November 10th, 2020 – Second Reading Podcast

[0:00:00 Speaker 0] welcome to the second reading podcast from the University of Texas at Austin. The Republicans were in the Democratic Party because there was only one party. So I tell people on a regular basis there is still a land of opportunity in America. It’s called Texas. The problem is these departures from the Constitution. They have become the norm. At what point must a female senator raised her hand or her voice to be recognized over the male colleagues in the room on Welcome Back to the second reading podcast for the week of November 9th 2020. I’m Jim Henson, director of the Texas Politics Project at the University of Texas at Austin. I’m joined again this week by Josh Blank, research director of the Texas Politics Project. Good afternoon, Josh. Good afternoon to you. It’s been just a few days, really, since our post election off schedule podcast last week that we we did on recorded on Thursday afternoon. Not a ton has changed, I guess. At the national level, they have called the race for for Joe Biden, though I think that was anticipated. The Trump campaign, to some extent the Trump administration are, are resisting that call thus far, and so that remains to be resolved. And I think is that continues. Maybe next week we’ll return to the impact of that on the legitimacy of our people’s views of the legitimacy of elections, which I’ve been thinking about a lot. But we must see how the next couple of terms of that going, what that looks like. I think today we wanna join the discussion on polling a little bit more fully and what went wrong in in 2020 which is a the most deja vu of deja vu conversations ever. I think, let’s say what went wrong in 2020. Well, let’s start with the polling. Eso Yeah, I think that’s there’s a lot of that going around in terms of the number of things people could talk about. I mean, we do want to kind of round up some of the usual suspects, and this is a very obviously very active discussion going on right now. Um, you know, come from stem to Stern E. I just want to throw something out there. I mean, I’m just thinking about this in the moment. I mean, we’re being dismissive of. It almost is a is a point of self deprecation. But, I mean, I do think that, you know, there’s Ah, there’s an important thing that, you know, in terms of focusing on this polling, which is just, you know, in interpreting what happened in the election and our understanding of it in the meaning of it, ultimately, you know the best, Really, the way that we’re gonna get the most depth of understand that is going to come from polling, which might sound like a joke right now, but that’s kind of part of the problem. And so it is actually not unimportant as really, you know, deal with what were what was not a very good day for polling. Yeah, you know, And when we say not a good day. I mean, it was not a good day. I mean, you know, early kind of compilations of of the state level polling showed the polls being off virtually everywhere and and in some key areas, some of the key states by quite a bit 7.5. If you look at the polling averages generally, the trump margin was underestimated by Aziz Muchas you know, 7.5 points Um and that was pretty uniforms, you know, not uniforms. That was pretty true. Across the board, a couple of states were reasonably close. Most were pretty far outside of what we would have think, you know? I mean, it’s a it’s an average. I pulled the polling averages of what we’re looking at. But if we were to put margins of error on those polling averages, they were outside of what you would say. Well, that was close. Yeah, and and the real problem here is and this will be the first, you know, a glossary term watchword. Here we go, but like, but it’s also systemic error, and that’s part of the problem here. It’s not just that there was error. It’s the fact that all of the error went in the same direction. Whether you were looking at all the national polling, whether you’re looking at the polls and individual states, in all cases, they were underestimating trump support. And that Z you know, the rial, the real problem here. Yeah, basically, I you know, I guess. And if you look at the love and competitive states, 10 over predicted Biden’s performance, right, right. And you know that’s not really gonna That doesn’t say much in in Texas. It waas you know the same phenomena and and in the same ballpark, right? If you only averages. Yeah, that’s right. And I mean I mean, even before we get to that, you know, you make, you know, we talked about 10 states live in states and something that, you know, we don’t have a lot of time to make in other venues for this, but I’ll make it here, which is, and it doesn’t really matter what kind of state we’re talking about. So if you start thinking well, was it just, you know, Republican states or just democratic states? Where was it? States with early voting, but states or states without doesn’t matter. Whole range of possibilities doesn’t look like the underlying conditions could explain the outcomes. Let’s just set that aside in Texas theatric jizz, You know, the polling average going to the election