Not Another Politics Podcast – Episode 42

You’ve probably seen a lot of surveys recently about how many Republicans believe the 2020 election was stolen, or that they support the January 6th insurrection on Capitol Hill, or that they don’t trust the vaccine. Do these responses predict their behavior in the real world? Or are they just partisan cheerleading?

Northwestern Political Scientist Mary McGrath looks into this question in her paper “Economic Behavior and The Partisan Perceptual Screen.” By combing through data about survey responses and spending patterns before and after presidential elections, she investigates whether partisans truly believe it when they say the economy is getting better when one of their own occupies the White House. If partisans do believe what they say, shouldn’t their financial decisions change accordingly? And if these decisions don’t change, what does that mean for how we should think about survey responses in general?

Listen on Apple Podcasts or wherever you enjoy podcasts.

Transcript

Wioletta Dziuda: 

I'm Wioletta Dziuda.

Will Howell: 

I'm Will Howell.

Anthony Fowler: 

And I'm Anthony Fowler. And this is Not Another Politics Podcast.

Wioletta Dziuda: 

One crazy thing that's been happening in the news for the last nine months or more is the polls.

Reporter: 

According to the latest Reuters/Ipsos poll, 60% of Republicans believe the election was stolen from Donald Trump.

Wioletta Dziuda: 

And every time I hear that, I pause and I wonder, "Is that really true? Do people really believe that? Is that what they mean when they say that they believe that? What's going on?"

Will Howell: 

There are different stories that one can tell. One is that they really do believe it and they believe it because they've been told that it's been stolen. Alternatively, it may just be about partisan cheerleading. And on our show, we've looked at a number of papers and thought a lot about what we learn from survey results. What are we actually measuring? One thing that one might want to do in thinking about what the value is, or the interpreted value is of a survey result is, to see how what people express in surveys correlates with how they behave outside of surveys. And Anthony, you talked to somebody who had explored precisely that topic.

Anthony Fowler: 

I did. I spoke with Mary McGrath, who wrote a paper on perceptions of the economy. There is a very well-known finding in political science that people change their perceptions of the economy depending on who's president. If you're a Democrat and Joe Biden gets elected, all of a sudden—you used to say, “the economy is terrible. I don't like Donald Trump, and the economy is doing badly.” Joe Biden gets elected, and all of a sudden you say, “the economy is doing great!”So, the paper is very much a response to a previous paper. There was a previous paper by Alan Gerber and Greg Huber that is asking this question of: “Okay. We know that people give survey responses. We know that people say the economy is doing better when their preferred president gets elected, do they actually behave differently?” So, they measure consumption. Their finding was that when your preferred party wins the presidential election, you all of a sudden you spend more money. And the implication was, people really do believe this. It's not just a survey artifact. It's that people actually go out and spend more money.

And Mary, essentially, just reexamined, reassessed that study and said that's a pretty striking finding to find that people dramatically change their consumption behavior just because of who wins the presidential election. Let's see if it's true. Let's try to push on it just a little bit more and see if we think... Do we really think that's a reliable result or not?

The paper is about how partisanship and elections might relate to spending behavior and economic behavior. Why is that something that you and other political scientists might look at to begin with? Why might we expect that partisanship would potentially relate to economic behavior?

Mary McGrath: 

Yeah. So, when we look at how partisans respond in surveys to questions about the economy, for example, they report very different perceptions of the state of economy. So, factual questions about the state of economy. You'll see Democrats and Republicans responding in very different ways depending on which party is in power, which party is in the White House. The question is, is this just something that people are saying in surveys or do people really believe that? Does partisanship affect not just how you're saying you think the economy is doing, but your economic behavior? How you're spending.

Anthony Fowler: 

Okay, interesting. So, my preferred party wins the presidential election. I say in surveys that I think the economy is doing better. And now you're asking, do I put my money where my mouth is? If I think the economy is really doing better, if I think that my wages are going to be higher in the future or just the economy is generally going to be more stable, I might go spend more. I might go buy a new TV or a new phone, in my case. That's the idea. That's the idea that if you really believe the economy is doing better, you should spend more. And so, we should actually see people putting their money where their mouth is.

Mary McGrath: 

Yes. That's the idea. There's another hypothesis that's brought in, which is that people kind of take the future into account when they're making economic decisions. And they're trying to look forward... Well, they're trying to avoid these big jumps in consumption by basically trying to smooth out their consumption over time depending on what they anticipate earning in the future. And so, when you're taking this hypothesis into it, when you have partisans who are... You see an election coming up and they're faced with this uncertainty about the future. Does the future hold this rosy economic landscape, a booming economy if their party wins, or do they see this grim economic outlook as would be the case if the other party wins? And so then the hypothesis is when they're trying to smooth out consumption, they basically are sort of hedging before the election. Trying to take this uncertainty into account.

