Many Americans are now voting before Election Day. Does early voting and voting by mail increase turnout or help one party? Does early voting data allow us to predict election results in advance? Michael McDonald says yes. He is the foremost tracker of early voting and turnout data. We review the results of the 2022 election and early voting in the last three elections and also discuss the new normal of high turnout and the effects of this redistricting cycle.

Guest: Michael McDonald, University of Florida

Study: From Pandemic to Insurrection


Matt Grossmann: How early voting is changing American elections. This week on The Science of Politics. For the Niskanen Center, I’m Matt Grossmann. Many Americans are no longer voting in a booth on election day. How does early voting and voting by mail change American elections? Does it increase turnout? Does early voting data allow us to predict election results in advance? And will election day voting become more anachronistic?

This week, I talked to Michael McDonald at the University of Florida, the foremost tracker of early voting and turnout data. He runs the US Elections Project and is the author of From Pandemic to Insurrection: Voting in the 2020 US Presidential Election. We review the results of the 2022 election and early voting over the last three cycles. We also discuss whether higher turnout is the new normal and whether redistricting turned out fairer this time. I think you’ll enjoy our conversation.

Let’s start with an overview of how people are voting in the US these days. How many of us are still going into the booth on election day and what are the major alternatives and how many people do each?

Michael McDonald: Yeah, so I first started collecting early voting data back in 2008, and I did that because the exit poll organization needed to know the size of the early vote to calibrate some of the exit poll numbers that were going to come out on election night. So they needed to know the size of the early vote in some of the states and they had been running early vote polls. So they do phone surveys in states where there’s high levels of early voting. Some states really have no option because places like Oregon and Washington, they don’t have any polling places to speak of. So you really have to do a phone survey in order to get a sense of what the opinions of people voting are.

So I started in 2008 and then just as a lark I just said, “Oh, I’ll just post this online as a spreadsheet. People might be interested in that.” And you know when you have a million hits to something that you’ve struck on; an idea or something that people really want, information that they want. And that’s what happened. And so I just started keeping track of the early vote over time and started noticing patterns in what was going on because I was collecting and processing the data and being able to see trends in realtime as to what was going on with the early vote.

And so that gets the long way around to answering a question, which is what’s been happening. Well, slowly over time states have been expanding early voting options and so we’ve seen a general increase of the percent of electorate that casts a ballot before election day. And so that trend had been continuing along pretty steady. It goes up a little bit more in presidential election, comes back down. It’s kind of a saw tooth pattern, but generally the trend is an upward trend.

And then suddenly the 2020 election happens and that trend is completely broken because people want to protect themselves in the midst of COVID and they’re really interested and eager to vote too. So I think it’s two things that were going on. And so that trend, we were about 40% of the electorate in 2018 voting prior to election day. Suddenly in 2020 where it’s 70%. And so it was a question. Is this a long term change? Is this just a one off? Are we going to go back to the trend line? And the answer is actually we went back to trend line. Again, we don’t have all the data yet, so it’s going to take a while to actually answer the question for certain, but we’re somewhere around 45% of the votes in this last election were cast prior to election day.

And again, if you were drawing out that trend line, that’s pretty much in line with the trend. And so I think as we look forward, we’re going to continue to see this upward trend, although I think we are reaching near the maximum of where the trends can go. There are only a few states left that are left-leaning states that might adopt early voting options that haven’t done so yet. And so we’re close to the end of it. And so maybe we’ll end up around 50 or so percent of the votes will be cast prior to election day in future elections.

Matt Grossmann: And I know we call it early vote as a whole, but we have some mail voting, have some in person. People are able to drop off their ballot on election day that they got earlier sometimes. What are the major pieces of that early voting pie and how different is it by region or state?

Michael McDonald: Yeah, so mail balloting and in person early voting are the big distinction here and it was primarily western states that were first adopting mail balloting first. Oregon was the first state to adopt at statewide in 2000 through a ballot measure and other states were following on that path. Actually can go way back to the 1910s. And Nevada was the first state to adopt all mail ballot precincts. Because some of those mountainous areas were, had such small population and it was difficult to set up a polling location for those locations that Nevada first innovated this idea a long time ago. Over a century ago. And so it’s really been that has been the driver out in the west, has been people live in remote locations. It’s easier for election officials to run an election by mail rather than set up at a costly polling location that would be difficult for people to get to anyway.

