Election expert Charles Stewart III, a professor at MIT, joined the podcast to discuss the results of his Survey of the Performance of American Elections, asking voters in every state and Washington, D.C., about their voting experience. His 2024 report from that survey offers a window into how Americans voted—not who they voted for—and if they were satisfied with their voting experience.
Since 2008, Charles Stewart III, a professor of political science at MIT, has conducted the Survey of the Performance of American Elections, asking voters in every state and Washington, D.C., about their voting experience after every presidential and midterm election. His report from that survey offers a window into how Americans voted—not who they voted for— and if they were satisfied with their voting experience.
Stewart joined the podcast to offer a preview of the 2024 report, that includes questions about whether voters cast their ballots by mail or in person, how confident they were in the election results and other matters.
Stewart, who is also the head of the head of the MIT Election Data + Science Lab, said some people might be surprised to learn that most voters are both satisfied with the voting experience and have confidence in the results.
Resources
Ed: Hello and welcome to “Our American States,” a podcast from the National Conference of State Legislatures. I’m your host Ed Smith.
CS: Nobody had ever talked to voters about their experience when they voted. So, we thought it was really important for somebody on a systematic manner to talk to voters about their experience on Election Day and the things around voting to see what they thought.
Ed: That was Charles Stewart III, a professor of political science at MIT for the past 40 years and head of the MIT election data and science lab. He is my guest on the podcast to talk about how people voted in the 2024 election and how they felt about it. Since 2008, Stewart has surveyed voters in every state and the District of Columbia to find out whether people voted in person or by mail, how confident they were in the election, how satisfied they were with the voting experience and other matters. Stewart joined the podcast to discuss the findings from his most recent survey. Some findings might be as expected, but others might be a little surprising such as the high level of satisfaction people had with the voting experience. Here is our discussion.
Professor Stewart, welcome to the podcast.
CS: Oh, it’s a pleasure to be here.
Ed: Professor Stewart, thanks for coming on the podcast. I wonder if you could start by explaining the genesis of your survey of the performance of American elections, how it works, how many people you surveyed – that sort of thing.
CS: The survey of the performance of American elections which I call S-P-A-E. Some people call it just SPAE which I dislike so I’ll call it S-P-A-E. It started in 2008 with some pilot programs before that, but the genesis of it came from conversations we were having in the voting technology project …, which arose out of the you know recounted Florida in 2000. We ended up in a really interesting interdisciplinary group of social scientists, designers, political scientists, electrical engineers, astrophysicists – all sorts of people looking at elections. And across the two campuses for many, many years. In fact, that project continues to this day, but in smaller footprint. One of the things we learned in doing our work was during the thoughts during that decade there was oftentimes a sense among policymakers, media, among voters that reflecting on Florida and the recount of what had happened there. America where it wasn’t really anything more than a banana republic. Elections were fraught. Voters were being abused and having a horrible time with things and the thing that occurred to us was that first of all we could quantify and agree to which certain problems did exist. For instance, some kind of generalized counting out the number of people who might be affected by hanging chads and those sorts of things. But that was a relatively small number. I mean it was significant when you rolled it up. It could be a couple of million voters in an election. In an election with 120 million voters though that is a small percentage. So, we could count up those problems, but nobody had ever talked to voters about their experience when they voted. So, we thought it was really important for somebody on a systematic manner to talk to voters about their experience on Election Day and when the things around voting to see what they thought. So that was the really the genesis of this and we received funding from the Pew Charitable Trust, which was in the area of philanthropy at the time. The Jet Foundation which ended up going under the Bernie Madoff scandal unfortunately. But we got the check before then. That’s when we did the first S-P-A-E and that was in 2008 and we’ve done it in almost every federal election since then. A couple of holes or pretty much every federal election since 2008. We structured the surveys so that it really emphasizes us being able to compare across states because one of the goals in addition to assessing just the general tone of voters and their experience was to be able to say something about whether this state was performing better than that state, whether implementing a particular reform in the state would affect the voter experience or turn on those sorts of things. And so, we ended up with a unique design in this space which is that we don’t do a national sample of voters rather we do 51 samples one for each state plus the District of Columbia of 200 voters in each state, which allows us to first of all 200. Some people think gee that is not really big, but it is big enough to do the sort of work we are interested in if we wanted to compare between states. And then we are able to roll up all of those responses into a national study if we want to of 10,200 voters.
