We have a presidential election between two candidates that few Americans wanted to see. As much as people argue and complain about polling, it’s one of the tools we have to understand what’s going on in the United States, and how politics has changed and could change further. What does Donald Trump’s base look like now, exactly? Who counts as an independent? When voters say they wouldn’t vote for Trump if he were convicted of a crime, should we believe them?
I spoke with David Byler, chief of research at Noble Predictive Insights and a polling expert and former writer at The Washington Post and The Weekly Standard, who told me, “We’re in this era where candidates are just always going to be in a position where one of them could catch up. We’re too polarized for anything else.”
This interview has been edited for length and clarity and is part of an Opinion Q. and A. series exploring modern conservatism today, its influence in society and politics and how and why it differs (and doesn’t) from the conservative movement that most Americans thought they knew.
Jane Coaston: Is the MAGA base more or less powerful than it was in 2020 and 2016?
David Byler: The MAGA base is more powerful than it was in 2016 because it’s acquired other parts of the Republican base. You can look at the difference between Trump’s numbers in this primary and the primary eight years ago among very conservative voters. There are groups that believe in conservative ideology or there are demographic groups like white evangelicals, especially church-attending white evangelicals, who once harbored skepticism toward Trump, that now have been folded into the MAGA wing with populists who were already there. Because there was no real primary in 2020, we don’t have a clean one-to-one test in that way, but I would say that Trump trades some of these suburbanite Republicans for other voters, sometimes Black and Latino voters. According to some polls, you’re getting a more and more MAGA Republican Party with younger Republican voters.
Coaston: So you mentioned people who are shakier Trump voters, people who, as you mentioned, considered a different Republican during the primaries or may have. Basically, what are the potential problem areas for Trump?
Byler: One really direct way to get at this is to ask people whether they’re more of a supporter of Trump or more of a supporter of the Republican Party. Maybe they live in the suburbs, maybe they’re college educated. There’s something that pushes them away from the current Trumpian G.O.P., but they hold conservative beliefs at the same time.
A lot of them, I think, come home to Trump by Election Day just because of polarization. We’re in an era where both major party candidates basically start out with 40 percent of the vote no matter what they do. You might call ’em reluctant Republicans, you might call them party-first Republicans. They’re different from the people who left the G.O.P. circa 2015 and never came back.
Coaston: Where do independents fall here?
Byler: It depends on whether you’re talking about true independents or independent-leaning Republicans. The elections are won and lost on true independents, which in this definition are people who you ask which party they favor and they say they’re independent. And then when you press: OK, but where do you lean toward? They still persist. A lot of them are the least tuned in to politics. They’re the sort of people who may start listening to the news six, seven months from now. In some of our surveys, Trump is winning these people, but by virtue of them being true independents they’re just less attached, they’re more able to be swayed.
Coaston: What’s the most important thing going on with conservative voters in this presidential election that you are seeing but you don’t think is getting enough attention or isn’t understood well enough?
Byler: This is a good question. Some of the racial realignment we’ve seen in recent elections and in current polls is an ideological alignment as well. Part of what’s happening to the Republican Party is that they’re gaining ideologically sympathetic voters. They’re gaining conservatives from demographic groups, be it Black voters, be it Latinos who had views that lined up with the G.O.P. already and are coming home to their natural ideological home.
Coaston: In 2022, you wrote that Trump had lost ground in the Republican presidential field and looked weaker than he appeared. To quote: “Trump used to take positions that helped him stand out from other leading Republicans. But he hasn’t done that for the 2024 primary. He is focused on the ‘big lie’ — an issue that’s less potent than it appears — and allowed Trumpian alternatives such as DeSantis to gain ground.” What has changed, or do you think that he’s still somewhat weak as a nominee?
Byler: Oh boy, that piece. Given what I knew at the time, I think that was solid analysis, but that did not age as well as some of my other writings. My big picture view of what happened in the 2024 primary was that Trump really did start out from a place of weakness because of the 2022 midterms. I think that he was correctly blamed for elevating candidates who had lower electability. A lot of people were seeing some of the antics of Trump as not particularly worth it. Then I think two things happened. One is just the simple passage of time: As the 2022 election receded from voters’ memories, they saw Trump as electable. And he settled into this position of incumbency where Republican voters who, by and large, incorrectly believe that he won the 2020 election, started to see him as electable.
The other thing is that Ron DeSantis really failed to capitalize on that opportunity and fell very, very far short of how he looked on paper in late 2022, early 2023.
And then there’s just one more detail here. This was something revealing. It was in a poll we did in January 2024, before the Republican primary was over. We took Trump’s primary voters and gave them the open-ended opportunity to say, “Hey, you voted for Trump above these other Republican alternatives, not even mentioning Joe Biden, just in the context of a primary.” And we said, “Why’d you pick Trump?” Just let them say what they wanted. Now, predictably, a lot of these responses were short. People would say, he’s my guy. He’s the one I like the most. But what really stood out when I read all, I guess, 600-something of these responses was that they thought that Trump did a good job. His primary voters thought that he performed well as president.
So if you put that data point alongside the point that Republicans by and large believe that Trump did win the 2020 election, you have the recipe for a successful primary candidate. You have someone who Republicans believe is going for a three-peat victory and they know delivers for them on policy. There wasn’t a lot of room for somebody else, in retrospect. And then DeSantis, and you could argue Tim Scott, maybe the only two people who had the right profiles to win over those voters, just didn’t make the most of any opportunity that might’ve been there.
