(music)

And Iowa based pollster with a national reputation has released her last election poll after a three-decade career. We'll talk to J. Ann Selzer on this edition of Iowa Press.

Funding for Iowa Press was provided by Friends, the Iowa PBS Foundation.

(music)

The Associated General Contractors of Iowa, the public's partner in building Iowa's highway, bridge and municipal utility infrastructure.

Elite Casino Resorts is rooted in Iowa. Elite's 1,600 employees are our company's greatest asset. A family run business, Elite supports volunteerism, encourages promotions from within and shares profits with our employees.

(music)

Across Iowa, hundreds of neighborhood banks strive to serve their communities, provide jobs and help local businesses. Iowa banks are proud to back the life you build. Learn more at iowabankers.com.

(music)

For decades, Iowa Press has brought you political leaders and newsmakers from across Iowa and beyond. Celebrating more than 50 years on statewide Iowa PBS, this is the Friday, December 13th edition of Iowa Press. Here is Kay Henderson.

(music)

[Henderson] Our guest today has been at the Iowa Press table often. Since 1987, she has been the Director of the Des Moines Register Iowa Poll. In 2014, she released a result that very closely hued to the results of the Joni Ernst-Bruce Braley race for the U.S. Senate. This past election cycle, just before the general election, she released a result that was wildly not indicative of the election outcome of the Donald Trump and the Kamala Harris race here in Iowa. Ann Selzer, we're going to ask about that and some other questions. Welcome back to Iowa Press.

[Selzer] Thank you. It's always a pleasure to see your smiling faces.

[Henderson] Also, joining the conversation, Katarina Sostaric of Iowa Public Radio and Erin Murphy of the Gazette in Cedar Rapids.

[Murphy] So, Ann, you're a self-described data girl and Kay mentioned the poll that came out right before the election. I'm sure you've been crunching the numbers since. What have you learned? What went wrong?

[Selzer] Well, I'm not here to break any news. If you were hoping that I had landed on exactly why things went wrong, I have not. It does sort of awaken me in the middle of the night and I think, well maybe I should check this, this is something that would be very odd if it were to happen. But we've explored everything. The Des Moines Register in an unprecedented move for transparency has put online our cross tabs, our weighting system and my analysis and that I've not needed to update because it was pretty complete. We don't know. Do I wish I knew? Yes, I wish I knew. But as I sit here today -- one of the political commentators before the election but after the poll said, they've got all of these pollsters that are fidgeting with their numbers and they're adding this into their weighting structure and they're doing this with their sampling and Ann Selzer has never had anything to fix. Well, until now.

[Murphy] Well, and I'm glad you mentioned that. Katarina is going to ask you a little bit more about methodology so I won't get deep into it yet. But, as Kay mentioned, you have a number of polls in the past that were remarkably accurate and not just when everybody else was, on occasions when others were not, like that Ernst-Braley poll in 2014. And maybe the answer is the same, but I'm just curious if you have any sense of why that methodology that has served you well for most of your previous career suddenly did not in 2024?

[Selzer] I wish I knew the answer to that. But, like I said, there wasn't anything that we saw that needed to be fixed. The reality is that more people supporting Donald Trump turned out. I'm eagerly awaiting the Secretary of State's turnout reports that will happen in January to see what we can glean from that. But there wasn't an adjustment to my data when we saw that it was going to be a shocker that I would have said okay, let's adjust it. It's not like I know ahead of time what the right numbers are going to be in the future. So, you kind of take the data designed to reveal to me our best shot at what the future is going to look like.

[Murphy] You mentioned, and real briefly I'll ask what some of the national pundits -- I know from hearing you in other interviews that your voicemail inbox, social media mentions has been busy let's say -- maybe just describe what that experience has been like. Were you surprised by that and some of the allegations that have been made regarding the way you approached this poll?

[Selzer] Let's deal with the allegations because I am mystified about what the motivation anybody thinks I had and would act on in such a public poll. I don't understand it. And the allegations I take very seriously. They're saying that this was election interference, which is a crime. So, the idea that I intentionally set up to deliver this response, when I've never done that before, I've had plenty of opportunities to do it, it's not my ethic. But to suggest without a single shred of evidence that I was in cahoots with somebody, I was being paid by somebody, it's all just kind of, it's hard to pay too much attention to it except that they are accusing me of a crime.

[Sostaric] So, you use a certain methodology in your election polling. Can you briefly explain your method and how you screen people for your poll?