expected a pretty narrow trump victory. Fivethirtyeight said that the trump was up on average by 1.1 points. Real clear politics. Another site that aggregates thes so that he was up by 1.3. It looks like he’s gonna lose. He’s gonna up winning by between five and six points. Now, do you want to see E? Do you want to keep talking about the Texas polling for a minute or Yeah, let’s do that. We’re gonna be insured in that, okay? I mean, you know, I think one really notable thing about that error is that it it got worse as we got closer to the election, which is a strange phenomenon, because generally we tend to think that polls don’t get more accurate. As we get closer to the election, we think it’s going to be the opposite. The opposite, right? But But if you looked at But if we just look at the average of the polls conducted just a simple average if your if your stats together, I’m just This is just a simple average of the polls conducted prior to early voting. So in the month of September and the first few days of October, before early voting started, Trump was up by 4.3 points on average. Should not far off from the 5 to 6 point, certainly within the margin of error of that polling. If you look at the polls conducted during early voting. Trump was up by 1.7% on average. Which, you know, I mean is kind of striking and also makes no sense. I mean, I’m just going to say that I think it’s, you know, in the sense that, uh, if there’s one thing we knew about these candidates in this campaign is that, you know, voters were decided. I mean, voters were clearly decided how they they knew how they felt about both candidates. They certainly knew how they felt about Donald Trump. And you know the idea that you had this sort of consistent Trump lead, you know, really again in this sort of let’s say, you know, 4 to 8 point range for much of the campaign and then in the last month, all of a sudden, despite the fact that every other poll up to that point said, you know, 90 plus percent of voters know who they’re gonna vote for, All the partisans were gonna vote for the partisan, you know, basically for their partisan candidate. And yet somehow, in that final month, the idea that the polls were tightened by 34 points is a little hard to believe it’s easy to say in retrospect, obviously, but also, it didn’t really make sense when it was happening. Well, I mean, you know, and I think this gets to a little bit to how we discuss these things, But, I mean, I think, you know, we’ll put a pin in the fact that we’re going to talk about you know, how people put together. They’re likely voter pools. And that seemed to be one of the things we were looking at. Like, Who are these voters that these polls are measuring that air producing such closer outcomes. And we talked about that a little bit of the time, you know? But, I mean, one of the things I you know that I’m struck by is I think back on the discussion of this is just a you say how much you know, the specific fundamentals of what was going on, you know, seemed not, you know, seemed to be telling us that Trump was going to do less well than he than he did in 2016, which turned out to be the case. But we kept looking, you know, You know, I feel like we were missing three or four points all along. I mean, I think on this podcast two or three weeks ago, or, you know, I don’t have the week before when you’re sitting there kind of trying to do the back of the envelope math and saying, Okay, so you know, he’s eroded maybe five or six points among Republicans, and he’s eroded. You know, there’s this sliver of independent voters that have, you know, flipped out. And we never could quite get that tow add up, you know, to on erosion of, you know, seven or eight points, which is what the polling averages were telling us, right? I mean I mean, yeah. I mean, I should add, you know what I was saying, You know, going into the election that everybody’s opinion, you know, clean seemed so demonstrably fixed. But I mean, the story of the Trump presidency in Texas to to a large extent, was the extent to which attitudes towards the president his behavior is the administration were incredibly fixed incredibly early. I mean, I remember a poll we conducted in one of the more turbulent periods of the Trump presidency after Charlottesville. I think what happened There was a lot. I mean, this was at the point at which it was Charlottesville. It happened. That was when Colin Kaepernick started kneeling during the national anthem and sort of the sort of fight with the NFL over that there were a lot of a lot of things going on, and we ended up doing a whole battery of job approval metrics, specific job approval metrics on how Trump was handling these things. And again, it was, I think the list was, like, you know, 10 or 16 was a battery. Yeah, and I mean, some of these things were demonstrably going poorly or were you know, clearly, you know, it’s a by just an objective standard, not a great handling of whatever the situation be. And, you know, attitudes were just so polarized and so solid that, you know, a majority of Republicans, I think, you know, express approval and relatively strong approval