That ends up being the largest survey by far of anyone commercial or academic looking at election administration and other issues like voter confidence and things that come up around election. That’s the long version. The short version is that since every federal election almost since 2008, we’ve gone into the field. We’ve asked 200 voters in every state plus DC about their experience on Election Day so this includes literally Election Day, the people who voted by mail, people who voted early in person. We asked a series of questions ranging from how easy was it to find your polling place. Did you have problems when you registered or went to the registration table, when you filled out your ballot, when you used the machine. How good were the polling officials. Those sorts of things and was the polling place run well. So, we asked kind of these questions of kind of quality assurance and then we ask other questions that are related to voting that aren’t core but give us insights into the voting experience. For instance, we asked voter who didn’t vote well why didn’t you vote. We asked opinions about reform about police about fraud. We will add questions that are topical. For instance, the last couple of elections we’ve asked voters to give us feedback about what they think election officials are doing to secure the vote.
In 2020, we asked in person voters what sort of precautions they observed when they went to vote in polls to keep infections down. We’ve asked about violence in the polls. This time, we’ve asked about whether people have voter ID’s. If they don’t have voter ID’s, do they have passports or birth certificates or those sorts of things. So that we can you know we can address a number of policy issues as well as the experience of voters. And then finally the thing I would say is that this is a survey that was put together by a bunch of academics so we also you know we pair those attitudinal questions self-reports about what was happening when they voted with demographics. We know the state, county, zip code where people live and so we can link all of this data to demographic, economic, political context in which voters are voting. That’s a lot for your listeners. But I think it is important to say that this is kind of the big robust survey of voters that can inform the nation, can inform policymakers at a state level. Can really be informative about a number of different issues that have to do with election administration.
Ed: Well, I read the 2022 report and you are right, it’s a lot. I wonder for the 2024 report if you could talk about what the topline finding is. What is your elevator ride description?
CS: The three toplines I would emphasize one is something we’ve seen every election and that is the voter experience was a positive one overwhelmingly regardless of whether you voted in person or whether you voted by mail. You know when we asked people for instance if they had problems filling out their ballot or with the machine, the percentage who said they had a problem was 1 or 2%. So, the percentage of people who say they have problems when the vote is really, really tiny so this is a long elevator ride. So, the experience is good.
The second one for 2024 was that not surprisingly but we quantified to a degree to which voting methods that mix between voting in person, voting by mail, voting early, voting Election Day it continued to evolve in 2024. We talk about it later with about the details but some interesting both trends, long range trends have been reestablished, but there are some short-term interesting things that we can talk about later. But this is an elevator ride and then finally there is a matter of trust. And there again, there is a lot more to say about this, but the short answer is that levels of trust at least reflected in these survey questions Americans are the most trusting of these election results that I have ever seen not only in the S-P-A-E, but in the commercial polls that preceded it going back to 2000. So, in one way, you could argue that this 2024’s voters were the most confidant that we have ever seen in the results of a national election.
Ed: Well, we will have to dive into that a little more. It’s very interesting. Let me ask you before we go into some of those details just what is the biggest difference you’ve seen in the voting behavior between mid-terms and presidential election years.
TM: 10:49
CS: Well, I mean I’ve looked at the turnout for probably actually every election going back a couple of broad things. Going back to sort of the beginning of the 20th century to the present, mid-term elections have had turnout rates of about 60% of presidential elections. That’s the main thing. And some other just some of quick factoids, there is a difference in the electorate between mid-term and presidential elections. The mid-term election is a more experienced electorate. It includes people who voted in a reelection kind of not surprisingly it’s generally an older electorate. It used to be a more republican electorate although that may be changing. That has consequences for the experience of voters compared to the presidential electorate which you know surges in people who are younger, first-time voters. People who only vote in presidential elections. If you vote in only kind of election, it’s generally a presidential election. That’s a less experienced voting crowd, people who are more likely to have problems, people who are actually more likely to vote on Election Day. There are these you know differences. The final thing I would say is that we’ve seen a pretty steady increase and turnout at the presidential level since the 2020 election so that now we are at a level that’s as high as it has ever been since say the passage of the 19th Amendment which brought women into the electorate. And that the midterm levels because you asked specifically about midterm on voting, it really started sagging kind of dramatically significantly in the late 60’s, early 70’s, but that has kind of come back and so we are now again it’s about 60% of the presidential, but considering the presidential turnout is at historic levels. And also means with the exception maybe of what the 2014 midterm. Midterm turnout now is also approaching historic levels as well.