[Selzer] Yeah, I will. And that might be a key to something that we can't just check out. So, thank you for that question. What do I know about the future electorate? Nothing. What do I know about the state of Iowa? Oh, I know a lot because we have a census. So, as we are interviewing people, we're going to ask a question early on about how likely they are to vote. They can say definitely, probably, might or might not, probably not or have you already voted. If you have already voted in our last poll, you're in and if you say you will definitely vote you're in. But for everybody else we keep them on the phone to get age, sex, what county they live in. That is going to give us some geography so that we can weight all of those people, I think there were 1,100 and some in our final poll, to look like the state of Iowa because that we know. And then we extract from that those people who made our likely voter screen and interviewed them. So, if they are proportionately more women who pass that screen, they are going to show up here but they are weighted at the statewide level. It's a little geeky. But one of the things I've been thinking about is with caucuses we take you if you definitely or probably will participate and maybe our screen was too tight. I don't know, there's no way to undo it and sort of check, well what would the probably people have said, because we very politely terminated the people who just said they would probably vote. But that has been our system since 2008.

[Sostaric] So, just to review, it sounds like you have likely voters and you get their age and their sex. Are there any other characteristics that you screen for?

[Selzer] Their county so we can weight by congressional district if needed. Our weights for this were tiny. Any pollster would look at what our weighting report said and said it's not like what we did by weighting our data to make it look like a cross section of all Iowa had something crazy in that. The weighting didn't change the horse race at all.

[Sostaric] And some pollsters have started looking at educational attainment of voters because there's kind of a split between college educated and non-college educated. Why isn't that something that you used?

[Selzer] We did take a look at that in previous elections and in Iowa we found that there were some interactions, a statistical term, between the variables we were using and weighting. So, it didn't make things better, it made some things worse. And then once -- this is what I call dirty fingers on your data -- once you see that it threw off the proportion of people who were married or not married, then you have to worry that it's actually harming your data rather than strengthening your data.

[Henderson] I've heard you say in a podcast that pollsters like linear, correct.

[Selzer] Yeah.

[Henderson] So, let's go back to 2020. Donald Trump easily won the state and then in 2022 Iowa was among the only states that had a big so-called red wave and republicans seem to be on the ascendency. So, why weren't you more surprised or said hey, there's something weird here when you saw what your data was telling you?

[Selzer] We did say something it weird here, we just didn't see that it was something in our data that was causing it. And the context for that answer about linear was looking at where Joe Biden was in June in the Iowa Poll and where Kamala Harris was in September and that was an uptick. I think she closed by 13 percentage points. I might not have that exactly right. But then you continue that line on the same angle and you get to where she was at plus three. So, it looked like there was a movement afoot and that this was, if she had jumped more vertically to an even bigger that would have given us pause. But Iowa has been, has voted twice for Barack Obama. So, it's not unreasonable to think that things might have been moving a different direction.

[Murphy] One of the things you said in that same interview and along these same lines when you talked about putting your dirty fingers on the data, you talked about well let's look at the data and if we did do that. And you said, if you had applied the 2020 electorate here in Iowa, if I'm remembering this right, it showed Trump plus six in your numbers. But that is, again, not your methodology, right?

[Selzer] So, it comes back to the question of how do I know before the election what the future electorate looks like? So, how do you know the right number? How do you know? We can't really go back and look at what the turnout was before because that might not be the turnout again. If we had done that, imagine after 2012 when Barack Obama was re-elected, things would look very different. So, in hindsight you say, wow why didn't you do that? Because it's not science.

[Murphy] Did you ever in any moment in that analysis consider killing that poll, not publishing it?

[Selzer] That would be a decision that the Des Moines Register would make. When we do a preliminary briefing the day before they get final data and we walked through it all because the numbers are jaw dropping. And I started by showing them the demographics and how they have lined up with previous polls that we've done. And I think by the end of that they were comfortable with it. And so, the question was never put on the table as to whether we would publish or not.

[Murphy] And that would be pretty rare, right, that the only time I'm aware of that that happened was in the 2020 caucuses with Pete Buttigieg and that was an error in the actual asking of the questions, not in the data.

[Selzer] At the phone bank, right, not in our shop.

[Murphy] So, that has never happened in your experiences otherwise?

[Selzer] Early on back in the Stone Age when I was on the staff of the Des Moines Register as the research manager, which included the Iowa Poll, I was away strangely as we were putting out the final data, I was in Washington getting training, and I became aware that there was concern over a large number of Iowans supporting Michael Dukakis. So, this would have been --

[Henderson] 1988.

[Selzer] 1988. And they convened. So, it's David Yepsen and Jim Gannon and Arnie Garson and Jim Flansburg and names that some of your older viewers will remember from the Register and they said, this is such a huge lead for Dukakis, should we publish it? And it was James Flansburg who said, what else do you see in this poll that looks unusual to you? And nobody could find anything. And he said, so we publish. But to my knowledge that is the closest that we've ever been based on how surprising the result has been. So, kind of part of the thing that I live with is that I've lived through enough of these situations where people thought this can't be right. So, one of my nicknames is the outlier queen. So, I've been there enough times that I can go back and say, well that Dukakis poll turned out to be okay. And this Joni Ernst poll turned out to be okay. But I don't think I was not vigilant enough. That is one of my conclusions.

