Explainer · Turnout
What Early-Vote Data Can and Can't Tell You
Every election brings a flood of early-vote charts sorted by party, age and county. Some of it is genuinely useful. A lot of it gets read as a scoreboard, which it isn't.
Early-vote data tells you who has voted and how, and in some states their party registration, age or gender. It does not tell you whom they voted for. It's most useful for turnout comparisons with a similar past election, and least useful as a forecast. In the 2022 midterm, about 54% of voters cast ballots before Election Day, by mail or in person, so early data covers a large share of the electorate but not all of it.
Early voting has changed how elections unfold. Instead of one day of voting, there are weeks of ballots arriving, and election offices publish running counts of who has cast them. Analysts, campaigns and reporters slice those numbers every day.
Because the data comes from election offices, it's real. The trouble is in the interpretation. This guide covers what early-vote data measures, what it can't, and how to read it in the days before and after an election.
01What's in the data
States and counties publish counts of mail ballots requested, returned and accepted, and of voters who cast ballots in person before Election Day. The details vary by state. Depending on what a state collects, the files may include:
- Voting method: mail or early in person.
- Geography: county, and sometimes smaller units.
- Party registration: only where voters register by party. Ballotpedia counted 31 states, plus the District of Columbia and the U.S. Virgin Islands, that let voters list a party and publicly report registration totals, as of August 2026. In the rest, there is no party breakdown at all.
- Demographics: age and gender in many states and, in some, race.
The best-known national compilation is the University of Florida Election Lab's early-vote tracker, run by Michael McDonald, who has published early-voting statistics since 2008. The lab describes the numbers as clues to expected turnout and hints of relative partisan engagement when compared with a past comparable election. That's the right frame.
02How big the early vote is in a midterm
The U.S. Election Assistance Commission's survey of the 2022 general election, the most recent midterm, gives the baseline:
EAC 2022 Election Administration and Voting Survey: 112,054,124 ballots counted; 35,316,617 by mail; 21,322,200 early in person; 50,860,863 in person on Election Day. Turnout was 46.8% of the citizen voting-age population.
So in a typical recent midterm, roughly half the electorate has already voted when Election Day begins, and the other half hasn't. Any read of early data has to account for the voters who aren't in it yet.
03What early-vote data can tell you
- Turnout pace. Compared with the same point in a similar past election, it shows whether participation is running ahead or behind, statewide and by county.
- Who is engaged. Shifts in the age mix, or in the share of registered Democrats, Republicans and unaffiliated voters where that's recorded, show which groups are turning out early compared with last time.
- Where ballots are still out. Requested-but-unreturned mail ballots show how many more could arrive, which matters for how long the count will take.
- How the count may unfold. Knowing how much of a state voted early, and when the state counts those ballots, helps you anticipate whether early returns will lean one way. See red mirage, blue shift.
- Problems to fix. Counts of rejected or flagged mail ballots tell voters and campaigns how many ballots need a signature cure. See cure deadlines.
04What it can't tell you
- Vote choice. Ballots are secret. Election offices record that someone voted, not whom they voted for.
- How party registration translates into votes. Registered Republicans can vote for Democrats and vice versa, and unaffiliated voters can go either way.
- Whether turnout is new or shifted. Early votes are often cast by people who would have voted anyway, just on a different day. A party that moves its reliable voters into early voting can post big early numbers without adding a single vote overall.
- Final turnout. Election Day voters, late mail ballots and provisional ballots are still to come.
- Apples-to-apples comparisons across states. Mail voting rules, early-voting periods and reporting differ by state, so a big early share in one state and a small one in another may mostly reflect the rules.
Voting habits by party change from cycle to cycle as parties change their messages about early and mail voting. A gap that looked meaningful in 2020 or 2022 can mean something different in 2026.
05Common misreads
| The claim | What the data supports |
|---|---|
| "Party X leads the early vote" | More registered members of one party have voted so far. That's engagement, not a vote margin. |
| "Record early vote means record turnout" | Possibly, or voters may simply be voting earlier. Compare total turnout once Election Day votes are counted. |
| "Young voters are surging" | Check the share, not just the raw number. If everyone is voting more, a group's share may not change. |
| "This county's early vote is down, so the party is in trouble" | Early voting periods, drop box access and mail rules may have changed. Compare like with like. |
| "Polls of early voters show who's winning" | Surveys of people who say they've voted are polls, with polling error, not counts. |
06A better way to use it
- Pick the right baseline: the same state in 2022 is usually a better comparison for a midterm than 2024.
- Compare shares and rates, not just totals.
- Treat party registration as a clue about engagement, not a vote count.
- Note unreturned mail ballots, and check the state's receipt deadline.
- After the election, compare early-vote patterns with the actual results county by county. That shows how much the early data was really telling you.
- For who voted and why, wait for surveys built for that purpose. See exit polls, VoteCast and the Voter Poll.
Early-vote data is a turnout measure. Used against a sensible baseline, it tells you who's engaged and how the count may unfold. Used as a scoreboard, it tells you much less than it seems to.
Background on how early and mail voting grew: Early voting, mail voting, and why Election Day is now election month and who votes in midterms.
FAQFrequently asked questions
Does early-vote data show who is winning?
No. Election offices record that a voter cast a ballot, not whom they voted for. Party registration, where it exists, shows engagement, not vote choice.
How much of the midterm vote is cast early?
In 2022, the Election Assistance Commission found 31.9% of voters voted by mail and 22.2% voted early in person, so roughly half voted before Election Day.
Why doesn't every state report early votes by party?
Not every state registers voters by party. Ballotpedia counted 31 states, plus D.C. and the U.S. Virgin Islands, that collect and publicly report party registration totals as of August 2026.
Can a big early vote lead disappear?
Yes. If one party's voters shift from Election Day to early voting, its early numbers rise without adding new votes, and Election Day ballots can change the picture.
Where can I find early-vote numbers?
State and county election offices publish them, and the University of Florida Election Lab compiles a national early-vote tracker.
Sources
Sources accessed October 31, 2026. Rules and dates can change; confirm with your state or local election office.