And then when that uncertainty resolves after the election, the election is decided. If that's what partisans are doing, then we should see consumption sort of bounce in these two opposite directions, basically, to recalibrate for that correction they were doing before the fact for the uncertainty. We see the winning partisans should be increasing to account for that sort of hedging low that they were doing. And we should see losing partisans decreasing to account for that. Anticipating the possibility that they might win.

Anthony Fowler: 

So, much of your paper is actually a reassessment of a previous study by Gerber and Huber. Can you briefly summarize for us? What did Gerber and Huber do in their study? And then how did you go about reassessing their result?

Mary McGrath: 

So, they collected county-level data on taxable sales. So, they had it 1992 through 2004. And I was able to get two additional elections of that same type of data. And I was also able to benefit from the fact that doing it later, there were more states that had that kind of data available online in an accessible format. So, I was able to add to their data a little bit. I looked at each election year individually when we have partisans who were hedging their bets before the election. So, if you had no electoral uncertainty, if people just knew the outcome somehow in advance, well, we wouldn't expect any change because they'd just be doing that consumption smoothing. And so, one of the things I was look at, "Okay. Well, let's take that electoral uncertainty into account." We should see the biggest bump where the election is most uncertain where people are hedging the most. And we should see the smallest change in spending where it's most certain. And one of the things that I noticed when I looked by election is that the effect that they saw in aggregate was really driven by this 1996 election. And when you then look at the electoral context of those four, that was the one with the highest certainty about what the outcome was going to be. So, that's actually where you would expect the smallest change in spending after the election.

And then also adding in those two other elections—you're just getting more data, getting more information and looking at those... Similarly, you sort of see that the evidence is no longer really suggesting that people are, in a meaningful way, at least as far as we can see from this data, changing how they're spending after an election.

Anthony Fowler: 

Okay, interesting. So, that's two really interesting pieces of new evidence that you offer over and above the previous study, right? So, one is this theoretical idea, which is, "Okay, if this is real, if elections do affect spending, they should especially affect spending in elections where we were uncertain about the outcome." And if anything, it seems to go the other way. It seems like you get this big result in '96 where the result was maybe not as uncertain, and you don't get the big result in 2000, for example, or 1992, where it was not clear at all who was going to win the election. So, that's interesting. And then the other piece of new evidence is just collecting more data. It's just, "Let's look at new elections, new data that's available since they published their study." And you also don't find anything there. So, what's your overall interpretation: is it that the original Gerber-Huber study was a good idea, and it just happened to be wrong? It just got an unlucky result? Is that the idea?

Mary McGrath: 

Yeah, basically. As I said, I think it was a really brilliant idea. I just had the benefit of time. I was adding more data. It's not like we have a clear behavioral evidence, this clear smoking gun showing that partisanship distorts your perception of reality to the extent that it's changing your behavior.

Anthony Fowler: 

Okay. So, the overall picture here... So, Gerber-Huber wrote this paper. The implication is that voters are... and Americans are so partisan that it actually affects their spending behavior. Their preferred candidate wins the election and they go out and they spend more money. And it looks like that result was a statistical fluke. It looks like maybe people are not as partisan as we thought from that study. We've had Greg Huber on our show before, and we actually talked about one of his papers that is relevant for this discussion. We talked about his work on factual beliefs of partisans. In light of all of this evidence that we have, what is your takeaway when people say something in a survey? When partisans say that they think... When Democrats say they think the economy is doing better just because Joe Biden took over, should we not believe them? Is it just partisan cheerleading, for example? That's one possible interpretation.

Mary McGrath: 

Yeah. Does this mean that respondents are lying when they say the economy is doing better? Some people are probably basically giving an answer to a different question. You have this survey which is this very sort of contrived way of communicating with people and people have complex beliefs. And so, when you ask them, "How's the economy?" They're basically taking the question as “How is the economy and what do you think that says about the people who are running the country?” We're not giving them the option to say, "Well, the economy is okay, but Biden is a terrible president or Trump is a terrible president." And to whatever extent, if the economy is doing okay, it has nothing to do with that terrible leader who is in office.

Anthony Fowler: 

That sounds like partisan cheerleading. I think that would be okay calling that partisan cheerleading. Maybe lying is not the right word, but I'm slightly misrepresenting my views because I want to give an answer that's favorable for my preferred party or for my preferred candidate. Not because I'm trying to faithfully answer the question as accurately.

Mary McGrath: 

Yes, exactly. Yes. So, I’d say that.

Anthony Fowler: 

So to follow, do we have any direct evidence? I mean, it seems like an interesting thing you could test. You could sort of change the way you ask the questions. Give people an opportunity to express just how much they love Joe Biden before you ask them about their perceptions of the economy. Does that change how they answered the economic questions?