on pretty much everything we tested. You know, the majority of Democrats, you know, expressed disapproval in most cases, strongest approval on most of the areas tested. This was you know, I think there was an expectation that the way that Trump Trump was approaching the presidency and really, you know, his sort of jettison norms and all these kinds of things would lead to some sort of backlash. And we never saw it. That was two years ago. I mean, at least, and there was little of it, and there was very little evidence of evidence of it, right up to the eve of the election. Nothing changed. And so, you know, I mean, it does you know, the narrative at the time and some, you know, in retrospect. And when we talked about this, I think some of the time that, you know, the narrative was that the Democrats, we’re gonna we’re gonna do one of two things that we’re gonna benefit from defections away from Trump, of which there was no evidence, or they were gonna change the composition of the electorate. And as we char and as we And as we started looking at the early voting data, I think there was a There was a real tendency to kind of say Okay, well, there’s this big group of voters that we don’t know anything about. That was those general election. Only voters or those of those voters with no primary records and I think it turned out that those voters didn’t look very different than everyone else. Well, I would say, you know, even beyond that, you know, even setting aside what we didn’t know, there’s also what we didn’t know, which is, you know, even within the huge surge in early voting turnout. And we’ve mentioned Derrick Crime last week, Republican consultant who analyzes especially early voting in particular and provides information on it. You know, it wasn’t as though Democrats were outpacing republic. I mean, I am. I see Democrats. I mean, Democrats have a consistent history of voting in Democratic primaries or voters have a consistent right versus voters have a consistent history of voting Republican primaries. Ultimately, we know those air Democratic and Republican voters the 90 you know, let’s say 98% plus of the time. And it wasn’t as though with the surgeon turned out, Democrats were rushing ahead of Republicans and, you know, creating a deficit that they would have to make up on Election Day. If the remainder of those voters turned out to be, let’s just say split 50 50. It wasn’t even looking that way. And so I do wonder I mean, one of the things I wonder I don’t have an answer for this is you know, to the extent that you know, pollsters were finding increasingly tight races in Texas, you know, and increasingly tight race in Texas in the final month. I wonder what justification they had for believing that I just I don’t know it, but, I mean, that’s an aside. We can come back to that. Yeah. You know, that may be an unanswerable question. Although it may have something to do with, you know, something we talked about last week, which was, you know, faulty assumptions about what they could expect from from Hispanic voters in Texas. Yeah, E. I mean, there are a lot of candidates to be fair. Yeah, it’s funny. I mean, with that, You know, we talked about this last week about the extent to which, you know, we consistently find somewhere between 30 and 40% of the Hispanic voting population supporting Republican kids. We found that consistently, you know, it was consistent before Trump came to office. Has been consistent since he’s come to office. But I think I saw, you know, I think this was from one of the other big national pollsters released, uh, basically the party I d. Breakdown of the vote choice breakdown amongst you, a large sample of Hispanics of different origins. And so whether the origin country or, you know, is Mexico vs Salvador versus you know, Puerto Rico, Cuba. Now you know what they found. Not surprising. You know, Cuban Americans were the most Republican group amongst Hispanics and, you know, it was funny. I was looking at it and amongst you know, you know, people with Mexican ancestry of which that’s the majority of Texas Hispanics. It was the same numbers. It was about 70 30 Democratic lead. And that’s again, that’s nationally. That’s not Texas. And so it’s like, you know, I don’t know why this was such a surprise to people the Biden wasn’t doing better or wasn’t doing. You know, again, I would say beyond reasonable expectations with Hispanics because it was beyond reasonable, and I just and I wonder to what extent, you know, again, this influenced people’s reading of the polls. If the assumption waas again this thing that again, there’s been no evidence for that. If we just keep putting more Hispanics in the electorate eventually that Hispanic group is gonna become more democratic. Well, I don’t know if that’s true. Yeah, I think there’s a lot of discussion of that e I don’t know if you thought I was gonna I was gonna send it to you. And I know if I remember that her colleague David Lay all has a co authored piece in the Monkey Cage blogging The Washington Post that talks about some of these issues and, you know, expectations getting ahead of reality. And I think, you know, it’s been