Ed: Let me ask you about voting by mail. From 2000 to 2022, as I understand it, there was a significant drop in voting by mail. I wonder if that happened again from 2022 to 2024.
CS: Well, you know, it is really interesting and actually caught me a bit by surprise, we don’t have the PowerPoint slides so I will try to navigate the slide. So, if we go back into the late 1990’s if we trace how the rise in voting by mail, it was a steady rise from you know around 10% of voters voting by mail back in 2000 to something around 25% or so voting by mail by 2016 and 2018 in there and it was just a couple of percentage point growth. It was steady and there’s a story to be told there really kind of organic growth states one by one are changing the laws to make it easier to vote by mail. A couple of states then began to have all vote by mail. I describe it as being organic. And then the 2020 election happened. It created this big shift and so that for the first time in American history, roughly half of Americans voted by mail. And we know the story there which was people were trying to avoid the big story was people were trying to avoid crowds and so they wanted to vote by mail. The smaller story was that there was kind of a political story such that democratic organizers were pushing Democrats to vote by mail. There was some countermovement among Republicans to discourage people from voting by mail, but on net it was the democratic organizers I think that had the bigger effect so a big increase.
So, then the question is we are waiting to see what was going to happen in 22 and what happened in 22 was this if we were to take remember I mentioned kind of this slow steady increase from 2000 up until 2018 of voting by mail. Imagine if 2020 had never happened that we just continue that line past 2020 to 2022, percentage of people voting by mail in 22 is about what we would have expected if the pandemic had never happened. It was just a couple of points above that. 22 there was this crash in voting by mail. What people don’t oftentimes don’t recognize is that the big crash happened among Democrats, right. They were the ones who got surged into it so they surged away. The interesting thing in 24 is that the percentage of voters voting by mail actually you know kind of increased a bit above trend and there the increase came among Republicans. And we know that there was a movement among some Republican strategists saying that you can’t weapons, good weapons on the table and so you just got to use everything that you have. And so that was kind of the wisdom there and so more Republicans came back to vote by mail and I think that’s what kind of kept the numbers up. But nonetheless, I mean voting on Election Day reestablished itself as this common way to vote, but now it is kind of neck in neck with voting by mail. And we will have to see now that it looks like the politics of voting by mail is beginning to settle down a bit. You know it could very well be that the convenience factor of voting by mail the strategic advantages that both Republicans and democratic candidates see in locking down their vote before Election Day will come into focus. We can very well see a slow, but gradual rise in voting by mail in 26 and 28.
TM: 17:02
Ed: Let me follow up on that about Election Day voting. My understanding is that in-person voting is still the most common way to vote whether that’s on Election Day or at a voting center or something like that beforehand. How did that turnout in 2024? Was it back to 2016 levels?
CS: Well, you know it’s actually kind of halfway so it was about 40% on Election Day from what I’ve gotten in this survey. It might be updated later on. It was about 40% on Election Day which in 2022 it was something so if it was about 40%, it was about 45 or 50% in 22 so there really was actually a drop in Election Day voting compared to 22. It’s just that there was more compared to 20. And I think the important thing is that it is no longer on the it didn’t return to a majority. So, it’s really 40/30/30. So, it’s 40% on Election Day, 30% by mail and 30% early in person. That’s nationwide. It certainly varies by state, but nationwide that’s how it is.
Ed: I’m sitting here in Colorado so my perspective is entirely skewed by the fact that I’ve voted at my kitchen table for quite some time now. You talked to voters about their consumer experience, if you will. How did people respond whether they voted by mail or voted in person?Did they feel like they had the options they wanted? Were they satisfied with the experience? My understanding is pretty much yes from what you said earlier.