Mary McGrath: 

Yeah. And I think Greg Huber has a paper looking just exactly at this and looking at... If you give people those outlets to express their partisan cheerleading, basically, does that change how they're answering these factual questions? And I believe they find that it does. Just giving that outlet. There are a number of different ways people have tried looking at this. Giving them incentives through right answers. Just sort of trying to encourage like, "This is for the sake of a scientific study. Please give the most objective answer you can." And with a number of different methods, we do see that partisan cheerleading reduced. But I believe even when you do that, I think... on a lot of questions you still see this residual difference. I mean, it's interesting to think about what that represents, because you're still dealing with a survey context and you're dealing with people who are not necessarily trying to put a whole lot of thought into what is exactly the right answer.

And so, there's something a little bit... I would call different from partisan cheerleading where they're just really trying to make their best guess possible with very limited effort at what is the right answer. And they're thinking... Okay. If you ask them a question about the economy, well, I don't know the answer. What's a good clue? Think about who's in charge of the country and does that person seem like a competent person or like an incompetent person. So, those people, there’s maybe an extent to which people are trying to get to the right answer. They're using cues that are available to them, and it would be leading to that same result where it looks like simply partisan cheerleading, but it's really a little something different.

Anthony Fowler: 

So, do you want to tell us more about what you've been working on in the last five-plus years since you wrote this paper?

Mary McGrath: 

Yeah. So, one of the reasons that I found this question so initially interesting: does what people say in surveys line up with their behavior in the real world? I wanted to try to understand what was going on with partisanship and beliefs about climate change. That's obviously another area where you see this as a very important issue. You have partisan responses to questions of fact, very divergent. So, you have Republicans who are responding to questions of fact about climate change in a way that is clearly colored by their partisanship, but when thinking about, “How did these beliefs manifest in behavior?”Well, you also have Democrats who are... I was looking just recently. 50% roughly of Democrats say that climate change is basically a top personal concern, but then when you try to identify behavior, even just asking them what behaviors they've done, what actions they've taken, people are not really taking actions that would reflect that belief. And so, this ties back to the economic question because, well, does the fact that people aren't acting on these beliefs, does that really tell us anything about whether those beliefs are true beliefs or not? And is that even the question that we're interested in or are we much more interested in, how does belief manifest in action?

Anthony Fowler: 

So, that sounds like an interesting thing. I'm sure I'm going to discuss with Will and Wioletta. Where do you come out on that debate? Are you saying maybe we should just not even care what people's true beliefs are because what really matters is what they do anyway?

Mary McGrath: 

Trying to discern what belief is matters to the extent that it gives us a picture about whether or not people are learning. And as long as people are learning, then I think there's hope. But on the other hand, I think that just drawing the conclusion from people's actions that you can tell something about what they truly believe is going a step too far. And I saw in a completely separate domain, another interesting example of this. There was this study on health insurance data and looking at birthdays during COVID. In the two weeks after a household birthday, COVID infections went out by something like 30%. And one of the authors of the study wrote an opinion piece in the New York Times saying, "What's interesting about this is there's no partisan divide on that. In Democratic leaning counties, we see the same birthday party effect as we see in Republican leaning counties." There's obviously been a very strong political polarization of questions about COVID and behavior around COVID. Does that behavior tell us something about those beliefs or do we just draw from that that people are fallible and their actions don't always line up with their beliefs? They make mistakes. They make bad decisions. Connecting belief and action is a difficult thing to do.

Anthony Fowler: 

No. Okay. I mean, I see your point, I think, but I think I probably would come on the other side on that. I think my own thought is if you say in a survey that you're really worried about COVID. You wear a mask when people are watching you, but in your own private life, you still have a birthday party hoping that... So long as the wrong people don't find out about it, that suggests that maybe you're not really so worried about COVID when the rubber meets the road and it's time to make an important decision where you'd have to give up something that you care about, you'd still rather have the birthday party. So, you must not be worried enough about COVID. That that would be my own interpretation of that. So, I wonder if you can kind of translate your findings from this economic literature to lots of other domains like reactions to COVID and people's views on climate change and so forth.

Mary McGrath: 

Yeah, I think so. And I think something that you said really hits a part of it. It's sort of... The question is, are you willing to change your behavior in a way that represents a sacrifice of some sort? And I think the answer to when people are able or willing to do that, I can be kind of cynical, but I think it really comes down to whether they think it's going to affect them. And so, that applies to beliefs about climate change. I think that people can truly believe that climate change is a major threat, but aren't necessarily going to make changes in their life that would be a sacrifice if they don't believe that to an extent that they think it would affect their own lives. That they think climate change is going to appreciably affect their own lives. And I think you see that with COVID too. I think people can truly believe that it's a major, obviously, tremendous public health crisis and that people's behaviors need to change, but that... Except if you think it's going to affect you in a very tangible way, likely to affect you in a tangible way, people are great justifiers and they always are able to find a reason to not make that sacrifice.

Anthony Fowler: 

Okay. Great. This is really interesting, Mary.

Mary McGrath: 

Well, thank you. Thank you again for the invitation. As I said, it's very flattering that you guys want to talk about it. So, I appreciate it.