a cliche of talking about the Latino, you know, the quote unquote Latino electorate forest, long as there’s been one really or one conceived as such, you know, I mean, if you had a dime for every time somebody said Latinos are not a monolithic group on either Fox or MSNBC, you know, you could retire yesterday. Um, but I think that’s really coming home to roost now, in terms of, you know, honestly thinking, you know, this has been a very controversial thing in kind of Latino studies and and the academic study of Hispanics. At some point, it is fair toe wonder where the conceptual groundwork or the conceptual integrity of Latinos as a group, as a group, identity comes from right and what utility it serves other than, as, you know, some kind of kind of gross geographic, central geographic slash linguistic generalization. And I’m just, you know, you talk about that a bit, but, I mean, you know, you know, I think it is an under appreciated piece of this. I mean, because if you had said, hey, we need to look at the Mexican American vote, the Cuban American vote and the Puerto Rican American vote taking, you know, the two the three biggest nationalities setting aside Central America again is a region which in itself would break you into something you would require you to break out Salvadorans and Nicaraguans, and you see where I’m going. I mean, on one hand, that makes a lot more sense. And you’re less likely to have this confusion over. Well, wait a minute. What about the you know, Miami Hispanics first is the, you know, etcetera, etcetera, etcetera. And I think that that has not been fully confronted in part because of you know, the way that we’ve settled on the fact that we talked about Hispanics and Latinos as a za racialized group. And one of the problems that we’ve talked about any that I’ve mentioned in here before is that you know, thean implicit comparison, which is sometimes, you know, very misleading with African Americans. Yeah, I mean, I think, you know, you even leave out of that. You know, the addition layered complexity of, you know, generational status. Right? So are, you know, are you a first generation American is a second generation. I mean what you know. And that adds another layer across the races and ethnicities in terms of, you know, corporation decisive is just anyway. Anyway, we could probably just talk about this for a while. Maybe we’re kind of dancing around, I guess. You know what? You know what we think? The potential challenges we’re here. Although, yeah, I think that’s right. And so you know, I mean, so let’s you know, let’s talk about the You know that quickly again. You know, the narratives that are out there, you know? So the return, you know, this week has been the return of the shy Trump voter, right? Did you Did you see the shy Trump voters driving down main street in their caravan. I did waving their flags to high. Yeah, Yeah, now you know, So, yeah, so we can lay out. You know, the shy Trump voter theory is, you know, basically, at least in its in one form, that people responding to polls that support Donald Trump feel social pressure to not admit that they that they favor him, that they would that they would vote for him, Right, You know, and this is I should say, is something is a theory. I feel like comes much more from pundits, uh, than from actual pollsters. I don’t really hear pollsters making this thing, because, to be honest, there’s almost no evidence for it. And in fact, you know, this was a consideration in 2016, pollsters went back. Look at the data. They’ve been doing a lot of stuff in the meantime, experimentally to figure out, you know, is this even? Is this plausible? Is this likely to be the case And you is the simplest kind of quickest example of why it’s probably not is it doesn’t matter whether you asked people whether they you know who they support with a live interviewer. Someone on the other end of the phone. Who’s a human being, who presumably is the person that you would be being in shy in front of, or whether you answer the question online or via text message or to a robot? It was the case, you know, Trump voters were shy about expressing their preferences, and you’d expect to see greater support for Trump in polls where the person is not where the respondents not actually talking to an actual person. But we don’t see that. I mean, there’s some other problems with this and also that are more complicated. But we don’t see this. And I mean, I think, you know, in the other piece of this is I got to say I really I almost despise this argument and despise is a strong word. But part of it to me is the extent to which pundits who don’t really conduct polls are talking about. What it sounds like is a group of people expressing a preference that they can’t imagine people would express, even though 70 million people express that preference like it’s not as though you know, there are trump voters in a lot of places and your joke earlier about the shy Trump voters kind of speak. So, you know, I mean, why on one hand it might seem, you know, even less likely this time, given that we’re not seeing a lot of evidence of shyness among many Trump voters