CS: Yeah, yeah, no as I said earlier that you know everything that we asked people about, they are feeling good. The one caveat about that and then one exception. The caveat of course is that my guess is that the voters who voted on the hanging dimpled chads of South Florida probably thought they were having a good experience too and so you know we need to always temper the surveys with other metrics of outcomes. And so, take this as a starting point, not an ending point. So that’s one caveat of that. The contrary point that I’m trying to socialize is that lines were longer in 2024 than they were in 2016 and I’d like to use this as an opportunity to remind people that after the 2012 election President Obama you know identified as people waiting in line to vote as being a problem. He appointed the bipartisan presidential commission election administration to study that problem and other issues of the election administration. The PCEA came back with a benchmark saying that people shouldn’t have to wait more than 30 minutes to vote and the percentage waiting more than 30 minutes to vote in 2024 was of course less than in 2020 because it spiked up for a variety of reasons, but it didn’t come down to where we had landed in 2016. In 2016, we really got the weight times down. There was a lot of effort among local election officials to work online length and so line lengths now are about where they were in 2012 again. I’m hoping that as people find this out that state legislatures and state and local election boards and election officials will turn their attention to what’s going on here. Certainly not a crisis, but it certainly doesn’t appear to be of the same form of long lines as before, but it’s you know it’s a performance area where it looked like we had kind of licked this thing and I think we are back to kind of square one or square one and a half.
Ed: Let me ask you for those of us who are vote by mail people about ballot tracking something I find extraordinarily efficient. What do other people think about that? What did your survey tell you?
CS: A couple of interesting things. One is it looks like the percentage of people using ballot tracking doubled from 22 to 24 from among those voting by mail from about 30% to about 60% which I had to doublecheck to make sure the data were right and they appear to be right so a lot of more people using ballot tracking. I think part of that is in the states that have decided to go all vote by mail or have a large number of voters voting by mail, they recognize the importance of the ballot tracking software websites, etc. and so there are just more opportunities for voters to use ballot tracking. I also think that ballot tracking is useful and you see this in the survey because of concerns about voter security. And so, people are motivated both by a desire that their own ballot be received and be counted, but also people who are also just concerned about the possession of ballots and whether they actually make it to the counting screen. That security concern is also driving some people to use ballot tracking just to make sure that you know there’s the flipside of this being wanting to make sure my ballot gets there is this worry that someone is going to steal my ballot or someone is going to steal other people’s ballots and so we have to be attentive here. It looks like the platforms that are being used are across the board kind of equal numbers or getting text messages or going to websites, getting emails and so it looks like there is no one way of getting ballot tracking. Coming from Colorado, you know you can get the application that basically tells you everything about your mail ballot from the moment it leaves the election department to the moment it returns back there and has been accepted and gone into the accounting stream. There’s that. There are other places where the amount of information is more limited and so I mean that’s something I think in the policy area ought to be looked at.
Ed: One thing you mentioned before, which I find interesting, is people who didn’t vote and why. I had a friend years ago who always wore a button on Election Day that said don’t vote it only encourages them. It’s not a sentiment I particularly agree with, but it does make me curious about those people and what did people say about why they didn’t vote?
CS: A couple of things. The first thing actually that has been said is that in a survey like this, about half the people who didn’t vote tell the survey that they voted which tells you something about the social desirability of voting. And although your friend kind of wore their non-voting on their sleeve or their lapel, roughly half the folks are kind of ashamed that they didn’t vote. It’s called social desirability by us in survey research. By the way, there is one feature of this survey that I should mention. We buy a module or we contract with the survey company where they would go out and they are doing it roughly right now that all of the voter registration lists have been updated. They know the identity of the people in the survey. I don’t, but they do. They are going and they are going through those voter registration lists to tag everybody who actually voted and that’s why I can tell you what fraction of people who said they voted actually didn’t vote. Among those who were willing to admit they didn’t vote; it’s kind of a dog stew of reasons. The biggest reason was I’m not liking the candidates. The top two typically are not liking the candidates, not being interested in politics which I consider to be kind of very, very similar and that’s about a quarter to a third of the non-voters. And then you begin to get into areas where it’s like 10% will say they were sick. Another you know 8% will say they were too busy. A few people will say they were out of town. Those sorts of things.
