ban mid decade redistricting when not required by court order.

B (32:52)

He's one of the ones who would be drawn out, drawn out in this California map, by the way.

A (32:56)

Yeah, and Mike Lawler too, right?

C (32:59)

Well, New York, I mean Mike Lawler is in New York. And New York, as much as they might like to play in this little game, is, is limited by their own state constitution where voters passed a constitutional amendment I believe in 2014 that specifically prohibits mid decade redistricting. So Michael Ahler is safe for now.

B (33:21)

He's in one of the three districts that voted for Harris that elected a Republican House member. So he's, he's not safe kind of inherently, but well, his district lines are safe.

C (33:31)

Should I say he could be a lot less safe? No. I mean I think that's a really interesting question though Galen, just in general because like this is one of the things that Democrats tried to pass in their HR1 in H R1, the like elections reform package during the Biden administration that did not go anywhere, I guess. And I think it's interesting now you've got some Republicans that are also on that same position a chance for some bipartisanship. But here's the like fundamental problem. Once all of these seats are completely gerrymandered, right. Like we go through the gerrymandering arms race that you just mentioned. If you don't allow gerrymandering, all the people who hold those seats will have to leave Congress. So what, what motive do they have to vote against it if the only reason that they get to serve is because of gerrymandering? I think that's like, I don't know, seems hard I mean, I do.

B (34:25)

I think that if, if you could force this bill to the House floor right now, I think it would have a majority of votes because I think you'd have enough of these kind of endangered, slash, moderate Republicans, plus most, if not all Democrats who would vote in favor of it. The issue is Donald Trump, right? Even if this bill passed the House, even if this bill passed the Senate somehow getting past the filibuster, Donald Trump would veto it because he's the one who kicked this off by pressuring Texas to redraw its maps. Because he correctly understands that basically the only way that Republicans can keep the House next year is if they draw new maps that strongly favor them in Texas and states like that. And he also probably understands that if this were to be taken to its logical conclusion, to its kind of full blown conclusion, that Republicans would come out and head because of the structures of the redistricting kind of reforms state by state, as we were talking about. So that is the fundamental problem and why I don't think that this is ever going to pass. And, and also, like Mike Johnson wouldn't put it on the floor because he's listening to Donald Trump. I think the real interesting what if is what if Democrats had been less ambitious and decided to put a gerrymandering ban, like a standalone gerrymandering ban, with all the stuff without all the extra election reform stuff about, you know, mail voting on all that stuff. If they had just tried to pass that back in 2021 and maybe that had gone through and maybe then we wouldn't be in this situation.

C (35:47)

Yeah, that would have been a good idea.

B (35:49)

How about that?

A (35:52)

All right, well, I have a feeling that we're going to come back to this, but let's move on to our GDBD Needy for today. Today's podcast is brought to you by you, the listener. This podcast is possible because of paid subscribers. At GDPolitics.com paid subscribers get access to about twice the number of episodes. They can join in our private chat and send in questions that we'll later answer on the show. And most importantly, they ensure that we can continue to try to understand politics and the world with curiosity, rigor and a sense of humor. When you become a paid subscriber, you can also connect your account to wherever you listen to podcasts and never miss an episode. Join the GD Politics podcast community today at GDPolitics.com in fact, go to GDPolitics.com right now and become a subscriber. We will still be here when you get back that's GDPolitics.com and thank you. To kick us off, I'm going to read from an Ipsos press release. It reads, quote, ipsos announces a groundbreaking partnership with Stanford University's politics and social change lab Pascal to pioneer the use of AI and synthetic data in market and public opinion research. The collaboration will center on creating and rigorously validating digital twin panels, virtual representations of real world survey respondents to dramatically enhance the speed, efficiency and security of data collection. IPSOS and Stanford's PASCAL are collaborating to build and test the effectiveness of synthetic respondents, synthetic respondents leveraging Stanford's research on digital twins that will expand its application to marketing while also identifying and mitigating potential risks and limitations. The project is built upon the foundation of IPSOS exclusively knowledge Panel. Renowned for its rigorous methodology and including diverse and hard to reach audiences. The project will first launch in the US Followed by a global expansion. So, Mary, I feel like this is a good task for you. Can you tell me what that means in plain language?

