Charles Franklin on Conducting Political Polling in 2024

Frederica Freyberg:
For months leading up to Election Day, we hear from Marquette Law School poll director Charles Franklin, whose poll was recently ranked number three in the nation. As polls dominate the news, we wanted to go behind the scenes to better understand them and the director behind the data. Marisa Wojcik started by asking how the polls are conducted and what has changed.

Charles Franklin:
From 2012 until ’20 through 2022, we did them all on the telephone, randomly dialing phone numbers in the state, both landline and cell phone. But that became increasingly difficult because more and more people don’t pick up phone calls. So beginning in 2023, we shifted to a sample based on registered voters. The list of registered voters you can buy from the state. And that allows us to send an email and a text message inviting people to take the survey online. Only the person we’ve sampled can take it. They can only take it once. You can’t volunteer to do the poll. So it’s still a random sample of the population. It’s just that we’re not randomly sampling phone numbers. We’re randomly sampling people we know are registered voters. Now about 80% of people complete the survey online, either on their phone or tablet or whatever is convenient, and about 20% if we haven’t heard back from them in a couple of days, we start phoning. So we complete those interviews by telephone.

Marisa Wojcik:
How has that changed the data or has it changed the data?

Charles Franklin:
Not in exorbitant ways but there are some subtle ways. There are some advantages to online in that you can ask a slightly more complicated question online that might sound really hard to understand when an interviewer is reading the question, but just reading it, it’s pretty clear to the respondent. The other thing is, we know a bit more about the person that we’re trying to interview. Because we sample from the voter list, we know where this person lives. And so we can be geographically more precise in our sample. When we were on the telephone, we divided the state into five big regions based on media markets in the state. Now we divide the state into 90 different regions because we know where these people live. And so that means those regions are defined by going from the most Republican to the most Democratic, say, within the Green Bay TV market. So we’re making sure we get the right percentage of people from the most Republican areas of the Green Bay market and the most swingy areas and the most Democratic areas. I wanted to show you Vernon County. So that’s one of the changes we’ve made to try to make sure that we’re representing the full range of opinion. It’s just really striking how much diversity and heterogeneity we have in opinion, even within a pretty rural Western Wisconsin county.

Marisa Wojcik:
Would you say that this makes polling more accurate or better?

Charles Franklin:
We think it does. The reason for making these changes is in part in response to the polling errors that we had in 2016 and others had in both ’16 and 2020. But the proof will come in November when we finally get the vote count, and we’ll see how close we are.

Marisa Wojcik:
2016 and 2020 was a tough set of cycles for many pollsters, not just the Marquette poll. Some of the explanation for that was people who wanted to vote for Trump didn’t feel potentially comfortable telling that to someone on the phone. Does that mean that kind of might be mitigated?

Charles Franklin:
I don’t think it’s so much the so-called shy Trump voter who somehow doesn’t want to admit they’re supporting Trump. Go to a Trump rally, there are not a lot of shy folks at a Trump rally. What I think it is, though, is that Trump appeals to a set of voters, some of voters, who have not traditionally voted in elections but are being motivated to vote because of him, and who are very skeptical of mainstream media, the established political parties and pollsters. So I think where the problem comes is not so much unwillingness to say who you’d vote for, as it is skepticism about doing a poll at all and therefore falling out of the sample.

Marisa Wojcik:
Beyond turnout, do polls ever convince voters who to vote for, or how they should feel about an issue?

Charles Franklin:
I’m skeptical of that. What I would push more is the value of seeing what the polls say about the people in your community, your state, who don’t share your views. It’s so easy to be trapped in the notion that everybody I know agrees with me on this. This is the obvious majority. And yet the poll can tell you that maybe 30%, 40%, maybe even 60% don’t hold the same view as you on that. It’s a bit of a pipe dream to wish people would use polls that way, but I really think that one of the most valuable points of polling is to talk about policy issues, not just candidates. And to try to recognize the wide range of opinion people hold on those things even within a county or certainly within a state that not everybody’s on your side on this.

Marisa Wojcik:
How should an average voter use this information? It can feel a little overwhelming. It can feel like there’s horse race polling and how do we get value?

Charles Franklin:
One fairly serious answer is try to ignore it all. Make up your own mind. Talk to your friends and do that. I think that’s a totally valid way to follow this.

Marisa Wojcik:
How has this election really shaken things up for you?

Charles Franklin:
As for me personally, I have rewritten so many survey questionnaires this year and this summer, especially on the last night I can make changes before we start interviewing. When Biden dropped out, I had to hurry to rewrite the questionnaire for Harris rather than for Biden. So it’s been a fun time for me that way.

Marisa Wojcik:
What do you think is the most important thing for voters to think about in this election as a whole, not in polls?

Charles Franklin:
I think I really go to the question of, this election is important, no matter which side you’re on, and paying attention to the candidates and making your own mind up and treating it as seriously as you can is really my main advice. It’s not about the polls, per se. It’s not even narrowly about how the candidate’s campaign. But I think this is just the political scientist in me. I think we’re a better society when more of us vote. We’re a better society when the majority wins. When we release our final poll before Election Day, the last thing we say is, remember, polls don’t vote. People do. The election’s in your hands, in the voters’ hands. And I still think that is the most important lesson.