One of the cautions about this study or any other study on non-voting is actually any study of anything where you ask people are there reasons for doing things is that we know that people don’t know why they do things. If we turned around and asked people who actually voted why they voted, they probably they would give you an answer but whether that was “the real answer” we don’t know. But to the degree we can tell, it’s a good number of people just aren’t interested in politics. They don’t like politics. They don’t the candidates. They don’t like something about it. And then a lot of the others either rationalize or they you know say out of town, sick, forgot. I’m always amazed you know 4% who will say I forgot about the election.
Ed: You mentioned before the issue of confidence in elections. I wonder if you could talk about that in terms of what people’s attitudes are about the competency of elections, whether they think the election is administered correctly. That sort of thing.
TM: 26:36
CS: When you look at surveys of my type, we see a couple of general patterns and we saw this in 2024, I would identify two really important patterns. One is the closer the target of voting is and I’ll explain what that means, the more confident you are. By target I mean if I ask you how confident are you that like your vote was counted as you intended or your vote was counted fairly or something like that. Something like 90% will give some flavor of confidence. If I ask you how confident are you that the votes in your county were counted properly or fairly or something like that, it would be like 80%. And then 70% for the state and 60% for nationwide. Also, as we go further away from you, the fraction of people who are confident become less and less likely to say very confident and become more likely to say somewhat confident. But the point I should just note here is that even in 2020 in which a lot of people got worried about voter confidence, most Americans still responded that they were some flavor of confidence in the votes counted in the 2020 election. Some like 60%. That’s thing number one. Thing number two is that if you look across time you know I’m thinking since I’ve been asking these questions in 2008, but I’ve also done research where we look at these questions when they were asked before 2008 starting 2000 now until the present. And that the overall level of all of those answers has been pretty constant across the last couple of decades, a quarter century. What has varied and this gets into you know the next point, what has varied has been patterns of partisan confidence. So, the and the one that people will resonate with the most I think is think about what happened in confidence from say 2012 through 2020 into 2024 where if you asked people the question about the national level, you know you had 60% roughly who would always say confident of some sort. But in 2012 when a democrat won the presidency, it was Democrats who might have been at 70 and Republicans might have been at 50. 2016 when the republican wins the presidency, those numbers flipped. 2020 democratic numbers went up a bit when a democratic won, but republican numbers really crashed down to about 30% so there was a dip in 2020. But because democratic numbers went up, it wasn’t as much of a national dip overall as you might imagine. And then in 2024, democratic confidence came down a bit and republican confidence rebounded quite a bit. And so that’s what we see at the national level in 2024 and when I say that we are at historic levels of confidence by this measure, what I mean is that this really high level for the Democrats. Sometimes I say Democrats what’s kind of lost in these patterns in recent years is Democrats have kind of become irrationally exuberant about how good election administration is like election administrators can do no wrong. And so, their confidence has really stayed really historic highs for Democrats even when they came down in 24. And so, when republican numbers come back up to meet where they were before, those two things together pushed the national number up into historical levels. And that’s similar patterns we see when we ask about the state, county, your own vote. Just the numbers don’t move as much.
Ed: When people say they are not confident, do they think there’s fraud or incompetence or voting machines that don’t work? Do you get a sense of why they don’t have confidence?
CS: You know that’s one of those things where we need more and better research. I’ll say there are two things we do know. One is that the people who say they are the least confident also are more likely to believe there’s a lot of fraud for instance. And so there appears to be a relationship between the two. But we don’t have any good what’s called panel studies where we observe people over time such that say Republicans in 20 you know was there an increase in the number of Republicans who thought there was a lot of fraud and so did the confidence tank because there was a big increase in fraud or was it that people Republicans even those who believe there wasn’t a whole lot of fraud thought there was something wrong in 2020. And it actually looks like this ladder that is to say the percentage of Republicans who believe there was a lot of fraud in 2020 went up a little bit, but didn’t go up enough to account for the decline in Republicans becoming less confident. And so, my hypothesis which needs to be tested is that in normal times and let’s leave 2020 out of this because I think it’s a special case and “normal times” there’s a mixture of reasons why some people will become less confident and what we know is that’s that belief doesn’t last very long. And so, I think it’s probably just kind of a disappointed so it’s more probably as psychological sense among some people being disappointed and they will answer in a certain way.