Of course we know Utah has adopted all mail ballot elections, but you also see all mail ballot elections being run in places like Nebraska and North Dakota for the smaller jurisdictions in those states. So this is not exclusively a blue state phenomenon. This is a practicality phenomenon for western states. And really mail balloting is the cheapest way to run elections and it’s the most convenient for voters when they live remotely. Notwithstanding people perceive it as being some sort of democratic plot to change the electorate.

Again, the research on it suggests that yes, mail ballot increases turnout and we can see it again in this election as just we’ve seen in prior elections. The all mail ballot states like Washington are going to have higher turnout than national average, but they didn’t really have lots of compelling statewide offices that might have drawn people to the polls.

So it really does look like it does bring people in, especially in lower turnout elections, like local elections where there may not otherwise be a stimulus to entice people to vote. Here, people get a ballot and we know there’s a lot of mobilization literature out there that says if you get a reminder to vote, you’re more likely to vote. It does seem like having that ballot in hand being delivered to you does stimulate turnout.

But here’s the ironic thing. Again, the academic studies that have been done on this suggest that it actually benefits Republicans. And I know people are not going to believe that. [inaudible 00:07:56] they’ll say, “No, this is a Democratic plot.” No, what that stimulus does, it activates people who are already high propensity voters. And so who are the high propensity voters? Well, by and large they tend to be more Republican than Democratic.

And so if you looked at the early studies and if you look around the debates of the adoption of Oregon of all mail balloting, there were a lot of voting rights groups and Democratic groups that were opposed to mail balloting back in 2000 because they were afraid that it would actually shape the electorate to be more Republican. Now of course, Oregon’s a blue state, so it’s not like Oregon suddenly became a red state. These are just tendencies.

So it’s really interesting that everything that happened with the 2020 election has shaped our belief about what mail balloting does. But the academic studies suggest it does exactly the opposite of what everybody thinks. But again, you can’t change people’s beliefs very easily and so that’s where we’re at in trying to explain this to people.

But for in-person early voting, that was actually first adopted in Texas in the 1990s and it was actually the Democrats who adopted it right before they lost control of the state government. They thought it might help them with turnout and it might help them stave off a couple of elections, that state trending to the Republicans. It turned out it didn’t do that. And there, the literature and the academic studies are more mixed on the effect of in person early voting and many people just find it’s more of a substitution effect that people will choose to vote at a different time. It still helps in some ways to manage the election because you have fewer polling locations. You can have what we call vote centers where anybody in the county can go vote at that location. It helps election officials manage their resources better because they can have their staff at these polling locations and they can help troubleshoot any problems that may occur.

So from an election administration standpoint, in person early voting is still a good way to go to help manage the volume and the workload of election officials. But in terms of its partisan effects or turnout effects, it doesn’t really seem to have a large effect there. It’s really the mail balloting and it’s really the all mail ballot elections that do seem to matter most to turnout.

Matt Grossmann: President Trump doesn’t seem to believe you. He complains incessantly about early voting and that seems to have had an effect on the composition of early voting. How much did it deter Republicans from using that option and is that likely to change?

Michael McDonald: Yeah, I mean part of what people believe is they look at the usage and they can see Democrats are much more frequently using mail ballots. But before 2020, in states where there was multiple methods of voting, mail, in-person, early and election day, mail ballotings tend to be more Republican than the in-person early vote. That all got upended in 2020 because of the rhetoric that was coming from Donald Trump.

And just as a side, that’s fascinating that you don’t often see behavior changes like we saw in reaction to a political figure saying something. And yet here we see large swaths of voters changing the way in which they vote because they are listening to their leader telling them what to do. And then other people also probably voting in reaction to that as well. So I mean just from that standpoint, it’s fascinating to see the direct effects on people’s behavior. We don’t often get to see that in political science, but here we do.

Now in my book, which is From Pandemic to Insurrection: Voting in the US Presidential Election of 2020, what we see there is that, look, Donald Trump had a choice early on in the pandemic and he decided to downplay it. We know that he’s on tape telling Bob Woodward that he didn’t want to cause a panic. And so any reasonable safety measures that might have been implemented like mail balloting to help protect people and election officials, he disparaged that because if he was to embrace it, he would’ve had to say, “I’m embracing the idea that the pandemic is a reality and that people need to protect themselves.”

And so I think what was happening here with meal balloting and Trump’s attacks on it weren’t so much about the fraud or anything like that. It was more about his vision and his rhetoric about the pandemic itself and the intersection of that with the safety measures that states wanted to implement and voters wanted as well. And so it became a partisan issue because of that. And you could see elected officials, politicians at the state level mimicking Donald Trump’s rhetoric around mail ballots in some of these states where they controlled the legislative process and/or governorships that would have to take action in order to implement the safety measures.