Wioletta Dziuda: 

So, to me, it seems like if you have a result that's a true result and you collect more data then the results should not weaken, should not go away, but the fact that adding just very few elections makes the result basically go away. That's a pretty convincing piece of evidence to me and the fact that everything is driven by one election, and even if that one election was the most surprising one, even if we disagreed on that, I feel like if the entire result is driven by one election, that's very suspicious. So, I think I'm inclined to believe that people don't follow on what they report in the surveys.

Will Howell: 

Yeah. I am too. What she did seems eminently reasonable. I mean, it isn't just that the positive finding holds for one election, it really is driven by one state in one election. It's really about what's going on in Texas in '96. You take that out and basically everything goes to zero or close to it. And so, what seemed like a really striking finding is striking in the sense that wow... When views about how an election turned out correlate with not just my generic outlook about the economy, but my own spending patterns. Okay. So, that finding doesn't hold, but what are we then to think about the status of survey responses, the status of beliefs, the nature of the relationship between beliefs and actions? Which is why I wonder like, what do you think is at stake in the analysis? What's at stake in our understanding of the meaning of survey results, in the status of beliefs? Are beliefs only real, only meaningful, if we can tie them to an observed behavior or beliefs have integrity unto themselves?

Wioletta Dziuda: 

Yeah. Before I can even answer your second part of the question, which I think is very interesting. Suppose I don't act on my beliefs, do only my acts matter or do my beliefs matter too? I think that's a very interesting question and let's go back to that, but before I can answer that, I think the first part of your question is interesting. So, what's going on? What are the possible interpretations of what we see? Before I even thought about this paper, my first interpretation of why do we see that Republicans think the economy is doing better when a Republican president is in power than when a Democrat is in power? I thought perhaps they do truly experience different realities. Republicans tend to be wealthier on average and to be more rural on average and so on. Of course, they might experience different realities than people who live in big cities. But now Mary comes along and she says, "Well, if they experience different realities, then no matter what they believe, they should be spending differently." She doesn't see that. So, I think that's an explanation I discarded. The second one is that they truly have wrong beliefs.

Will Howell: 

But if somebody... Wait, wait. You discarded it just by virtue of the fact that she doesn't observe a correlation. We shouldn't believe there's a correlation between what you say in the survey and consumption patterns, but you can imagine all kinds of other kinds of economic activities that aren't measured like investment patterns, that we don't know. We could still hold out that there, there's a big effect. Those studies haven't been done yet.

Wioletta Dziuda: 

So, let me correct myself. I did not discard it. Okay. Because I don't think we can discard any of the explanations as you say. Maybe she's looking at the wrong activity and so on, but at least looking at her results, I think that the explanation that people report different economies is not because they experienced different economies. They still might, but I think this explanation is less likely than what I thought before.

Will Howell: 

Because one thing you might look for is a correlation between your views and consumption patterns. We don't observe that. That doesn't mean that there aren't other things that might hold true, but having observed this, we should downweigh our... I get that, but in terms of outcomes, this strikes me as the least likely economic correlate of the views you express in the context of a survey.

Wioletta Dziuda: 

Why?

Will Howell: 

Well, if I am a wealthy Republican and I see that a Republican has won the White House and I therefore think my world is getting better, and I generalize that if my world's getting better, it must be better for everybody else, I'm wealthy. Am I going to go out and suddenly spend a lot more? I mean, maybe. I'm already wealthy. I'm already spending a lot. What I'm more likely to do is adjust investment or the are other kinds of economic activities that we could point to that would be much more meaningful because I anticipate changes in the tax code or changes in my ability to hand off money to my offspring. And what is that going to mean for where I keep my money? I just think there are all kinds of economic indicators and do I end up going out to dinner more often? Maybe. I don't know, but these other ones strike me as very real possibilities.

Anthony Fowler: 

And, I guess, so you're imagining kind of the... There's people at the corner solution, basically. There's people at the extremes already. There's some people who have already maxed out their spending. They have so much money. They can't even think of what to do with it. They already have the biggest yacht they will ever want. They don't need a bigger yacht. I mean, probably there are. As far as I can tell, it seems like rich people always want to find ways to spend even more money to make more money. So, I'm not sure how true that is.

And then on the other end, you can imagine people on the other end where they already are spending as little money as possible because they just don't have enough. But I think, in general, the idea that if you really believe the economy is doing better, we should see it reflected in your spending behavior seems like a totally reasonable one. If you really think that your future earnings are going to be greater, then you should spend more money and you should behave differently in some way that should be reflected in your consumption behavior. Even if it is just on the margin. If it's just, "I'm going to buy a slightly nicer car or a slightly bigger house or go out to dinner a little bit more often," or whatever it is.

Wioletta Dziuda: 

Yeah. And I sympathize with Will. As a person who does theory and doesn't do data, I always feel like, "How could you find anything in the data?" But there is research that looked at this question outside of the surveys like, "Is it true that when the economy is doing worse, people adjust their behavior?" And this research seems to suggest yes. They don't adjust their behavior completely rationally and perfectly in line with our theories and predictions, but they do seem to react. So, when there was the housing crisis in 2007, 2009, places where the prices of houses dropped more, spent more.

