What this is
An independent, non-partisan project tracking the partisan composition of election turnout during the 2026 U.S. elections, plus live results for the U.S. House, Senate, and Governor races and past results back to 2018. It runs as a static site that refreshes from official sources every ~10 minutes (every ~2 minutes on Election Day for results).
What “partisan composition of turnout” means
For states with party registration, we show the registered party of voters who have already cast a ballot (by mail, early in person, or on Election Day) — not how they voted. A county shown as “R+10” means registered Republicans make up 10 points more of the ballots cast there than registered Democrats. It says nothing about which candidate those voters chose.
States without party registration (e.g., Georgia, Michigan, Wisconsin, Texas) are shown in turnout-only mode: total ballots cast and turnout %, with no partisan breakdown.
Voting methods
Where the source distinguishes them, turnout is broken out by mail, early in person, Election Day, and provisional ballots, plus mail ballots outstanding (sent but not yet returned) where available.
How it updates & freezes
Each state has its own connector that reads the official feed and writes a standard data file; a scheduler refreshes them. States are added as their 2026 feeds go live through October. When a state’s results are certified, its data is archived (a permanent final snapshot) and updates stop — see the feed-status page for what’s live vs. pending.
Comparisons
Florida includes a 2022 vs 2026 partisan-lean comparison. After the election, a turnout-vs-result comparison overlays each county’s partisan turnout against the actual Governor/Senate result margin, to show which counties over- or under-performed their registration.
Past final turnout. Every state page (and the national page) can compare ballots cast so far with the final turnout in the 2018, 2020, 2022 or 2024 general election. Statewide totals are the official ballots counted compiled by the UF Election Lab (Dr. Michael McDonald; CC BY 4.0); where the Lab gives an estimate for a state that doesn't report total ballots, the votes cast for that year's top race are used and labelled. County comparisons use each county's votes for that year's top race (president, or the biggest statewide race in midterms) from the MIT Election Data + Science Lab's official precinct returns (CC0) — a little below total ballots cast. 2020 county totals are presidential votes from MIT's County Presidential Election Returns 2000–2024. A state-year's county comparison is shown only when its county totals add up to the official statewide count (a few states' precinct files don't).
“If every county voted like…” Under each state’s map is a geographic estimate that works whether or not the state registers voters by party (Virginia uses its localities): pick a past race (president, Senate or governor, 2018–2024), and each county’s 2026 ballots so far, in the selected category, are split the way that county voted in that race. Adding the counties up gives the implied statewide margin, shown next to that race’s actual result; the difference is how much more Democratic or Republican the counties casting ballots so far were in that race. It shows where ballots are coming from, not how anyone voted: within a county, early and mail voters can lean differently from its voters overall. County results come from the results page’s county maps.
Who hasn’t voted yet. In states that register voters by party, a card under the statewide summary takes each party’s registered voters (the same state registration statistics as the Registration vs Turnout page, with their as-of date) and subtracts the ballots cast so far by party, leaving the registered voters who haven’t voted yet and how they lean, next to the lean of registration overall and of ballots cast. It’s statewide only (county registration by party isn’t available for most states) and groups nonpartisan and minor-party voters together, because some states’ ballot counts do. Registration is a periodic snapshot, so people who registered after its date aren’t counted; the card isn’t shown when the snapshot is more than a year old or the party groups don’t line up. It’s the pool of voters left, not a forecast.
Results & balance of power
The results page shows a national map and scorecard for the Senate, Governor, and House, colored by the current margin, with a live balance-of-power bar (seats needed for control, including the seats not up in 2026). Results are wired from official state election-night feeds.
Past results. Use the year menu on the results page to see the final results for President (2020, 2024), U.S. Senate and Governor (2018–2024), plus the last two governor elections in the odd-year states (Kentucky, Louisiana, Mississippi, New Jersey, Virginia). They use the same race ids as 2026 so 2026 results can be compared with them. Sources: the MIT Election Data + Science Lab's U.S. President 1976–2024 and U.S. Senate 1976–2024 statewide returns (CC0), with runoffs (Georgia) shown as the deciding round; governors from MIT's official precinct returns (CC0), used only where the same file reproduces that year's statewide president or Senate result (or, with neither on the ballot, where the leading candidates match the state's official totals), and the 2020 official results in tabular format (CC0). Precinct-summed totals can differ from the certified statewide totals by a few votes. Everything else, including every odd-year race, is entered from each state's official results and links to them (open a race for its source). In the few cases where a state's own results couldn't be reached or are no longer posted (Arkansas 2022, Utah 2020), the certified totals are taken from a republisher and labelled. Where MIT's statewide file disagreed with the official count (Pennsylvania's 2018 Senate race), the official count is used. A race with no verified source would be listed under the map instead of being guessed.
County maps. Click a state (on the map or in the table) for a past race to see its results by county (towns in Connecticut, parishes in Louisiana): each county is colored by its winner and shaded by the margin, and hovering shows the votes. County figures come from MIT's county presidential returns and official precinct returns (CC0), the 2020 official tabular results (CC0), or the state's own county tables (each map names its source). A county map is shown only when its counties add up to the statewide result within 3% (the map says so when they're not exact); otherwise the panel says why there's no map. Alaska reports by state house district rather than borough, so it has none.
Who’s on the 2026 ballot. Before results arrive, each Senate, governor and House race lists its general-election candidates, from the race tables in Wikipedia’s Senate, gubernatorial and House election articles (CC BY-SA 4.0), which cite each state’s certified candidate list. The statewide ballot measures (name, type, subject and a one-line summary, each linked to its page) come from Ballotpedia’s 2026 ballot measures list. Both lists refresh once a day until Election Day, and a refresh that comes back incomplete is discarded so the previous list stays. Official candidate names and vote counts replace them as results are reported.
House district maps. On the House map, click a state to see its congressional districts (zoom with + / − or double-click, drag to pan); hover a district for its candidates and click it for the race. Lines are the ones in use for November 2026, from the U.S. Census Bureau’s TIGERweb 120th Congressional Districts (January 1, 2026 vintage), which already include the 2025–26 redraws in Texas, California, North Carolina, Ohio, Utah, Florida, Tennessee, Louisiana and Alabama. Missouri is the exception: its 2025 map is on hold, and the U.S. Supreme Court’s September 25, 2026 stay keeps the 2022 map for November, so Missouri uses the Census 119th-district (2022) lines. Virginia’s April 2026 redraw was struck down, so its existing lines stand.
Sources
Primary source for each state (full notes on each state’s page):
Loading…
Disclaimers
- Party registration is not a vote choice or a projection; independents/unaffiliated voters are shown separately where reported.
- Figures are unofficial and preliminary until certified; feeds can revise or lag.
- Some states report by municipality or precinct and are aggregated to counties; a few use estimates where noted.
- This is not affiliated with any state, campaign, or party.