And so the fallout from all that is that Republicans are now in places where there’s multiple methods of voting. Again, I have to add that little caveat on there. You see that where in the past Democrats voted in person early, suddenly it’s Republicans voting in person early and where Republicans voted by mail it’s suddenly Democrats voting by mail. So the whole way in which people voted has been upended. And it’s not universal. There’s some variations across states on this, but it’s just fascinating.

One other thing about this though, if you go look out in California, there are some Republican house seats, house wins that were won under mail balloting. And although we’re focused on places like Arizona and Nevada because of their slow counts, there’s still slow counting going on in California and Republicans aren’t complaining about that because they won the elections out there. And remember what I said earlier, in a low turnout election, it may be actually Republicans who were benefited by mail balloting.

So it comes down to the wire in a couple of those races. It’s not inconceivable it was the actual mail balloting that helps the Republicans get control of the House of Representatives. Again, no one’s going to believe that, but I mean that’s what the academic research suggests.

Matt Grossmann: So 2018 and 2020 had historically high levels of turnout and it looks like 2022 is going to come close. What are the major reasons for that and is this a new normal of higher turnout?

Michael McDonald: I think we are in an era of new higher turnout. There are a couple reasons for that, but I think the biggest reason is polarization of our politics. People believe that the choices matter more than they used to and when people believe that choices matter, they’re more likely to take action. If you look back at the big arc of history and the last time that we had high levels of polarization … Now we don’t know if it was in the electorate, but we could see it at the elite level. It was happening in the 1800s and during that period of time, we also saw high levels of turnout as well, at least among the white males who were able to vote.

So we’ve yet again entered another area of polarization. There’s that old curse that many live in interesting times and that’s what’s happened is that we now live in interesting times. People care about politics and they care about the choices and so they want to take action and vote. Presidential turnout rates have been going up for a while and it was curious why the midterm turnout rates hadn’t moved up with the presidential turnout rates. Again, a long history, you look at it. There’s a pretty good correlation between presidential turnout rates and midterm turnout rates.

So something happened where there was a disconnect and suddenly in 2018 I think the disconnect went away and we saw a more inline correlation between the presidential and midterm. So as long as we continue to live in these interesting times, I think we’re going to continue to have high turnout. And it matters because I think if you look at some of the poll misses, and we’re even seeing some Republican pollsters [inaudible 00:17:19] on the election forecasting that they were doing with their polling estimates, they believe that turnout was going to be low.

And I kept saying before the election, you really need to look at those likely voter models. If you see something that gives you an indication that there’s a likely voter model which is forecasting a very low turnout, you should be skeptical of it. And I think that’s part of what happened in the polling misses that some of the Republican pollsters thought we were going to have a pre 2018 midterm turnout and instead we got something that was not quite 2018. We’re down from the 50%, we’re about 46.5, somewhere around there. Still have data in Dallas to count. But we’re somewhere around there and that’s much higher than the 40% that we’ve been in the last 50 years of midterm turnouts.

So I think some people just didn’t believe that we were going to have another high turnout election. They believed the past was predictive and 2018 was a one off and 2020 was another one off. And instead, it looks like we’re just in a pattern of higher turnout for as long as people believe that their choice matters.

Matt Grossmann: And what do we know so far about the 2022 turnout? Did it differ by state and subgroups in the same ways that it has in the past?

Michael McDonald: Yeah, absolutely. I mean, you see these very common patterns in midterm elections where the states that had the competitive elections have higher turnout than the states without competitive statewide elections. So that’s a primary driver. And in states where there was no statewide election, maybe because of timing of the Senate elections or because they hold their accumulatory elections off midterm years, their turnout really was very low.

So it’s a very common pattern though. You need that marquee race to pull people to the polls. Again, it probably mattered because places like New York and California, the voters there were driven more by a more traditional retrospective evaluation of the president. And if you look at the competitive races, it was much more a choice that people were expressing between the visions of the different political parties. If Democrats had somehow managed to nationally convert the election into a choice between competing ideologies, that probably would’ve saved them control of the house.

But it’s hard to break through. The national media and of course the donors pay a lot of attention to those competitive races because they matter to control the Senate. And it’s just hard to break through on individual house races to say, “This house race is really, really important. Look at this crazy character who’s running for house.” It just doesn’t resonate the same way and the media have a harder time following those sorts of stories. There were a few breakout stories like that, but for the most part it’s a different narrative. And unfortunately I think for the Democrats and fortunately for the Republicans is that narrative didn’t break through nationally with the US house.