So, there's evidence that in the aggregate, we can pick up how the economy is doing by looking at the spending patterns of people. Yes, you might be right that investments respond more and so on, but I think the fact that she didn't pick up anything makes me think that this is not the likely explanation. That there's something fishy going on with those beliefs that people had or expect.

Will Howell: 

Yes, and that's, I think that's what this gets down to. It's that what they're saying about in the context of these surveys is less meaningful. We’re measuring something, but we're not measuring people's actual views of the economy, right? That's the suggestion that comes out of this paper, because that we don't observe this behavioral correlate of an expressed survey response suggests that the response itself is indicative of something that isn't a real belief. I guess that gives me... This is a way of kind of validating survey responses, generally. It gives me some pause. Maybe I'm abstracting too far away from the paper, but I can think of lots of things that... And I don't think we want to... Does it then reduce strictly to partisan cheerleading? That that's what it is because we didn't see a behavioral correlate? I don't know. In the same way that if you ask people in surveys, are you worried about climate change? And then you get data on whether or not they actually recycled individually and you show that there's no correspondence, are we then to conclude that what they say in the context of a survey about their concerns about climate change don't mean anything? Because, well, of course, if you really did think that, you would, in fact, go and recycle.

Anthony Fowler: 

There is a difference. There's an important difference between those two examples. So, in the recycling example, there's the collective action problem. And there's the fact that my recycling behavior has almost no effect on the long-term climate change problem. And so, it might be totally reasonable to say, "I'm worried about climate change," but it's not like I'm going to stop driving my car because I know that me stopping driving my car is going to have basically no effect. What I really want is massive regulation that gets everybody to drive their car a little bit less whereas in this example, there's a difference, which is, if I really believe that the economy is doing better, that should have an effect on me directly in some very tangible way. And that should free me up to spend more money if I want to.

Wioletta Dziuda: 

Of course, there's a possibility that people think, "Well, I'm doing fine. So, I'm not going to adjust my spending pattern after Donald Trump is elected, but I believe no one else is doing fine." There might be this sort of strange cognitive dissonance that' happening, that I know my reality and I'm responding to the real state of the economy for me because I experience it. So, I'm not fooling myself about it, but I think other people are hurting. It's possible. I don't know whether that would be partisan cheerleading. I don't know how we would call it, but I can see how this will happen.

Anthony Fowler: 

Okay. So, you know what I see. So, everyone is sitting around. So, they have wrong beliefs in this funny way where they know that their personal economic situation hasn't changed, but they believe that their preferred president's policies are going to just help other people without necessarily affecting them. That's the thought. That seems possible.

Anthony Fowler: 

If you're interested in hearing more or contrasting views on today's top political issues, be sure to check out Politics & Polls with Princeton professors Julian Zelizer and Sam Wang. The show features some of the best minds in politics, journalism, academia, and even the arts. The result is a respectful, intellectual, and oftentimes fun, debate and discussion on the history of politics and current events. The show is part of the podcast enterprise of the Princeton School of Public and International Affairs. Listen to the latest episode at politicspolls.princeton.edu.

Wioletta Dziuda: 

If you're getting a lot out of the important research shared on the show, there's another University of Chicago podcast network show you should check out. It's called Entitled. International lawyers, Claudia Flores and Tom Ginsburg, have traveled the world getting into the weeds of global human rights debates. On Entitled, they use that expertise to explore the stories and thorny questions around why rights matter, and what's the matter of rights. Subscribed to Entitled, part of the award-winning University of Chicago podcast network.

Will Howell: 

I completely agree that we should be really careful about how we interpret these survey responses. And it may be that they're answering a different question than the question they're actually being posed. That could be true. The thing again, though, that gives me pause is, is it going to generalize from this paper? To assess the validity or the integrity or the meaningfulness of people's express beliefs in surveys by looking for a behavioral correlate? Let me give you a different example. Can I try a different one?

Okay. I see your point on the recycling one. And then I thought, "Well, what about if it's... I asked you, do you believe in a Christian God?” And then I go out and I get data on whether or not you behave in ways that are consistent with the Bible, right? And I find out that there actually isn't a relationship between them. And then on the basis of that, should I then conclude that when you come out and say you believe in a Christian God, you really don't believe in a Christian God, or you're saying something altogether differently. I don't know. That gives me pause. I think there are lots of other things you could say about why you don't observe that behavioral correlate, but that there still is meaning in your response, even in the context of a survey that's worth grappling with, that's associated with whether or not you believe in a Christian God.

Anthony Fowler: 

I love that we're talking about religion. We already were talking about—we've focused on politics. That's our show. So, we're already as polarizing as can be. And now that we're bringing religion into it, we'll just—

Will Howell: 

Let's go all the way.