C (37:54)

Sure. So the idea here is to use real survey respondents. So like, Ipsos has the knowledge panel, it's a probability based panel. They have people that have responded to multiple surveys. So they have a bunch of information about these respondents and they're gonna feed that information into a large language model of some sort and then tell the model, I would like you to answer new questions as if you are this person I just told you about. So like, dear robot, please pretend to be Galen Drook and answer questions the way Galen Druke would answer questions. Knowing what you know about Galen Druke.

A (38:31)

That little of a robot, I don't.

C (38:32)

Want to make that robot that way. They can use that instead of conducting traditional polling. They don't have to call Galen to find out what Galen thinks. They can use their Galen robot to find out what Galen thinks.

A (38:46)

Is this good data, bad data or not data?

C (38:51)

I think that depends on what you're using it for.

A (38:55)

Okay, I feel like I get that answer a lot, but go ahead, tell us.

C (38:59)

I want to be really clear. This is not polling day data. This is not going to be, I think, a viable substitute for conducting an actual poll most of the time. I think this feels to me a little bit more like modeling, right? Like you give the robots a bunch of information about the world, all the opinions that people have expressed in the past, and then those robots can like do, like some of times they can do other stuff too, like go create a news feed that might look like Galen's newsfeed based on what you know or Whatever. Like if they have data about what, what what people are consuming in news, they can then have a robot construct a news feed that looks like respondents news information and then try to predict their future opinions in real time based on what the robot thinks that person knows about. That's an interesting idea about modeling the world and it seems to be more or less like okay at doing it. So there's a, a polling startup, polling hard air quotes. If you're listening to this. A polling startup in the 2024 election. I think it's pronoun aaru and they built a forecasting model that basically like the same kind of forecasting model that we had at 5:38 that Nate Silver has that all these different places have a forecast model. But they did it entirely with synthetic respondents. And they were just trying to see like could we do synthetic. Use synthetic data, synthetic poll respondents to do basically the same thing that all the polling based models are doing. And guess what? They had basically the same numbers as all of the major actual polling models. Right. They had the election at 53:47 for Harris. Something close to a coin flip. It doesn't seem to be doing anything better than what polls can do, but there are cases in which it seems to be doing things quite a bit worse that we can talk about. But I'll stop talking now and let Nathaniel get his thoughts.

B (40:50)

I agree. If anything, I'm probably even like harder against it. I would say this is not data. It is, you're not, you're not talking to people. Like the whole point of polling is to talk to people and to ascertain public opinion. And this is not doing that. This is not public opinion. It is just kind of like made up. Like yeah, like it's secondhand stuff and like maybe you can kind of get to it. But like if you send a bunch of journalists out to diners in Iowa and have them interview people for two hours and then ask the journalists who they think are going to win the election. That's what we were doing for a lot of time and it was just very vibes based and just not data. E and I just don't think that that is not to say that to Mary's point, it's not just predicting who's going to win the election. You know there are a lot of uses for polling and you know, polling hard air quotes other than horse race. But yeah, I just, I do not think that you can call this data. It is making, it is, it is extrapolating stuff that may or may not be True.

C (41:51)

Yeah, I agree. I don't think it's. I think it's a mock. That's what I. That's like the word I want to use. Instead of data, I want to say it's a model.

A (41:59)

So I was in Wisconsin over the weekend, speaking of diners and did you.

C (42:05)

Ask people who they're going to vote for?

A (42:07)

No, but I asked people what they thought about this synthetic respondent gambit.

B (42:13)

I'm sure they were real up in arms about it.