Matt Grossmann: And did the subgroups look the same? Did the electorate look the same demographically? I know that Democrats, they’re more educated now, but they’re younger now, so maybe that washed out. Anything we can say about that so far?

Michael McDonald: It’s hard to say. I mean we really haven’t gotten all the data yet. Some states have started producing incomplete vote history files, and we can start to look at this, but the reality is that we’re going to need to wait until a few more months before we can really get a good demographic profile of the electorate using something like the Census Bureau’s current population survey voting and registration supplement, which has a long time series on this and is a fairly consistent survey. So we’re going to get pretty good read on it.

Here’s what I expect though. Whenever turnout goes up, it goes up for everybody. There’s usually not one group that’s left behind, but those high propensity voters, yes, they vote at slightly higher rates, but that’s not the driver. The turnout increases are going to come from groups that are moderate to low propensity voters and they’re the people who are activated and participate.

So that’s what I expect. We’re going to see turnout down from 2018. So a lot of people will make comparisons back to 2018 and they’re going to say, “Hey, youth turnout went down,” and everything like that. But again, in the big history, if we look back the last 50 years, I think you’re going to see things like younger people were voting at a higher rate than they have in the last 50 years, excluding 2018. And other groups that had lower turnouts, not the pointy hat academics like ourselves, we voted like 90% PhDs, but everybody else, you’re going to see that their turnout rates were up comparatively compared with the last 50 years.

Matt Grossmann: So as you mentioned, Florida diverged a bit from the national patterns and moved more red than other states. Any particular reasons for that and anything you think is indicative for the future?

Michael McDonald: Well, I think it has a lot to do with DeSantis having a $100 million war chest and Charlie Crist only put up a token campaign against him. Donors were not attracted to his campaign. They knew that DeSantis was likely going to win and Chris was a former Republican. So it’s hard to contrast DeSantis with a Republican-like candidate that the Democrats put up.

So I think that’s really the story for Florida is that Democrats were not enthused to vote in the election and you can see it. I mean, DeSantis got about the same number of votes that he did in 2018. It’s really the difference here is that the Democratic candidates got a lot less. Now, is it possible that there’s some mobilization, some persuasion going on here and that some people are persuaded to vote and crossover? Yeah, absolutely. But I think most likely the primary driver for the diversions from 2018 has more to do with turnout than it does with persuasion or people changing their preferences.

Matt Grossmann: So as you mentioned, we’re still waiting for votes from several states and that’s sometimes blamed on early voting or the early voting ballots, the mail ballots taking longer. What are the big determinants of why some states take so much longer to count the ballot and how much does that matter to people’s perceptions of the election and any way to fix it?

Michael McDonald: Yeah, well there are two things going on here. One is do states allow election officials some time to process ballots that they’re ready to be counted or even pre-counted before election day? And some states don’t do that. And so places like Pennsylvania and Michigan, and although there was a change to law there, but I think only in one county in Michigan said that they were even going to be able to implement the law to allow them a couple more days to process ballots, and Wisconsin. So these three key battleground states, this narrative comes out of with ballot dumps and drops and everything else out of the big cities as they’re processing those ballots late in the evening. That’s one factor that’s at play here.

The other factor at play though is place like Arizona, which is a state that allows voters to drop off their mail ballots at polling locations on election day itself. And California does that as well. You get a lot of ballots coming in on that last day. And so again, it’s just a processing problem. You’ve got to verify that those ballots were legally cast, that they’re not deficient in some way, you’re not going to reject them, and you have to give people a time to cure those rejections. Every state allows that even a couple days after the election. And so it’s just a processing issue for some of those states.

And then the last dynamic that you see is that there are some states like California, Washington, Iowa and some others that allow ballots to be counted if they are postmarked by election day. So they’ll continue to count those even up to two weeks after the election. And so it takes a while to figure out how many ballots were cast in California because they’re going to have a million ballots dropped off on election day itself, and then they’re going to have these ballots coming back in.

I’ve heard Washington election officials joke that their highest turnout day is the day after the election, not election day itself because that’s when they’re getting all these ballots back in the mail. So that’s the other thing that tends to happen. Fortunately for California and Washington, we didn’t have any of those hot statewide races that would’ve been drawing a lot of calls of fraud. But we did have some other places like Nevada, which does allow ballots to be counted if they’re postmarked by election day, if they’re received five days after the election or four days after the election. And so they had a Public Veterans Day came in. So it was actually the Monday rather than the Friday after the election. But so they had to wait.