[Sostaric] So, election polling is a snapshot in time, as you have said, and it's not necessarily meant to predict the outcome of an election. But it's kind of not necessarily covered that way in the media. How much do you think national, local, all kinds of media are kind of at fault in all this for just elevating the importance of polling and doing that kind of horse race coverage?

[Selzer] They can't help themselves. So, they like to argue about it, they like to figure out the spin on it, they like to say well what does this mean going forward? I don't blame the media for being focused on horse race polls at all.

[Sostaric] And when polls are published there is, of course, a margin of error. Do you think that should be emphasized more in polls and in coverage of them?

[Selzer] You know, there's probably a long answer but I'll give you a short answer. I think most responsible media do talk about the margin of error. But I want to qualify that the best estimate of what that race is, is exactly the numbers that we show. And so, if you want to say it's a two-point race, if the numbers show that, or you want to say well that's a statistical tie, it is equally likely that that race is further apart than that. So, I think you just say what the numbers are. You can say that is within the margin of error if you like. But I think there is a pressure to make races look closer than they end up being.

[Henderson] Was one of the flaws in this party, poll rather, the fact that we who cover Republican Party politics continually hear republican politicians and officials criticize the Des Moines Register? Obviously, it's the Des Moines Register Iowa Poll. And some polling nationally indicates that republicans are less likely to answer a polling question. Is that part of the problem with the data?

[Selzer] It could be. We do not mention the Des Moines Register's name until the very end when we're asking if they're willing to be re-contacted by a reporter. So, hello, I'm Ann Selzer, I'm calling from the Iowa Poll. And our response rates are actually better than the national average. So, I think Iowans are a little more wanting to do that. But of course, that is a possibility. And of course, there is a possibility -- we finished interviewing on Thursday night -- that things happened along the way, those final days, that pushed the republican turnout or depressed the democratic turnout, like I'm eager to look at those reports. That happened in 2004 that we showed John Kerry winning Iowa and it was George W. Bush who ended up winning in a tight, tight race. And former Governor Terry Branstad explained to me that over that weekend, after we were finished polling, there was this monumental effort in northwest Iowa to up the turnout among republicans. So, things can happen. I don't know if that's what happened this time at this point.

[Henderson] In November of this year after the poll was released, you announced that this was the last election-related poll that you will do. Some people have doubts that you had that in the making. Now that this result has happened and it was so wildly off of the end result, is there part of you that says I want to do another election poll?

[Selzer] You know, I have been working on this plan and there are key people who know that this was in the works for a long time, including the Des Moines Register, who was very aware that they were going to need to find their own succession plan for the Iowa Poll. So, there are numerous other people that I could point to. It looks like I got beaten and I'm out. I understand that. It's a correlation, it's not causality. Does it make me want to do one more? No. I think I was committed to my plan. Would I have enjoyed being here after a spectacular win and say that I'm out? Yes, I would have preferred that. But this is the way it is and this is my journey to sort of get through and see what it is that I want to do in the future.

[Murphy] If hypothetically you were doing one more, would you go ahead with the same methodology? Or would you reexamine that ahead of 2026?

[Selzer] See, I think that's a question that makes me nervous because there are a lot of polling organizations that redesign their polling methodology after they have had a miss. So, I don't even know what I would do differently if we were going to do one more poll, as I sit here today.

[Henderson] But you did change your methodology in 2008, right?

[Selzer] Well, things happened in 2008. It wasn't an overhaul. But we hadn't needed to weight our data before then. We hadn't needed to worry about cell phones before then. So, there were just some technical things that we did. And I was able to say, you know, how to weight data in a way that would be convincing to me that this is the right way to go. The example is the democratic caucuses in 2008 where our polls said 60% was going to be their first time and it turned out to be 57%. But everybody, all of the politico says it's never been higher than 20%. Why did you assume that? And I responded, I didn't assume. Nobody would decide apriori that it's going to be 60% first time caucus goers. We turned out to be right. And that just said to me, look, what my method is, is to get out of the way of my data revealing to me what the future electorate is going to be like.

[Murphy] In our last few minutes here we wanted to kind of back out and talk big picture. As Kay mentioned at the top, you've been doing this for a long time. I'm curious to get your sense of what value do you believe that the Iowa Poll has provided to the state, to Iowans, to the discourse here?