A (42:15)

Yeah, honestly, I was doing man on the street interviews about the ipso Stanford collaboration on synthetic panels. You know, I wanted to get a sense of what real America thought. And so I ended up Talking to an MD, PhD, who's a resident at the University of Wisconsin Hospital, who uses machine learning in healthcare setting. I'm kidding. This is an actual friend of mine who does live in Wisconsin, but it was not a man on the street interview. And I was like, okay, what do you think about this stuff? This is not my exactly my purview. And he was like, all right, let me take a look at the study that Stanford did. And a couple things stuck out to him that I'll just copy. First of all, one of the things they did was they did a two hour interview with a respondent, then they fed the entire interview into the large language model, and then they asked the model to complete the general social survey known as GSS. And the bot completed the GSS with 85% accuracy. To how the actual respondent, the real person who did the two hour interview you did, that seems to be framed as success in this research. But like, for our purposes, having 15% rate of error seems to be pretty significant, especially when things like elections are won and lost around the margins these days. And like you said, Nathaniel, if you literally just don't screw the two hour interview. If you told me me, okay, right, we're gonna, we're gonna try to interview a white college educated woman living in a metropolitan area in the United States of America. And then you asked me to respond to the general social survey as that person. You could skip the whole two hour interview and I could probably get it 85% right. You know, so. And on top of that, they also do sort of quote unquote focus groups with these bots where they like, they're doing market research. So I don't know, say they're presenting a new brand of gin and they want to know like, okay, bots, like, what do you think about this new gin? And maybe the stakes aren't as High when we're deciding how to market, you know, a new kind of alcohol, a new celebrity's alcohol of the month. But, like, one of the most valuable things from focus groups is the unexpected. And when you have an unexpected response that tells you something. Ooh, like that gives me an idea for how I might market this. Or that gives me an idea for how I might. I present this campaign to the California public about whether we should change our redistricting process for the next three cycles in this case. First of all, you may not get the unexpected because you're literally training it to act like a, I don't know, stereotype or something like that. But also, if you get something that's unexpected, it's like that. Was that a bot hallucination? Was that based on a real world experience? Well, it can't be based on a real world experience because these bots don't have real world experience. So, you know, I was getting a decent amount of pushback from somebody who's actually in this space now. You know, maybe down the line we'll talk to the folks at Stanford and they can make their own case for why this makes sense. But I just wanted to represent some of the skepticism that I heard from somebody who uses machine learning in, like, hospitals.

C (45:24)

I think it's very telling that you had to think of a random product idea and the first thing you thought of is alcohol.

A (45:32)

Mary, are you trying to model me? Stop it. Stop it.

B (45:37)

It.

C (45:38)

I actually, I want to provide some data from another study that I think actually like, supports some of those points too. Galen. So I found this study from some researchers at Vanderbilt was conducted in 2023. So, like, the models may be a little bit better now than they were in 2023, but what they did was they prompted the LLMs with like, basically the kind of demographic data that you were just saying, Galen. Like, imagine you are a white, like 25 year old woman from a rural community, blah, blah, blah. And then they asked these robots, after they gave them their demographics, to rate different politicians on the feelings thermometer. Like, how do you, how warmly do you feel about this person? And they compared it to the CES data, the Cooperative Election Study. So it was actually like decently okay at re approximating the top line, but if you look under the hood, what? One of the things that they found is that their synthetic respondents were significantly more partisan in their responses than normal respondents. Like actual humans. Right. Much more likely to just go along with the, like, partisan line rather than have their own real opinions. And to that point, Galen, that's the sort of unexpected thing that's important to know, right? Maybe there are a number of Democrats that have positive feelings toward, I don't know, J.D. vance. And that's like a useful piece of information to know that a bunch of Democrats kind of like J.D. vance. I'm not saying that that's true. It's just the first name that came into my head. But that would be like a useful piece of information to know. And if you replace real humans who are idiosyncratic and a little weird with robots that are going to be like, ah, the correct response is my partisan response response all the time, you're gonna miss those nuances that really actually shape quite a lot of how people make decisions.

A (47:38)

I think that's so important.

B (47:40)

Yeah, no, totally. And I just think that like, you can avoid all these problems and like be just a lot more confident in your data just by actually interviewing the people. So why not interview the people? And like, the bigger question to me is like, what problem is this trying to solve? Right? And like polling has problems. We know this, we've talked about this on the podcast. You know, I think, think among those problems are like trying to get people who are maybe more inclined to vote for Donald Trump and are more anti institution and stuff like that. But like, is this gonna do that? It just seems to me like it's a way for them to save money by not contacting, spending money on, contacting people. And like, if you could show me that it's gonna help with the non response bias, then okay, maybe I'm listening. But like right now it's like, it's just, you're just kind of like cutting out the person we're actually trying to study, which are people.