And so what could we do about this? Well, one thing I think at least to deal with the ballot dump issue that plagues places like Pennsylvania, Michigan and Wisconsin is there’s no urgency for us to know the election results in real time. I think we can wait. I think it’s reasonable for us to wait until, I don’t know, noon the next day. Let’s just give election officials a bit of a chance to count the ballots, make sure that there’s no errors, because that can happen too. That happened in Antrim County with Michigan and it launched several conspiracy theories about Dominion voting systems.

If they had just had a chance to realize, “Oh yeah, we had a programming error on the tabulation and it was because of a local election that had been dropped off, the ballot or the tabulator and we had to reprogram it.” They had the mail ballots. I mean they had the paper ballots so they could go back and verify the votes were counted. So there’s nothing nefarious that happened there. But because there was this one programming error because they, at the last minute, had to change the tabulation and they made that error apply countywide instead of the one precinct where it needed to be applied to, that launches billion dollar lawsuits and it undermines the credibility of our elections.

And again, I just have to keep saying this. It’s just an error. It’s not a conspiracy. People make errors. When 150 million people do something, there’re going to be some errors. And that’s not any indication that there was some widespread conspiracy to throw the election one way or the other. Just the simplest explanation usually is just human error and that’s what happened there. And if we had a little bit extra time to check over the results, look for those sorts of human errors and then release them, there are going to be fewer of them and there aren’t going to be the nefarious ballot drops. It’s just going to be votes being counted.

Again, we’re going to continue to see some votes counted days afterwards so it’s not a panacea for everything, but at least I think it would dispel some or at least debunk … or not debunk, but take the teeth out of some of these conspiracy theories.

Matt Grossmann: So tell us about the US Elections Project and what you have going on and are going to be doing in the future.

Michael McDonald: The US Elections Project was just a way for me to post my updates to my turnout data originally and then I started posting the early voting numbers on there as well and other data resources. And then for the redistricting, we produced a precinct level boundary data with statewide election results data throughout the entire country as well. And so we’re continuing to do those things. It’s hard for people to believe this too. I process all that early vote data myself, so it’s a grueling one month trip through data formats and then some election office changes the format and your code breaks and it is grueling. And getting all those numbers out and the God forbid you make a mistake, then everybody’s like, “Well, what’s the problem with that number?” Again, human error. It’s the most likely explanation. It’s my error.

So that’s something, putting my turnout numbers out there so people can use them. Minnesota actually uses them as their official turnout rate because they’re number one always in turnout so they like the numbers. And other people use them for various purposes. I’m even starting to see people use the early voting numbers for academic work. Please don’t do that. Please don’t use my turnout numbers now. They’re not finalized at all. And even the early vote numbers are not the final numbers. I stopped collecting that data and some states don’t report at all.

In any case, what are we going to do in the future? Well, we did not get funded to produce the 2022 precincts and the boundary data in the statewide results, but we still have a commitment to do that. Over the years I’ve been doing a number of consulting projects through a vehicle that we have at the University of Florida to do different projects for paying clients and so we built up a reserve of money and we’re just going to fund ourselves doing the precinct collection and then we’re going to sell it to clients unfortunately. I mean, there’s sustainability issue here and how do you sustain data collection?

It’s a thorny problem across just about every academic enterprise on data collection. How do you sustain it? So we’re going to move to this subscription model or we’re going to sell the data. I hope that if we can get enough money through that enterprise and other work that we do, that by 2028 we’ll make all that back data available so that it’s available again when we do redistricting. In the 2024 election, there’s going to be even more interest in election analysis and election data analysis. And I think the places like Larry Sabato’s Crystal Ball, which is University of Virginia organization or Cook Political Report or the others, they all failed on the election forecasting because they over-predicted Republicans success.

And myself and others were out there saying, “That’s not what the early voting data is telling us. It’s telling something different.” So I think there really is an appetite for this counter narrative that some of these organizations tend to do, which is they tend to be center right on their forecasts. And so I think that there’s this opening that’s been made for another organization to come in, provide a different data analysis. It’s not going to be looking at polls so much. It’s going to be looking more early voting and doing other sorts of election analyses of voter registration data and those sorts of things and it can provide a different context.