Anthony Fowler: 

So, give me an example of a kind of behavior that you think might be... This hypothetical research you have in mind, what's the behavior they would look at? Because lots of behaviors that are consistent with Christianity, you might say, like treating people the way you'd want to be treated, that's also consistent with non-Christianity. That's consistent with just being a nice person, even if you don't believe in God.

Will Howell: 

No, no, no. So, that's true. That's true, but if at the margin you say you believe in Christian God, and it is a sin to commit adultery, then we ought to see some correlate.

Anthony Fowler: 

And you think if you don't believe in God, then you would have no reason to not commit adultery? It's just a free for all? There's no morality or no contract with your spouse or anything? Okay.

Will Howell: 

No. Obviously, I don't believe that. I don't believe that, but I think what you... I mean, this is part of the reason why this exercise I find problematic. I mean, as a way of validating beliefs. The researcher I have in mind who would undertake this exercise, would go in and say something that is identified as a sin, and professes to believe in a book that recognizes a particular behavior as a sin should at the margin, be significantly less likely to engage in that behavior. And so, the way to assess whether or not they take that seriously is to see whether or not they—

Anthony Fowler: 

There certainly are people who do all kinds of quite costly things because of their religious beliefs. You could find lots of examples and there are things that non-religious people do not do. Maybe we don't want to... We don’t wanna go into the details of all of them, but there are people who...

Wioletta Dziuda: 

Let's list them!

Anthony Fowler: 

There are people who, for example, mutilate their genitals because of their religious beliefs. If you pick the most extreme one, but there's all kinds of... So, when I look at some of these examples, I'm like, "Yeah. That's someone who probably... They really do believe it." So, I think this kind of analysis works in the religious setting too. I mean, I'm sure there are lots of people who just... They go to church because it's kind of the... That's a social thing and whatever, but when push comes to shove, they're not actually going to sacrifice their firstborn son when God instructs them to do so, but there are some people who do. And I think that is a way to distinguish who really believes in this stuff versus who's— who’s to say you do?

Will Howell: 

Not where we’re in the context of surveys, right? Where you get interrupted and you're talking to a stranger, but when you're talking to somebody, right. Here we are the three of us and Wioletta comes out and says, "I want to tell you something. I really feel bad about the puppies. The puppies all around the country, many of them are abandoned, right?" And we say, "Well, Wioletta, what are you doing about that?" right? And if she says, "Well, I'm not really doing anything about it." Are we then we going to say say “Well you don’t really believe—”

Anthony Fowler: 

Then that’s a recycling example. That's very much like your recycling example again, right? You've got the collective-action problem—

Will Howell: 

That's true. But if you really believe it, you would give. You would give. You would say what it means to strongly believe in something is to take some kind of action, no? You would expect to see something. And if in the absence of seeing any kind of activity whatsoever, if only adopting more pets or something like that, right?

Wioletta Dziuda: 

What you're saying is there's a belief and there's an optimal action given the belief. If we see then this optimal action, then that's supportive of the belief. If we see that the action is not optimal given the belief, someone is disclosing, then we should question the belief. So, there's some examples as Anthony mentioned where I can believe that global warming is real and that puppies are suffering, and I might cry about this every night, but the optimal action is not necessarily to go and adopt a puppy or start recycling. The optimal action might be actually do nothing. Just vote for the politicians who promise policies that are actually going to resolve the problem. To bring this back to the paper, I think the question you're asking is, what's the optimal action? Is the optimal action for me to increase my spending and for you to decrease your spending? And you sort of question, maybe not. And I say, "I don't know." The economic theory would say it is. The optimal action for me to increase my spending for you to decrease your spending.

And this seems to be evidence that on average we... If our beliefs are really strong, we should pick it up in the data. We are not going to resolve this issue. Perhaps given those beliefs, people are still behaving in a correct way, and we shouldn't be seeing anything, but I will fall on the side of... I think, given how strongly they express those beliefs if they truly hold them, then we should see something in their actions.

Anthony Fowler: 

And we are talking about something that we're saying we should expect to see it on average across lots of people. Of course, for any individual spending decision, there's all kinds of idiosyncrasy, right. I bought a new phone recently mainly because my old phone was really old and broken down.

Wioletta Dziuda: 

Yay! Our recordings are going to be—

Anthony Fowler: 

And I'm now recording on my new phone.

Wioletta Dziuda: 

Better quality.

Anthony Fowler: 

So, I hope the sound quality will be better. I don't think that was because Joe Biden is president. I mean, it might have been on the margin, but I think it's pretty unlikely that that spending decision was influenced by Joe Biden. So, we're not saying that... And similarly, if I hadn't bought a new phone, you wouldn't be saying like, "Oh, you're lying when you say that you think Joe Biden's handling the economy well." But across many, many people, if it's really true that a million people all think the economy is doing better, we should expect them to spend a little bit more on average, and same thing for... So, I think I'm willing to disagree with Will on that. That I think we should expect to see some genuine difference on average and behavior if this is a genuine belief.

Will Howell: 

And so, what you take away from this is that what is being said in the surveys is not a genuine belief. It's the partisan cheerleading story. Maybe a genuine belief, but not a belief about the thing that they say that it's about.