given boat parades and caravans and, you know, Maga merchandizing. And but, you know, I mean, it’s it’s fair to say that those, you know, I mean, the media coverage of those events, those those those folks may not be representative of your average Trump voter. Just to be fair, you know, you know Aziz, we think about this, and we were just talking. Became on about The New York Times, Nick Cohen’s piece on kind of explaining, You know, his first cut, this which he ends by saying, You know, as we all are, we’re gonna have to wait and see as he’s kicking around the ideas. But, you know, you know the idea that there might be something that’s that’s a different that could overlap the existence of shy Trump voters idea of non response bias that you know that that Biden supporters are more likely to respond to polls and Trump voters are less likely to respond to polls. I think I think it was in the Nick Cohen piece. You know, the suggestion that Thea anti institutional bias of some section of trump voters and particularly the anti media bias, is so strong that they would not want to respond to pollsters. Maybe, you know, I’d like to test that more. You know, You know, I mean, it would be interesting, you know, to see trying to think if we had any items that we could really test that on, you know? But what people have done this in some other ways, I mean, one of the things and again, we won’t get too complex here. But I mean, ultimately, we call this non response bias, and it’s something that’s common in polls. And we just there some groups of people who are just more likely to respond to polls than other people. We saw this in 2016 in the sense that you know, the big the big alleged problem of 2016 was the overrepresentation of college educated voters. College educated people are just more inclined to sit on the phone with a pollster or with the call center, rather for you know, sort of time talking about what they think about stuff, and that’s just part of life. It’s something that we deal with. And so, you know, this idea of not responsible for something that was being kicked around in the professional community? I think you know, even I recall the last time I looked was at least, you know, 67 months ago were coming to the election. People were thinking about it, come up, and then what they did was, you know, because a lot of polls will Ascot the end of the poll, who they’ve voted, who you voted for in 2016 and number one There wasn’t, like a huge evidence of, you know, shyness. They’re about reporting you voted for. But one of things you could do is you could wait the results of your survey back to the demographics of the 2016 electric and the vote share. And in doing that in most polls that that looked and most pollsters who looked into this it didn’t change their results. That was the thing. It barely moved it all and really have no consistent direction. So there’s it’s the preliminary evidence on that front is also relatively weak. But we’ll say that will get looked at moments, but the most, the most common or not the most common. The most plausible thing is you and I have been talking about it does go back to something again that we’ve talked about here before, but it’s kind of in a way we we talked about it a few weeks ago, mostly in the realm of carping, I would say in terms of talking about people not disclosing their likely voter models and and, you know, the whole the you know, the art science. And, you know, I think to the point when we talked about it, the mystery of people’s likely voter models turns out that I think you know that’s gonna loom pretty important here, uh, in terms of people estimating what the actual population was, and we kind of talked about it on here before on the podcast before. But it’s becoming apparent that all the uncertainty in the electorate in terms of the increase in turnout, who’s gonna vote, you know, just just the size of the electorate. You know, writ large really underlines the idea that we really got to think a little bit, Maura, About how we’re defining likely voters and how we’re communicating that. Yeah, that’s right. I mean, I’m not allowed to carp. Is that right? No, no, no, you carp. Okay, good. Yeah. I mean, this is something that I think the basic premise here is that every estimate of of an electoral outcome has some underlying electorate basically assumed in that estimate. Now again, posters, maybe Mawr or less explicit about about how much they want to guess at the electorate. But ultimately, it’s a guess, right? And this is the key thing here. Is it Normally, you know, I’ve been saying for, you know, at least last week or so It’s frustrating because, you know, election polling is just a subset of all surveys. And in most cases, when we conduct surveys, we know Ah, lot about the population were serving. In fact, we know, you know, usually you know very well usually very good information about the populations were serving. So if you ask me and we know more and more all the time now, given the use of databases and yeah, exactly what unquote right? And so But I mean in the simplest sense. You know, if someone tells, you know, pollster to go conduct