I don’t know what we’ll find in 2024. Look, if you look at my prediction, which I did post on Substat, please do so and you’ll see that I was right. I called every election right where we could use the partisan registration data to forecast the election using the early vote. And that’s been consistent over several years. So yeah, I made a couple errors early on when I was trying to devise the difference and difference approach, but I think by now it works fairly well. And places like Nevada and Arizona, we could clearly see that the Democrats had some strength there and in conversely, North Carolina did not look good for the Democrats. It looked good for the Republicans and certainly Florida. Florida was an easy call for me to say that DeSantis [inaudible 00:34:51] going to win Florida and I was all able to do that looking at a different source of data other than the polls.

Matt Grossmann: So what are the best ways that you can use early voting data before election day to forecast outcomes and what are the potential problems with doing that?

Michael McDonald: Yeah, it’s this difference and difference approach. So you don’t want to look at just the slice of the early vote at one cross section. You need to be able to compare it to a past election and see what’s changing. Now again, there’s lots of caveats, which is that the election has to be run under very similar circumstances as it was in the prior election. And if that’s true, then you can say, well, maybe this change that we’re seeing is perhaps a change of behavior. It’s a change of enthusiasm.

And again, if I can go back to the easiest example, and this would be Florida. Registered Democrats tend to “win” the early vote in Florida, but when they don’t, Republicans win. And so in 2014, I knew that things did not look good for the Democrats because Republicans actually more registered Republicans had voted early than Democrats in 2014. So you knew it because that’s the history. Republicans, if they’re losing, the Democrats are at least okay. But if the Democrats are not winning the early vote, then it’s not going to be good for the Democrats.

And that was even more starkly evident in ’22 in Florida so it was an easy call. I think when all said and done was 300,000 more registered Republicans had voted early in Florida so it wasn’t even close. So that’s the sort of thing that you do. And guess what? The polling averagers do the same thing but in a different way. They do difference and difference too. They look at the polling misses that the organizations had in prior elections and they adjust the current estimates that they had from pollsters based on the performance that they did in the past. That’s exactly what I’m doing with early vote. It’s just a different data source is all that’s being plugged into the model. And so I really don’t understand people who say that early voting data information is worthless, don’t look at it because there is information there. And quite frankly I think you’re a fool if you don’t look at it because you’re just disregarding another source of information that might help you reach a better conclusion about what’s happening with the election.

And I look at polls too. So it’s not like I’m exclusively looking at early voting. You want to think about the whole gamut of everything that’s going on to try to understand what’s happening in an election. Just take on all the data, take on all the information. I think that’s your best way to move forward when you’re doing any sort of analysis.

Matt Grossmann: So I think the hesitancy was that we only have a couple of elections in some of these states with large early voting numbers, or at least the last one that we had was in pandemic conditions. So I guess do we now have enough data to make these comparisons in most places where we’re going to see substantial early voting?

Michael McDonald: Well, I did not make a forecast out of Pennsylvania because the law had changed. That’s part of the difference and difference. You have to assume that there’s been no change to the conditions under which the poll was run or the election was run. It’s same thing with the pollsters. They assumed that all the major polling organizations didn’t tweak their polling methods because of the misses that they had in 2020 or in prior elections, which is just ludicrous because the pollsters, their whole reason for being is to make accurate predictions. So if they flubbed a prior election, they’re going to look at themselves and they’re going to say, “Hey, what did we do wrong? We need to improve our polling methods.” And so I think that’s where people made some mistakes. They just assumed that there was this Democratic bias in some of these major battleground states when no, there wasn’t really a bias there. The pollsters had changed themselves.

So to go back to it, I mean in a state like Pennsylvania, major change to law, and they used to have excuse required absentee voting in 2018. Now they have no excuse. You couldn’t really make a comparison of 2018 to 2022 to know what that information was telling you about the behavior changes because it was driven by a change to the law, not just a change in behavior. But going forward to 2024, it’s going to be really hard to look back at 2020 and make comparisons of 2024 to 2020. So my sense is going to be that our best bet … Remember, it looked like we reverted back to the trend line that we saw in the early vote so I think the best comparison’s going to be for 2024 to go back to 2016 and make that comparison where we can make that comparison where the election hasn’t changed in some fundamental way too.

Matt Grossmann: So this was also the first cycle after redistricting, and you’ve done a lot of research in that area as well. It looks like nationally that it’s going to be pretty proportional between votes and seats in Congress, but obviously with a lot of variation across states. So what do we know so far about the effects of the latest redistricting cycle?