Coaston: So I want to talk about the 2020 election denial. It seems that those who believe the 2020 election was illegitimate have hardened in their views, and you mentioned that, but is that true? Is that a real belief?
Byler: I think this is a real view. The 2024 primary might have panned out differently if voters were saying, yeah, Trump won, in an effort to just stick it to Biden, and then they really believed in the back of their minds that he didn’t win, and were freaked out about Trump not being able to win in November 2024. There would’ve been a bigger opening for a DeSantis or for a Nikki Haley or a Scott or whomever.
For a lot of voters, Donald Trump is a trusted source of information. If Democratic leaders say something, Democratic voters also tend to follow. You could argue the strength is different. But there’s a situation here where Trump has broadcast the 2020 election denial widely, and a lot of Republicans take him at his word for it. If you look at those same polls that we’re talking about, a nonzero percentage of people who self-identify as Republican will say, yeah, Trump lost.
Coaston: In polls this past year, a number of people have reported that one thing that might keep them from voting for Trump when they otherwise might is if he’s convicted of a crime. First of all, how do you think about those results and people’s predictions of their own behavior? Is that something polling can even measure?
Byler: The first question is: Are people good at predicting their own behavior? The second is: Is this a real liability for Trump? I would say no, people are not great at predicting their own behavior. There’s probably some social desirability bias here. People would say: Oh, I’m a respectful person. I’m a reasonable person. Of course I would never vote for someone who’s a convicted felon. And then the actual events unfold and they hear messaging from both sides, including people that they’re sympathetic to. Some of these people who are projecting that they would not vote for Trump end up coming back around to him.
The 2016 election is sort of the perfect case study for this because you had so many statements from Trump that were wild and that were at that time basically unprecedented for a national politician, and you had people leave him at first and then come back. People are not always great at predicting their own responses to events.
What I would say is that at the same time, a conviction would be a liability for Trump. And the reason that I think that is because if you look at sort of deeper questions, not just “Who do you prefer?” but “Who do you prefer, Biden versus Trump on issue X, Y, Z or on character traits A, B, C, D?” Biden and Trump have very different strengths.
You can look across different polls and see that on questions around attributes like honesty or integrity or things in that vein. Biden is often a winner. So if a court case were to go badly for Trump and suddenly corruption or morality or some issue where the candidate’s personal morality is injected into the race, I think that is bad for Trump. I think that is good for Biden.
Coaston: That polling result about not voting for Trump if he’s convicted often seems to be about Jan. 6 and the cases related to his time in office. Do you think those polling results also apply to the Manhattan hush-money case?
Byler: I don’t know. I think that’s a good question.
Coaston: If there is a Jan. 6 trial this year, which at this point is unlikely, are there any historical or political events in terms of shaping public opinion that you think could compare? Or would you expect it to be more like a standard polarized political event?
Byler: I think the experience of Covid and the 2020 election is informative. I’ll be honest, I thought Covid would affect the race a lot more than it did in the end. It was a truly earth-shattering, unprecedented event where one person, Donald Trump, was in charge of essentially emergency response, and the other person, Joe Biden, was not. A lot of the events that would be comparable in magnitude occurred during a time when there was less polarization. When we’re talking about the effect of a Trump case, I think Covid is in some ways our best precedent. It should have shaken up an election and honestly probably did harm Trump — but it didn’t turn the entire thing upside down into some double-digit landslide.
Coaston: There’s been so much change in communications technology and polling in the last decade. What do you worry that polling is missing now?
Byler: Response rates have been decreasing for a long time. You can go years and years to when they had response rates we’d only dream of right now, and people were still biting their nails, saying, “I don’t know. Is this a sample that we can trust?” This worry has been around forever, and so far, election polls are kind of still good enough. We haven’t had a sort of true earthquake that can’t be recovered from; 2016 and 2020 were not great, but they weren’t so abnormal that you’re shaking the foundations of the industry. I guess I’m just worried that if the response rate goes down further, is there a tipping point? Do we hit that at some point?
Coaston: Should voters and readers trust polling?
Byler: Define “trust.”
Coaston: See polling as factual data?
Byler: I feel like there’s a pretty well-rehearsed spiel that people in my world give about this, which is that polling is imprecise and that there’s always a margin of error. And all of that is true. We are interviewing people and we are absolutely making our best effort to figure out what a representative slice of a certain population believes, thinks and wants to do in a given moment.
We’re in this era where candidates are just always going to be in a position where one of them could catch up. We’re too polarized for anything else. There’s never going to be a 100 percent certainty on a probabilistic model given how polarized things are, and polling is going to reflect that. In terms of election polling, should people trust that one candidate is ahead or one candidate’s behind? Yeah, I think that’s the best estimate that we have right now, and I think that could change, and I think there’s big uncertainty around it.
But the other thing here is that if we don’t trust polling, what exactly is our alternative? And in a lot of past elections, polling has beaten the vibes. With a poll, I can ask people: “OK, do you think abortion should be illegal in all cases? Illegal in most cases, legal in most cases, legal in all cases?” I don’t really know how I would glean the same thing from looking at Google search data. When it comes to trust, I would say polls are what they advertise themselves to be. We’re really interviewing people and we’re really trying to make the population representative. And if you want answers to these questions, it’s kind of the only game in town.