a poll of adults in Texas, we could go to the most recent census data and the most recent census system it to tell us what share of the older population should be male between the ages of 18 and 30. With a college degree, we can know that we actually have really good data about that kind of stuff. We have no data besides prior elections and what people tell us when it comes to deciding what the actual electorate is gonna look like. And so, ultimately it forces pollsters to come up with, you know, again, there’s a likely voter models says Okay, well, how are we gonna include people in this electorate? But then underlying that again is a is a composition of the electorate, and I think you know that makes a big difference. I mean, on I was joking with you last week, but I mean, I feel like when I started in this role, people would ask us all the time. They said, Well, what we put out an election is when people would say, Well, what’s your that? What’s your you’re likely voter samples, Hispanic, you know, And it was often Democrats would ask. But the assumption being that we didn’t have enough Hispanics and likely voter sample was kind of what the underline This was seven or eight years ago and you go and say, Oh, it’s this or that seven or eight months ago. Yeah, that’s true. But then things I feel like those questions have kind of kind of stopped. I mean, and I don’t know whether it’s the volume of polling that took place. I mean, certainly in Texas there’s an increased volume of point, but there’s a volume of polling everywhere, and to a large extent, you know, I’m not blaming the media for this, but the reporting of it was just say, Here’s the estimate. It’s by this, you know, by this person, this group, this institution, it’s of likely voters in Texas. And that was about the extent of the information and the thing that I think we’re always curious about, and we’re always looking into when we see these new results is, well, one. How did you define a likely voter? And to what? What is What’s the composition of this electorate? then that’s producing this result. And in a lot of cases, that information was not available. I would say most cases, actually that information was not made available. Anybody and I think you know, you combine. You know, I think the difficulty of handicapping what an electorate would look like Like in Texas. That adds, You know, 2 to 3 million voters depend on whether you’re comparing it to 2016 or 2018 again, the likelihood of AH pandemic and and a economic crisis pulling more voters into the polls because of the fact that politics all of a sudden becomes more pertinent and more salient to people if it’s affecting their livelihoods. And I think there’s a lot of uncertainty around that. And I think you know the polling. I’m gonna say the entirety of the polling enterprise, from pollsters to aggregators to the media and the way that they report it really glossed over all the uncertainty inherent in that and just started saying, you know Well, Biden’s up. Eight. Yeah, we have a 90% probability of winning. Yeah, and and not to be, too, you know, I don’t know if this is abstract or not. to be, too, whatever about it. But I mean, like, the bottom line is that, you know, you were kind of, you know, kind of laying this out. Is that you know, we’re actually estimating a population. We’re estimating we’re taking a sample from an estimated population that actually does not exist at the time that we’re polling, right? Because because people are, you know, because we’re estimating what the electorate is gonna look like after the fact. Yeah, that is a, you know, difficult that is not captured That is not captured in this standard margin of error measure no. And the standard margin of error because you’re in and of itself is actually sampling. That’s only sampling error. So when you hear you know the margin of error in the poll, what you’re talking about is you’re talking about sampling or the rial overall error with you know what they call the design effects which account for non response bias. And the waiting is usually about twice that. So So just from just from the beginning of this, you know we are is a collective whatever group of group of people pushing from this information out. But you know we’re basically decreasing the amount of uncertainty that we say that we have by about half, which, you know again, uncertainty is part of the enterprise. It’s a sample, and again, as you point out and this is the key thing, it’s a It’s an estimate of the preferences of a population that doesn’t exist, which you know when and when you say it that way. It’s crazy. It’s crazy. I mean, it’s also like, why, you know, we gotta we often go election polling over. Let’s get back. Thank goodness the session is coming and we could talk about policy where we know what the population we’re estimating is stable. It exists and and in that the interesting thing about that kind of work is it’s really more about asking the right kinds of questions. We’re in election polling. The question is fixed more or less. I mean, there’s some ways in which it might not be, but the question is fixed. Is this candidate that