Michael McDonald: And just to underscore it. Virtually everything that we know about partisan composition of the districts across the country and the partisan effects of redistricting are due to the effort that we did at the University of Florida. Everyone was using the voting and election science team data that we created and with just a few exceptions. States might have created their own databases. So what does it tell us about things that said … Look, reformers over the last decade; and it’s been really been going on for a few decades now; have really pressed for redistricting form across the country and they’ve been largely successful. Not a 100% successful. There were some places where there were Republican gerrymanders and there were some places where there were Democratic gerrymanders, and some of those are very consequential. Maybe not for the overall partisan composition of Congress, but certainly at the state legislative level it’s hard to look at Wisconsin as being a democracy anymore when it comes to its state legislature. The Democrats simply cannot win control of that chamber or legislature because of the way in which it’s been gerrymandered.

I know the US Supreme Court looked the other way on this, but it’s hard to think that they really have a representative government in the state legislature of Wisconsin anymore. So there are some very consequential consequences for the state governments but as a whole, really we’re looking at a successful reform effort. And are there some problems? Yes, we had three voting rights violations in Alabama, Georgia, and Louisiana. Unanimous decisions that even Trump appointed judges said that these were shut and dried cases of Section 2 Voting Rights violations. And the Supreme Court basically stayed to the decision at Alabama and the other courts just followed suit at the lower level and said, “Well, we’ll wait for the Supreme Court to make a judgment on this.”

So it could be that we’ll see the Supreme Court neuter the Section 2 as they’ve already done to section 5, the Voting Rights Act. That is a possibility and I think a very real one where there other places like here in Florida where I think there’s a violation of Florida’s constitution and it’ll be up to the Florida Supreme Court to decide if there was a violation of Florida’s constitution. So unfortunately there were some places like that.

On the flip side, we had places like Illinois gerrymandering as well for the Democrats, but other places where there were Democratic gerrymanders, places like New York and Maryland. The courts took the lead and overturned those maps. And that was also going on with states like Pennsylvania and North Carolina. And so the courts at the state level have been good about intervening. Could we get courts everywhere to do it? Because the Supreme Court basically has kicked at least partisan gerrymandering claims back to the state courts. And hopefully I still hope that people will do the right thing and they can see that there’s clear intent to dilute the vote of African Americans in Alabama, Georgia, and Louisiana, and in violation of not only the Voting Rights Act, but the 15th Amendment. But I’m not going to hold my breath for the US Supreme Court to rule otherwise.

Matt Grossmann: So what are the first things you’re going to be doing with the 2022 data and what should we look for going ahead to 2024?

Michael McDonald: The 2022 day I’ll continue to process it. Again, it’s a grueling thing. I can’t constantly scan. I’m only one person so I’m not sitting here every evening looking through all the state websites to see if we’ve gotten the certified results yet. So I generally take my time. Some states we don’t even have the final numbers. I’m just guessing based on the vote that we do have and what I think might be outstanding within those states. So we’ve got to get the certified results. Once we get the certified results, again, one of those consulting projects that we’re doing is we’re providing maps and data to the Almanac of American politics. So we’ll be providing all of the election results to them. And then we’ve got a couple of paying clients for the precinct data at the moment. And so again, I’m hopeful that we’re going to hit upon a sustainability model with this so that we can make it available. And then again, in the future, we’ll embargo it for now, but then we’ll make it available down the road.

I’m interested in some questions. One of the questions that Dan Smith and I here at the University of Florida have been working with Matt Barreto at UCLA on is Hispanic cohesion in South Florida. The Cuban American population looks much differently than the non Cuban American population there. They’re all Hispanics. I mean, Miami data is 75% Hispanic. It’s an incredibly diverse Latino community. And so looking at that cohesion and what does that mean for voting rights issues. That’s an interesting question that people haven’t addressed yet. So we’re looking at that. We have paper we already presented on that, and we’re going to crunch more data once we get the 2022 data through.

And that’ll be able to answer some of the interesting questions like what really went on with Hispanic voting patterns in Florida and Hispanic voting patterns elsewhere? I know people have already started to look at this looking at county level data and they’re seeing a reversion back to before the 2020 election in Hispanic voting patterns in places like Nevada, Arizona, this Texas border. And I thought that was what was going to happen. I thought 2020 was a one off because when was the last time we saw a big move towards the Democrats among communities of color? That was 2004 when Bush was an incumbent president.

And so I think there’s a matter of incumbency that plays into some of these voting patterns among communities of color. And remember back to 2024, Karl Rove was heralded as a genius because he figured out how to capture Latino and African Americans, and this was going to completely transform the Republican party. That didn’t play out when Barack Obama came along. But that’s not, again, a surprise because again, presidency has a lot of powers and perks to it.

