On Tuesday, early voting continued across the country, including in key battleground states like Georgia, where more than 300,000 people cast ballots on the first day of early voting.
While Virginia isn’t a major battleground state this year, there are already more than two weeks of mail-in and advance in-person voting in the books, including hundreds of thousands of ballots that have already been cast. We closely monitor early voting trends to determine expectations for election night results, and a clear pattern has emerged in Virginia that may be part of a national trend.
Regular voters — those who regularly show up to vote in most elections — made significant use of early in-person voting in Virginia, particularly in areas that tend to support Republican candidates. As a number of other states begin early voting, we will be watching closely to see if similar patterns occur elsewhere, as early voting has important implications for how election results can play out on election night.
Follow live updates on 2024 elections
Virginia’s first two weeks of mail-in voting and early in-person voting suggest the gap between early voting in Republican and Democratic states may be smaller this year than it was in 2020.
The table above shows how many votes and absentee ballots were cast in 2020, based on the jurisdiction’s overall tilt toward Democratic or Republican candidates. In 2020, about 2.8 million, or 63 percent, of Virginia’s approximately 4.5 million votes were cast before Election Day. However, there were significant differences in early voter turnout based on the partisan leanings of a city or county. In cities and states with strong support for Democratic candidates, 70% of voters cast absentee ballots. By comparison, only about 50% of voters in Republican states and cities cast absentee ballots.
The table below compares the 2,024 ballots returned by October 11 in the same precincts. TargetSmartselective data vendor. So far, voting has been more likely in places that lean Republican than in places that lean Democratic.
One way to capture the trend is to calculate the percentage of the 2020 vote in a precinct that has already been cast in 2024 through ballots cast in 2024. It ranges from about 12% in areas that lean heavily Democratic to about 22% in areas that lean Republican.
TargetSmart also reports how often such voters turned out in the last three state general elections (the 2020 presidential election, the 2021 gubernatorial election, and the 2022 midterm elections).
Voters who participated in most or all of these elections are almost certain to vote in the upcoming presidential election—which means that observing these voters vote after, rather than before, polling day doesn’t tell us much about what turnout will look like. as However, if early voters are less consistent or new voters, this could indicate a change in election trends before November.
So far, it appears that about 90% of Virginia voters are repeat voters who have participated in at least two or three of these elections. There is no meaningful difference in the shares of consecutive voters based on the partisan bias of the district.
While it is tempting to interpret the patterns as evidence that consistent Democratic voters will vote at lower rates than Republicans in Virginia, we believe such a conclusion is premature. To see why, the chart below uses the Virginia Public Access Project to show how voters cast absentee ballots in the state’s 2022 election.
Specifically, the table shows which part of the general ballots issued 25 days before the election is registered. Data shows that at this point in 2022, only about 21% of early in-person ballots and 39% of mail-in ballots have been cast. Note that a higher percentage of absentee ballots were cast in areas that support Republicans more. in areas that support more Democrats.
Why might voters in more Democratic precincts vote later than in more Republican precincts? One possible explanation is availability “Satellite Polling Places” in Some Virginia Jurisdictions. These are additional locations outside of a county or city’s registrar’s office where voters can vote early in person starting in late October. Such satellite polling locations can make early in-person voting available in areas with larger populations that tend to lean more heavily Democratic. We’ll be watching closely over the next few weeks to see if these differences by area even out or reverse as satellite voting begins.
If we continue to see a lag in voter turnout in Democratic districts compared to Republicans after satellite voting begins, it would prompt further investigation into whether consistent Democratic voters are less mobilized than they were this election cycle.
Why are these samples important to the NBC News Decision Desk?
Part of the reason the NBC News Decision Desk follows outside voting so closely is its impact on the consistency of how votes are reported on election night. Democrats are more likely than Republicans to vote by mail and early in person, making our work on the Decision Desk more difficult than usual in 2020.
We are used to the fact that polling stations report election results at different speeds on election night. In some states, smaller, mostly rural counties report their votes first, while in others, larger, mostly urban counties report first. We have long used models that attempt to explain such disparities by analyzing whether states that vote more heavily tend to lean more Democratic or Republican in past elections than states that vote less. Models help assess what current reporting results suggest about likely final outcomes.
Before 2020, such models did not adequately account for the fact that states did not report different results with the same speed voting regime. In some states, counties first report mail-in and early in-person voting and then Election Day votes. In others, the order is reversed.
As it became clear in 2020 that Democrats would disproportionately vote by mail or in person early, and Republicans would vote disproportionately on Election Day, we had to quickly refine our models to account for differences only in reporting rates. counties, but also for differences in which voting modes are counted. While we are in a better place to address this issue now, it will take us longer to project races in states where there are differences in voter use by mail, early in-person voting, and speed. report with voting mode.
However, seeing the data from Virginia makes us cautiously optimistic that we may see less partisan differences by voting mode in 2024 than in 2020, at least in states like Virginia with extensive early in-person voting.