In 2020, there were a couple of things going on there. One was the psychological disappointment. The second was a belief that there were policy changes made in the election that were illegitimate. I think for those of us in the commentary at kind of the chattering classes, we oftentimes fail to distinguish different reasons for the loss of confidence. I would argue to be legitimately unhappy with some of the policy changes that were made in 2020. It’s not crazy to say ah if not for those policy changes, the outcome would have been different. I don’t believe that. I think evidence is contrary to that. But that’s not a crazy hypothesis to be made. That’s different from other things like a strong belief that you know things were stolen. You know it’s a lot of different things that get packed into this lack of confidence. The one thing I will say that for 24 that there are not many people left who think there is that outright stealing of votes right now and more I think is kind of the emotional disappointment about your guy losing which can easily be overcome by your guy winning next time.
TM: 34:04
Ed: You spent an awful lot of time studying this and I wonder for the legislators listening, do you think there are reforms that are needed in our election administration system or do you think it is basically solid and maybe needs some tweaks?
CS: I start with an answer with a question like answering a question like that by turning around and asking well what are you trying to accomplish. So, for instance, I get a lot of questions about well what can we do to change policy to increase turnout for instance. Are there reforms that we could do and my answer is well yeah. I mean there’s one reform that if you really wanted to goose turnout by 1%, 2%, have Election Day registration, but it’s not going to get you a whole lot of turnout. Anything else that you would propose is probably not going to increase turnout. It might be a good reason the thing to do for other reasons, but it’s not going to increase turnout. You might be interested in making it more likely or less likely that people will have registration problems on Election Day. You know there are best practices there. Something like automatic voter registration, Election Day registration, other policy things could minimize that. And so, there are evidenced basis for different reforms. My lab has been collecting the evidence base of that and we published on the MIT election data and science lab website. A series of white papers that talk about the relationship between research and policy outcomes, best practices where the evidence and the research is the most solid. My favorite good government ones are really kind of boring things like having a really good auditing regime, post-election auditing regime. Risk limiting audits are the gold standard, but even short of risk limiting audits having good audits, good practices that are results are communicated to the public, there is some evidence there that it increases confidence. But you have to communicate the results to the public and most states it looks like do not communicate the results of their audits to the public. So, in that sense, it kind of raises this is answering your question by not answering the question. Sometimes there’s reforms. But sometimes there’s practices that really are as important or even more important so it’s important that in a robust election auditing regime and that can certainly help you spot when there are problems. But if you want to use post-election auditing to increase confidence in the elections then you have to inform the public about the results. And like I said, most states aren’t doing that so do that. That’s not a reform since change in administrative practice. Do that.
Ed: As we wrap up, anything else you would like to share with listeners about the survey or any other thoughts?
CS: We are in the midst. This is you know the end of January as we are recording this, we are in the midst of beginning to write up the final report so for 2024 we will be issuing that I hope in another month or so. Encourage people to come to the website for MEDSEL, my lab and find the report there. We will be publicizing it on the socials and everywhere your reports are sold and actually in general, pay attention to what we are doing over at MEDSEL we are doing a lot of other research to try to highlight as practices and nonpartisan fracture than ideas about improving elections.
Ed: Professor Stewart, thanks so much for taking the time to do this. I found it fascinating and I’m sure our listeners did as well. Take care.
CS: Thanks.
Ed: I’ve been talking with Charles Stewart, III, a professor of political science at MIT and head of the MIT election data and science lab about the 2024 election. Thanks for listening.
You can check out all the podcasts from the National Conference of State Legislatures by searching for NCSL podcasts wherever you get your podcasts. This podcast “Our American States” dives into some of the most challenging public policy issues facing legislators. On “Across the Aisle” host Kelley Griffin tells stories of bipartisanship. Also check out our special series “Building Democracy” on the history of legislatures.
TM: 38:44