[Selzer] Well, I think it has been a remarkable tool for reporters. But I think it has also been a remarkable service to the people of Iowa. The Iowa Poll was started in 1943 with fellow Iowan George Gallup coming in and helping them design it. They used to send interviewers out on a monthly basis and pretty much publish a story every day, an item from the Iowa Poll. So, it was launched to be a big deal on the leading topics of the day in Iowa. And I think if you look at our legislative polls to kind of help Iowans understand where their fellow voters and Iowans are on key issues the legislature is debating, I think that has been a very valuable service. I'm proud of the work that we've done.

[Sostaric] Why did you become a pollster?

[Selzer] I knew early on that I liked gathering data. I like that idea of people talking to me and turning that into numbers and then seeing an idea. And when I was in graduate school looking at communication theory and research, I had a fellowship to Washington and worked on Capitol Hill as a congressional fellow. And so, politics and gathering data, it just sort of matched up with my interests. So, I was born to poll and I've been doing it about as long as Iowa Public Television has been in existence.

[Sostaric] Do you feel like after this the polling industry has to be redeemed at all after this election? And if so, how do you do that?

[Selzer] Well, I'm not going to be a part of redeeming the polling industry. I'm privy to all sorts of internal conversations about what's right and what's wrong. And it's kind of why I was ready to be done, that I don't know that there is any consensus in the polling industry. And the polling industry is predicated on getting people to pay money for their products. So, there's a lot of conversations that are self-serving about my way is best because blah, blah, blah, blah, blah, but if you dig into what they're doing it doesn't really stand up. But I can't predict that future.

[Henderson] You started at the Register as an employee. Then you formed your own company, Selzer and Company, in 1994. You have said you are no longer going to do election polling. Will you be closing the company?

[Selzer] I'm not closing the company.

[Henderson] What are you going to do next?

[Selzer] Well, we'll see. I have a couple of clients who got in touch and say, talk to me in January. I've got some things that are going on. Earlier this week I gave a presentation to a local client who said, we're going to need more help next year. We'll see. The support that I've gotten has been truly heartwarming. I had a CEO of a large Iowa company call and his first words to me, we did work for him fifteen years ago, how can I be of service? Because he felt that the work that we had provided for them was invaluable and changed the direction that that company was headed. I enjoy that work. It has been out of the public spotlight, so people don't normally think of me that way. But there are many things that I am thinking about continuing to do without the election side of that and chasing that business.

[Henderson] The person who is the previous moderator of this program, David Yepsen, worked in concert with you at the Des Moines Register, he was their top political reporter and an opinion writer as well. I know that one of his favorite Iowa Poll questions of the 1980s and '90s was, do you believe in God? Do you have some favorite questions or favorite Iowa Polls in the past that have been insightful or just something that sticks in your memory?

[Selzer] Well, thank you for that question. I agree with David, that was an interesting question. I don't know what the answer was right off the top. But we used to call these vegetable polls. So, they're not issues so much as they are akin to an early question asked about here are some vegetables. Is this a favorite? Is this not a favorite? Do you hate it? And what came out was lettuce was -- they said the favorite vegetable, I would say the least objectionable vegetable that way. But we had a question about whether you eat your corn on the cob horizontally or in a vertical sort of fashion.

[Murphy] I'm sorry, I have to know the winner of that one. Do you remember how that went?

[Selzer] No.

[Murphy] Oh, no.

[Selzer] And it might have been in me file of would-be poll questions. But corn on the cob, of the foods related to Iowa, number one. Then pork chop. Then third, surprise, Muscatine melon. So, I think of these questions as kind of revealing what I call the soft underbelly of the Iowa psyche. So, we asked a question about do you own a pair of bib overalls? Here is the national image of Iowa. And it was I think thirty some percent. Do you think bald men are as attractive to men who have a full head of hair? Would you marry the same person again? How would you rate your marriage on a scale of 1 to 10? And 45% gave their marriage a perfect ten. Now, is that a big number or a small number? And those sorts of questions that are just good for a good conversation over coffee.

[Henderson] Well, we're not drinking coffee here, it's water, but our conversation at this point is over. Ann Selzer, thanks for being here today.

[Selzer] My pleasure.

[Henderson] You may watch episodes of Iowa Press at iowapbs.org. For everyone here at Iowa PBS, thanks for watching today.

(music)

Funding for Iowa Press was provided by Friends, the Iowa PBS Foundation.

(music)

The Associated General Contractors of Iowa, the public's partner in building Iowa's highway, bridge and municipal utility infrastructure.

Elite Casino Resorts a family run business rooted in Iowa. We believe our employees are part of our family and we strive to improve their quality of life and the quality of lives within the communities we serve.

(music)

Across Iowa, hundreds of neighborhood banks strive to serve their communities, provide jobs and help local businesses. Iowa banks are proud to back the life you build. Learn more at iowabankers.com.

(music)