Wioletta Dziuda: 

Yeah. So, I think to me the most convincing story is the one that Mary mentioned in the interview that: I live my life. I make economic decision based on my checkbook and my budget and so on, but I don't sit down and ask myself the question, "How is the economy doing?" So, when you come and ask me this question that I'm not asking myself on daily basis, I want to answer very quickly. And I'm looking for something that will have me answer this. And the best cue that I have is with the president Biden. "Oh, Biden seems like a reasonable guy. The economy must be doing well." Even though I might have been fired a second ago, it's just you're asking this general question that I don't think about on a daily basis. So, that doesn't sound like a partisan cheerleading. That sounds like you're forcing me to take a stance. I'm going to take a stance, and that's the easiest thing for me to say, the most reasonable thing to say without any information.

Now, having said that, I'm not sure that this is not problematic. Perhaps it's as problematic as partisan cheerleading, but sounds a little bit less insidious or less... Gives me more hope. People are not just saying things because they want to stick it to the other side. They just say it because it's the easiest thing for them to say, but it might have exactly the same negative consequences. I don't know what you guys think about this.

Anthony Fowler: 

This is related to the discussion we had with [inaudible 00:36:06] and with Greg Huber in some of our previous episodes where we were kind of like, how troubled should we be if people survey responses are super partisan? Does this behavior that we're observing in surveys actually spill over into regular life? If it's just a thing you say in political surveys because it's kind of the easy, fun, cheap thing to say, it seems to me that it's not that big of a problem normatively for democracy. But if whatever we're picking up in surveys spills over into people storming the Capitol building or people screaming at their relatives over Thanksgiving dinner table, then that does seem like a problem. So, we don't see behavioral implications in terms of actual spending. Do we see behavioral implications in some other ways? Interpersonal communication that's breaking down or things like that.

Will Howell: 

So, the obvious one here given what's happening in the world is to think about COVID and masking. If you ask Republicans, do you think COVID is real? They're going to be much less likely to call it a real thing and a real threat. And, in fact, we see levels of masking and vaccination rates being much lower among Republicans. So, do you two then want to say by the same criteria that you're putting forward and Mary's putting forward? And then I'm sort of I'm struggling with that then that's evidence that they really do believe it.

Wioletta Dziuda: 

I think you can't really generalize because masking example has an additional complication that I want to signal who I am and perhaps I'm scared of COVID and I'm going to put the mask when no one is looking at me, but my neighbors are not masking. And I want to tell them that I'm not one of these crazy Democrats who are just freaking out about this crazy COVID. So, the vaccination rates, I think this is a little bit more of a serious data point because you can vaccinate yourself in secrecy. So, here, I would say if someone is not vaccinating themselves, they truly believe that they shouldn't be vaccinated. They might believe COVID is real, but they shouldn't truly believe that they are better off not vaccinating themselves.

Will Howell: 

But this is an instance where if we bring these criteria to bear and that we observed Democrats and Republican saying at very differential rates whether or not they think COVID represents a market threat. And then you see this behavioral correlate in vaccination rates. That's a point in favor of the view that, in fact, Democrats and Republicans in some ways are living in different worlds. No?

Wioletta Dziuda: 

Yeah, I think I would agree with that.

Anthony Fowler: 

I think that's right. I think I agree with that. Yeah. I agree with Wioletta that masks are not a great test of that because of the other complicated social signaling thing. Vaccines are probably good tests of that. Mary mentioned this other study in her interview and I don't know the details of the study well, but she mentioned a study about outbreaks that happen after people's birthdays. That even in pretty democratic places, there are still big COVID outbreaks after birthdays, which suggests that there are maybe people kind of lying on both sides. Maybe there are some Republicans out there who say they're really worried about COVID, but then they still secretly go get the vaccine or say they're not worried about COVID, not a big deal, but then still go get the vaccine. And maybe there are lots of Democrats out there who say, "COVID is a huge thing. Everyone's gotta shelter in place," but then when their birthday comes around, they still have a party with their close friends. So, I think this is a useful way to try to get at whether or not people really put their money and put their actions and put their lives, in this case, where their mouth is or they're actually changing their behavior. I like that way of trying to answer this question.

Wioletta Dziuda: 

So, she's responding to a paper by Gary Huber that showed that the effect exists. That they found actually a positive correlation between the beliefs and the behavior. The original paper was published in APSR, which is one of the three journals that every political scientist really wants to publish. Mary's paper is published in an excellent journal that we all know publishers of the very good papers, but a journal that is not considered a top journal in the profession. So, I find if you have some thoughts on what this tells us about how our profession works and how we seem to value and appreciate the first surprising result. And by all means, I think the first paper was great in that I think it was a great idea to look for whether there's correlation or not. And they found the correlation given the data they had, but I feel like we jump on this first finding that's cool and fun. And then when someone does their job and tries to really see whether the finding was correct or not and do this hard work, and then we value this contribution a little bit less.