So I’m not surprised by this reversion. I don’t think it necessarily played out in Florida in the same way that it played outs in some of the other states. But that’s one of the things that we really want to do, is collect all that data up and be able to say something meaningful about these patterns of voting that we’ve seen using the precinct data rather than polling data to come at measurement in a different way.

Lastly, we’ve also collected a whole bunch of data on redistricting and we’ve collected every redistricting plan that was produced, at least publicly, where we can get the electronic version of them, I should also say. Because some states didn’t really make them available very easily. So we’re going to analyze those and we’re going to talk about the transparency, the redistricting process. I’ve supportive of these reform efforts over the years, especially in the terms of their transparency and accountability issues that plague redistricting when it’s done behind closed doors.

Matt Grossmann: What can we learn from the new book?

Michael McDonald: Well, a lot of what’s going on now in terms of the voting patterns are just a continuation of what started in 2020. And when I first wrote that book, it was really in the midst of early on in the pandemic, and I was getting all these reporter calls asking me, “How does mail balloting work?” And so I’d spend two hours on the phone with a reporter explaining to him how mail balloting worked. And then I talked to Rick Hasen about this, and Rick said, “Well, you should write up a little explainer so that you don’t have to repeat yourself every time you talk to a reporter.” I said, “Hey, yeah, that’s a good idea.”

So I started writing up a little explainer and that’s one page, two pages, three pages. It’s like a whole book chapter. It’s two book chapters. It actually ended up being two book chapters. And so I was just like, “Yeah, this is just a book. This isn’t a little explainer.”

And so then as the election progressed, I just continued to archive news stories as they were happening in real time and doing data analysis in real time that I could see had something interesting to say about the death of a poll worker or an election worker, I should say, in Fulton County, how did that affect the ability for the election office to conduct the election? And sadly, you could actually see operations affected in Fulton County.

And so to go back to it though, the book, we have to give a lot of credit to the election officials who managed to run the highest turnout rate election that we’ve seen since 1900. That’s really remarkable and they did it in the midst of the pandemic. Some people really died. I mean, there’s no other way to say it. I mean, they died from the pandemic and the conduct of their job. Other people got sick. We don’t know how many other people were sick, their loved ones or other people that they knew.

So in some ways it’s a triumph, but it’s also a tragedy because we didn’t have to expose people. We could have run all mail ballot elections nationally and tried to reduce exposure, but the legal battles that went back and forth and with Donald Trump disparaging mail balloting, believing that the Democrats were somehow going to benefit from it, that unfortunately took its toll. And the other thing, and we’re still seeing this to today, the conspiracy theories that say that somehow elections are fraudulent. Some people decided to take action on that and certainly the insurrection was part of that, but also we see threats against election officials and they’re continuing, even in this election. People watching polls and intimidating people. That happened.

I know there were a lot of high profile stories in Arizona about that, but it was also going on in North Carolina. And there were two poll workers in Georgia that had to be removed from a polling location because of the conspiracy theories. And now we have the Maricopa County elections official who had to go into hiding because of death threats against him. I mean, this is insane. And there are good people being run out of their positions and it’s very troubling. I know the election deniers lost by and large at the state level, but they’re still active at the local level. And I think as we’re talking about this on November 21st, here’s a prediction. I think we’re still going to see problems with certification in Arizona and it could be some other places in Nevada too that we might see some certification issues and could be other places like Texas had some problems with it in prior elections, and New Mexico.

So I don’t think we’re entirely out of the woods yet and there’s still mischief that can be played at the local level, even if we withstood some of these attacks on our democracy that were happening at the state level. And so the story, the book really isn’t done yet. I had to finish it at some point, get it off to a publisher. But unfortunately, I think we’re going to continue to see some of the themes of the book play out over the 2024 election. In some ways, again, some of these people are just saying the 2022 election was a dress rehearsal. It wasn’t the main event. They believe that 2024 is going to be the main event. So yes, it’s good that we managed to get through without a major incident in ’22. That’s great but I really don’t think that we’re out the woods yet for 2024.

Matt Grossmann: There’s a lot more to learn. The Science of Politics is available biweekly from the Niskanen Center. I’m your host, Matt Grossmann. If you liked this discussion, I recommend checking out these episodes highlighted on our site: will Trump Anger Motivate Black Turnout? How Voters Judge Congress, Did the 2022 Election Show Us How Democratic Campaigns Win? How Information About Politicians Persuades voters, and How Much Are Polls Misrepresenting Americans?

Thanks to Michael McDonald for joining me. Please check out From Pandemic to Insurrection and then listen in next time.

Photo Credit: iStock