This page is a record. It gives the sources and methods of the 2026 forecast and of the pages of the 2026 election. The sources and methods of the whole site are on About the data.
Reading the data file…
·
Some counties count mail and early ballots for the whole county, not by precinct, and the State Election Board reports those ballots on a line of their own for the county. For a statewide race, those ballots stay with their county: the county totals on this site include them, so the counties add up to the certified statewide total. For a legislative district, we split each candidate's county-wide ballots among the district's precincts.
Every contest on the map is drawn on the 1,984 precincts of 2026. Precinct lines changed in 2023. For older elections we estimate the vote of today's precincts from where people live. Each 2020 Census block (about one city block) belongs to one precinct in the old map and one in the new map. The votes move with the people who live there. County totals are exact. A precinct that was split is an estimate. Some counties count mail and early ballots for the whole county, not by precinct. We split each candidate's county-wide ballots among the county's precincts, and we did the same for the runoff of 25 August. In the Republican-only and Democratic-only contests, gold marks the candidate listed first and teal the second. Those two colors carry no party meaning. The general elections are colored by party.
Every share on this site is a candidate's share of all votes cast in that race, other candidates included. A lead is how far the winner's share is above the runner-up's, in percentage points of all votes: 55% to 45% is a 10-point lead. We write it with the leading party first. R+3.2 means the Republican won by 3.2 points, and D+3.2 means the Democrat won by 3.2 points. The layers that compare an office with the governor race take one office's lead and subtract the governor race's lead on the same ballots. D+ means that office's Democrat did better than the Democrat for governor. R+ means its Republican did better than the Republican for governor.
Registration and the number of voters the forecast expected come from the State Election Board's list of registered voters. Our model gave each registered voter a chance of voting on 3 November, and we checked it against who voted in 2024. Added up by precinct, those chances give the expected ballots and their mix of parties. Turnout is measured against all registered voters, including those the Election Board lists as inactive (they can still vote), which is how the State Election Board counts it. The model ranks voters well, but its statewide total is only one estimate. The three midterms on record put it somewhere from 0.99 million to 1.36 million voters. Shares of all votes change color every 10 percentage points. Leads change color at 5, 10, 20, 30, 40 and 50 points.
The What-if tool works out results from settings you choose. It makes no forecast of its own. It starts from the Forecast page's numbers in every precinct and lets you change the voters. In "Set the voters by party" you set, for Democratic, Republican and independent registered voters, the share who vote and the share of their votes that go to the Democrat. The page then works out every precinct again: ballots are registered voters times the turnout rate, and the Democrat's vote is the ballots times her share, both from the forecast's own precinct figures. Each precinct moves from its published forecast by the change the sliders make, so at the starting values every county, district and TV market matches the Forecast page. The starting values are the forecast's own rates and shares. The 2022 and 2018 rates in the lists are voters divided by the people registered today. People who have left the list since are missing from both counts, so those rates run high, most of all for 2018. They show the gap between the parties more than the true turnout of that year.
In "Set the result" you set the statewide lead in the governor race. Each precinct then moves as far as it has moved, in the past, for each point the whole state moved. That figure comes from twelve statewide races between a Democrat and a Republican from 2020 to 2024, with recent years counted more (2024 weighted 0.35, 2022 0.30, 2020 0.15). Where a precinct's own history says little, its figure is pulled toward 1, a precinct that moves the same as the state. Then all precincts shift together so the state matches the result you set. Little Dixie and the southwest have figures near 1.5, and the cities near 0.9, so a statewide change moves the rural southeast about half again as far as Oklahoma City. The starting sets are named, and each one states its source. The chart "What decides it" moves one slider across its full range with the others held, and shows the change in the governor race's lead. Each other office then follows the governor result, with its own comparison with the governor race and its own share of skipped ballots.
The Early vote page counts the ballots cast before election day, by the party each voter is registered in. Early in-person votes come from the State Election Board's reports on in-person early voting, read at each report the Board published. Mail ballots come from the Board's absentee ballot totals, read each weekday: the ballots requested, and the ballots received back by a county election board. Every table lists all four parties and then the Democrats' share of the Democratic and Republican votes. "Same point" means the same number of finished early-voting days. Early voting ran Wednesday to Saturday before the election. So day one of 2026 (28 October) sits beside day one of 2022 (2 November) and of 2024 (30 October). A ballot cast by a party's voter is not a vote for a candidate, and the page never calls it one. The forecast did not use this page's counts.
The Forecast page gave one estimate for governor statewide and one inside the 5th Congressional District, each with a range. The statewide figure started from past governor results. We then adjusted it for the mail ballots returned by 22 October, and by our own judgment. The range came from how much past governor races varied. Every candidate is counted. The forecast gave the three independent candidates 2% together, less than minor candidates averaged from 2014 to 2022, because no Libertarian ran this year. The lead is the Republican's share minus the other main candidate's. Figures on the Forecast page are rounded to the nearest percent. The final figures were fixed on 2 November 2026.
The Forecast page shows the governor race.
The precinct map and the tables show each candidate's share of all votes cast. In every precinct the forecast gave the three independent candidates together their statewide share (2%) and split the rest between the two main candidates by the precinct estimate. It does not try to show where the independents run stronger. On the change view, we compare each office with the last general election for that office with the same party in the race. That is 2022 for most offices, and 2018 for Attorney General, Insurance Commissioner and Treasurer. For the 5th District House race, we compare with the governor vote inside the district. The figure we compare is that candidate's share of all votes, others included. A precinct with fewer than 50 votes in that year is left gray on the change view.
Half of the precinct map came from a model of single voters and half from precinct history. The model is built on the list of registered voters. We tuned it to each party and region and to the 2022 results by region, and adjusted it for public signs of support where we knew of them. The history is 12 statewide races from 2020 to 2024. The precincts add up to the statewide figure, so the map spreads that single estimate across the precincts. People who chose to answer the survey or send a yard-sign photo are not a random sample. The statewide figure does not come from them.
Each page, contest and map layer says why it is there. The list below collects those notes.
The Political ads page combines eight public records, and each one is measured from its own documents. Ads bought (station orders filed with the Federal Communications Commission, and the Google and Meta archives) can be added across channels, because no two of them describe the same dollar. Payments in campaign finance reports are shown apart. They pay for those same ads, weeks later. The details below come from the page's data file.
The party of the member who held each seat comes from the election that last decided the seat. That is usually its last general election. When no general election followed, it is the primary or special election that settled it. A seat with no deciding election on record shows a dot in that column. Senate Districts 23 and 41 were not on the 2026 ballot. State Senate terms are four years and staggered: the 24 even-numbered districts were on the ballot in 2026, plus Senate District 17 for the rest of a term. House District 92 also had an election for the rest of a term. So it appears twice on the Legislature page: once for the rest of the current term and once for the next full term. When candidates of only one party filed for a seat, that party's primary (or the filing period, if only one person filed) decided it, and there was no November vote. The State Election Board calls the winner "deemed elected". Nominees, filings, withdrawals and the results of challenges to candidates come from the Oklahoma State Election Board. The past results of each district are certified precinct results, added up inside the 2026 district lines.
On election night the page read the State Election Board's results as they came in. It checked a set of questions written before the count began. As the count came in, it also projected the final lead in the governor race from the votes counted so far. Every figure on it was unofficial until the State Election Board certified the count.
To report an error, write to the address on danthedemocrat.com.