Will Howell: 

This is red meat for Anthony Fowler and him to complain about the state of discipline. Anthony, go.

Wioletta Dziuda: 

That’s what I thought! So I thought I want to bring it up and then—

Anthony Fowler: 

Yes. Oh, yes. So, first, let me just say that the Quarterly Journal of Political Science, which is the journal that Mary published in is the best journal in political science.

Will Howell: 

There we go.

Wioletta Dziuda: 

It’s excellent. We all agree. We all agree.

Anthony Fowler: 

If you just asked, which journal can I trust the most? Which journal can I pick up and read a random paper and the result is most likely to be reliable? It's the Quarterly Journal of Political Science. And I'm not just saying that because I'm an associate editor of the journal, but I genuinely believe and I believed that even before I was associated with the journal, but that being said, Wioletta's right that the original paper was published in what would be viewed by most people as a much more prestigious journal. And yeah, I think that's a huge problem. I think you point to a huge problem, which is, especially true for these... What's a really catchy, exciting result? What's the kind of thing that's going to get people excited and be published in the top journal and the discipline? It's not going to be the most reliable, most credible, most carefully conducted study. It's going to be the splashiest study. It's going to be the one that's like, "Here's a huge new idea. And it's a really splashy result, and it's the kind of thing that the New York Times is going to pick up, and it's going to be a headline result and people are going to keep talking about it forever."

And that's a disease in our discipline. It's a disease and that we are attracted to these kind of exciting findings, whether or not they are reliable or not. And then the fact that the correction is published in a better journal, but slightly less prestigious in the minds of other people journal is also a disease in our field in the sense that the editors of the APSR don't care that much about what's right. What they care about is the flashy finding. So, I don't know the history of this paper. I don't know if Mary had submitted this paper to the APSR, what they said, et cetera, but I suspect if she had published her paper to the APSR, they would have said like, "Oh, this looks really good. It's probably right, but it's not really exciting. And so, we're just not going to publish this kind of paper." Which is embarrassing. It's embarrassing because they should want to correct the record and they should want to... Gerber and Huber didn't do anything wrong. There's no allegation they did anything wrong. They might've written a perfectly good study. The result happens to be wrong. The APSR should want to publish that correction. People should want to publicize it.

Anthony Fowler: 

And I'm guessing if you poll people in political science, a lot more people know about the initial Gruber-Huber finding and they cite it more often than people know about the fact that Mary McGrath has come along and shown that it's probably wrong. So, there we go. So, that was the red meat that Will talked about. I got to rant for a little bit.

Will Howell: 

It's a real thing. It's a real thing.

Wioletta Dziuda: 

I think this paper... And this is my bottom line. I think this paper together with the paper that we've discussed previously on paying people to answer the question in a more correct way, I think they all sort of are these warnings that really when we hear about this crazy poll that 60% of Republicans believe the election was stolen. We shouldn't immediately jump and say, "Oh my God, they are so crazy. How can they be so crazy? They're insane." We should say, "Perhaps they're crazy and insane. Perhaps they really believe that," but there are many other explanations that are possible. And there's evidence that frequently... There doesn't seem to be this tight correlation between what people express and what they truly do or really believe. So, for me, that's how I view the paper.

Will Howell: 

What about you, Anthony?

Anthony Fowler: 

I agree with all of that. I find the paper to be pretty convincing in that there is not any really compelling evidence that people change their spending behavior in response to presidential elections in the way you'd expect if they really did think the economy was doing better, but just because their candidate won. And I think that does suggest that people don't really mean it in surveys when they say that the economy is doing way better just because their preferred president is in office. If you just look at the surveys, you're going to get a lot of wrong answers. And you're going to think that people are actually way more partisan than they really are.

Will Howell: 

So, I agree with both of your points on lots of things. And I have this lingering discomfort. I think that the finding that she's putting forward is real, and it's an important corrective to a previous paper that was really provocative and important. I guess I want to hold out the possibility that the absence of a correlation here doesn't mean that we're right back into a world strictly a partisan cheerleading because I can imagine scenarios under which I, as a Democrat, say, "Yes, I think the economy is going to be better, but it's going to be for people who are not me. And so, my consumption isn't going to change one way or the other, but I think it's going to be better for a set of people who I care about." Or the story earlier about people who are very wealthy and their consumption patterns aren't going to change, but other kinds of economic decisions that they make might alter.

Your point that both of you make is that, no, this is in fact on average. I need to keep scratching my head on, right? So that we don't even observe it all on average is important. It's worth grappling with, but I want to keep thinking about how we assess the validity of beliefs, how they relate to actions when we don't observe a correlate that's associated with actual behavior that then that forces us to rethink the belief. I want to keep thinking on.

Anthony Fowler: 

Thanks for listening to Not Another Politics Podcast.

Wioletta Dziuda: 

Our show is a podcast from the Harris School of Public Policy, and it's produced by Matt Hodapp. Thanks for listening.