A (48:30)

Well, the other thing is in order for this to work over time, you have to keep going back to the real people because you're twinning a person with a bot. And if something new happens, I mean like Mary said, they can try to approximate how this person would respond to a new news event. Like say we're doing this in the middle of last summer and it's like there's an assassination attempt on Donald Trump. The Republican National Convention happens, Biden drops out of the race. Like all of these things are happening. You could, I guess you could tell the bot that all of these things happened or show them like an ex feed of somebody similar to that person and be like, how would you react to these? But in order to get things like, well, there seems to be softness amongst, you know, like black men in the electorate or there seems to be softness amongst this group or that group or, you know, changes that have not been seen in past polling or whatnot. I think you have to keep going back to the people and you have to find those more idiosyncratic people to represent that. And so you, you still have to keep doing polling. And so all of the challenges for polling are going to get. Get baked into this data or not data or these models anyway. So you aren't actually solving the response problem for underrepresented groups and pools.

B (49:58)

Exactly.

C (49:59)

Nathana mentioned, you know, people are trying to do something that's cheaper. Like, I don't know in the long run how much cheaper this is going to actually be for people to do. Like, because you, like, first of all, you have to have the humans that you train your models on. So, like, you're still gonna have to do that big polling stuff. And also you have to, like, run these models which are actually pretty expensive. And the only reason they seem cheap to us right now is because we've got this like giant AI bubble in San Francisco with like, infusions of capital into these companies that are running at a loss. So, like, it's gonna get more expensive.

A (50:35)

All right. I like the. I like the energy that we have on this podcast right now. I mean, look, look, I would say I'm open minded and I'm curious to see how this goes. Like, this podcast is not gonna stop this from happening. And in fact, I would probably say this should happen. But let's like, be skeptical of it and find out with time how it does. But I guess, final question, say FiveThirty is still up and running and you have to make decisions, Mary, about what information to include in the 2028 averages or forecast model or whatever. And we just start getting all of these, like, synthetic surveys. Is there any chance you're putting them in the. Including them in the polling averages?

C (51:17)

Not right now. Not right now.

A (51:19)

Okay.

C (51:20)

I am still aggregating polls and stay tuned. There may be some news in a couple of weeks on that. But I am still aggregating polls and I do have a methodology statement of what I do and don't include up on substack. And I do not include polls that use AI for anything other than replacing, like, phone calls. Right. Like you can use AI to be like the person on the phone calling rather than have an actual human do it. I will let you do that.

A (51:45)

Mary Will allow it. Okay. All right. So that landed somewhere between. Was it like a. Somewhere between all of them, there were like, maybe some good aspects, there were some not data aspects, and there were some bad data aspects. No, Nathaniel's like, none of this is good.

B (51:59)

I would just say it's not data.

A (52:01)

It's not data.

B (52:02)

And I think Mary agreed that. She said it was a model, which I think she's. Yeah, I think Mary maybe thinks there's a little bit more. More. Not to put words in your mouth, Mary, but it seems like you think that there are more potential applications for it than I do. But it seems like we both agree that it is not data, at least the way that we have come to understand pulling data.

C (52:20)

Yeah, I mean, I think I can appreciate why this might have some applications to the kind of low stakes stuff like Galen was talking about, like selling gin or whatever. Great, go ahead. But I think on political issues, these kinds of things are going to struggle because things change really fast. As Galen was saying, people change their minds about stuff, and I don't know that these models are going to be able to do a good job of capturing those kinds of idiosyncrasies that we see in polling data all the time.

A (52:53)

All right, let's leave it there, Mary. Nathaniel, thank you so much for joining me today.

C (52:58)

Thanks, Galen.

B (52:59)

Thank you, Galen.

A (53:00)

My name is Galen Drooch. Remember to become a subscriber to this podcast@gdpolitics.com and wherever you get your podcasts. Paid subscribers get about twice the number of episodes and access to the videos for the podcast. You can also join our paid subscriber chat and pass along questions for us to discuss on the show. Most importantly, you ensure that we keep this podcast going. Also, be a friend of the podcast and give us a rating wherever you listen to your pods. Maybe even tell a friend about us. Thanks for listening and we will see you soon.