candidate and whoever else is on the ballot? But the rial nub of how accurate your estimate is going to be is going to come down. Thio getting the mix of groups in the electorate close to right. But even even even even that even even that question there were some polls that, you know, didn’t you know, didn’t do trial ballots that got in other words, they asked, Would you vote for the Republican Donald Trump, the Democrat, Joe Biden or someone else? Now? So many problems with that? Yeah, and there’s huge problem. And clearly there were people doing that. We won’t name any names, but if you go back to our polling blogging, you follow some links. You also you’ll see people that did that. And it’s problematic, right? So that’s just, you know, that’s ah, just the thing that occurred to me that, you know, we could if we had to really, like, take those things apart. So I mean, as we think about this, I mean, watch for a not bad that Josh and I think we’ll probably be circulating later in the week and unless they don’t like and it’s just a very, you know, short opening gambit at this. But I you know, we’ll talk about this more and and we’ll have some more stuff with website going forward. But you know, two things here one disclosure and more transparency. That is, You know, there’s been a discussion that’s been going on for 15 or 20 years now about better norms of disclosure, better standards of practice in terms of how poll results are reported out by the people doing the polls and the entities that are reporting them. And, you know, in the last for the last decade or so, aggregating them quote unquote. In other words, they are the real clear politics and the 5 30 eight’s of the world. We’ve really got to revisit that in terms of the issues we’ve been talking about today, you know, And and and there’s gotta be an intermediate piece of that that is about better explanation and better contextual ization. And I think that, you know, one of the things that’s happened as an unhappy result of the presence of poll aggregators and, you know, sections of The New York Times and The Washington Post. And, you know, the you know, the major media outlets that say, Hey, this is our kind of data science data journalism section where, you know, basically, and this is the key part, if you’re interested, you can learn a lot more about this that has not stopped the beat reporters and the and the daily kind of horse race coverage from simply reporting the numbers and moving on without any context. And something about that has got to change that. That, I think, acknowledges the fact that look, people complain are gonna complain about this. Poland is not gonna go away because that’s what you got. And it’s too deeply ingrained. So it’s gotta be how we do it that has got to get better. And it’s got to be reporting. Some of it is reporting on our end. But a lot of it is also gonna be how it’s propagated. Yeah, I mean, ultimately, you know, something that I’ve been saying throughout the cycle is that, you know, these are just estimates. These were not, you know, perfect representation is thes air. The best attempts of you know, I think usually, you know, straight shooters who are trying to do the best work they can because there’s really look, I mean, this is at a point like there is no, I mean the idea that, like, there’s this question of like, Oh, is this just liberal pollsters? Well, look as a There’s two things I kind of point out against that. Number one. There’s no point in doing a bad job professionally to advance your career. That doesn’t really work. And so a bunch of people doing a bad job for some kind of, you know, perceived abstract good, let alone the fact that I don’t believe that, you know, poll results. Being out really changes behavior much in an election, and especially in this election. But ultimately, pollsters don’t get more work by doing bad estimates, so I just don’t really buy that. But the other thing is that people are talking about the liberal, you know, liberal media bias or the, you know, liberal bias and wanting to see thes kinds of election results. But the truth is, it wasn’t as though the campaign pollsters were doing a ton better somewhere. But if you look for example, you know, at the final days of the campaign for both Trump and Biden, Trump campaigned in Iowa on the Sunday before the election, even though he won the state by eight points and Biden campaigned in Ohio on Monday before going on to lose that state by eight points if their data was so much better and so much, you know, less problematic than the data that was publicly available. I don’t see. Presumably they have made different decisions, presumably. So just to set that aside, so I think we’re gonna wind it up there. We will, thanks to Josh, thanks to the folks in the liberal Arts development studio in the College of Liberal Arts at the University of Texas at Austin. Thank you for listening, thanks to all the people that voted, even though they made the polling much more difficult, at least in part. And we will be back next week. Second reading Podcast is a production of the Texas Politics Project at the University of Texas at Austin