Reported results
Actual reported votes. Forecast estimates appear separately.
Waiting on the first candidate list from the feed. Oklahoma votes on Central time, so the state closes at 8:00 PM ET, in CLOSED.
No votes counted0% est. reporting
Oklahoma runoff · August 25•Statewide forecast · county projection
Oklahoma Governor Republican Runoff
Gentner Drummond and Mike Mazzei finish a runoff neither led outright in June. The TPSI model separates them by less than a point and a half — inside its own margin of error — with Drummond ahead on the strength of rural and western Oklahoma and Mazzei holding the two metros. Whoever finishes first takes the nomination; there is nothing after this.
Forecast snapshotModel metrics update as new results arrive
PLURALITYWin probability
57.1%Drummond
Top outcomes
Drummond wins57.1%
Mazzei wins42.9%
No third option—
Model-estimatedPercent reporting
0.0%reporting
Model projection · not actual results
Drummond projected ahead
The model has Drummond ahead by 1.4 points, well inside its margin of error. Oklahoma County and Tulsa County together cast under a third of the expected vote and the model has Mazzei narrowly carrying both, which means this is decided in the seventy-five counties outside them.
Projected margin+1.4Drummond over Mazzei
Projected final vote shareMuted bars indicate a forecast, not certified results. They use the same 0 to 100% scale as the reported-results bars, and are the sum of all 77 county projections.
Forecast onlyProjected turnout~375,000
Votes remaining375,000
Margin SD±7.9 pts
Margin range (95%)−14.4 to +17.2
No history snapshots · live data only
DrummondMazzeidepth = margin · hatched = too close to callscroll or pinch to zoom · drag to pan
County projections are demographic estimates, not local polling: 460 interviews across 77 counties, a median of three per county, and 74 of 77 with a 90% interval that crosses zero. We do not call counties. They are here to show where the outstanding vote sits — projected turnout is drawn from registration and past runoffs, and is far more reliable than any county share.
| county | region | reporting | turnout | drummond | mazzei | margin |
|---|---|---|---|---|---|---|
| Oklahoma | Oklahoma City Metro | 0% | 55,733 | 48.3%26,936 | 51.7%28,797 | +3.3 |
| Tulsa | Tulsa Metro | 0% | 54,828 | 49.6%27,189 | 50.4%27,639 | +0.8 |
| Cleveland | Oklahoma City Metro | 0% | 26,121 | 46.3%12,091 | 53.7%14,029 | +7.4 |
| Canadian | Oklahoma City Metro | 0% | 16,164 | 48.5%7,838 | 51.5%8,326 | +3.0 |
| Rogers | Tulsa Metro | 0% | 12,672 | 53.5%6,781 | 46.5%5,891 | +7.0 |
| Wagoner | Tulsa Metro | 0% | 9,698 | 53.6%5,203 | 46.4%4,495 | +7.3 |
| Payne | North & Northeast Green Country | 0% | 8,276 | 51.8%4,284 | 48.2%3,991 | +3.5 |
| Comanche | Western I-40 & Southwest | 0% | 8,017 | 57.9%4,644 | 42.1%3,374 | +15.8 |
| Pottawatomie | South Central & Southeast | 0% | 7,759 | 52.2%4,053 | 47.8%3,706 | +4.5 |
| Creek | Tulsa Metro | 0% | 7,629 | 50.5%3,849 | 49.5%3,780 | +0.9 |
| Grady | South Central & Southeast | 0% | 7,629 | 48.5%3,699 | 51.5%3,931 | +3.0 |
| Garfield | Northwest Plains | 0% | 7,500 | 51.9%3,890 | 48.1%3,610 | +3.7 |
| Logan | Oklahoma City Metro | 0% | 6,853 | 48.6%3,329 | 51.4%3,524 | +2.8 |
| Washington | North & Northeast Green Country | 0% | 6,078 | 49.6%3,016 | 50.4%3,061 | +0.7 |
| Kay | North & Northeast Green Country | 0% | 5,560 | 52.0%2,891 | 48.0%2,670 | +4.0 |
| Mcclain | Oklahoma City Metro | 0% | 5,560 | 46.6%2,589 | 53.4%2,971 | +6.9 |
| Pittsburg | East Central & Kiamichi | 0% | 5,560 | 53.0%2,949 | 47.0%2,611 | +6.1 |
| Lincoln | East Central & Kiamichi | 0% | 5,431 | 54.7%2,973 | 45.3%2,458 | +9.5 |
| Stephens | South Central & Southeast | 0% | 5,302 | 49.6%2,631 | 50.4%2,671 | +0.8 |
| Carter | South Central & Southeast | 0% | 4,914 | 51.2%2,516 | 48.8%2,398 | +2.4 |
| Le Flore | East Central & Kiamichi | 0% | 4,914 | 51.3%2,521 | 48.7%2,393 | +2.6 |
| Mayes | North & Northeast Green Country | 0% | 4,784 | 48.9%2,339 | 51.1%2,446 | +2.2 |
| Muskogee | East Central & Kiamichi | 0% | 4,784 | 48.7%2,331 | 51.3%2,453 | +2.6 |
| Osage | Tulsa Metro | 0% | 4,526 | 51.9%2,351 | 48.1%2,175 | +3.9 |
| Delaware | North & Northeast Green Country | 0% | 4,267 | 51.5%2,198 | 48.5%2,070 | +3.0 |
| Bryan | South Central & Southeast | 0% | 4,009 | 50.8%2,037 | 49.2%1,972 | +1.6 |
| Cherokee | East Central & Kiamichi | 0% | 3,879 | 51.2%1,987 | 48.8%1,892 | +2.4 |
| Pontotoc | South Central & Southeast | 0% | 3,233 | 51.5%1,665 | 48.5%1,568 | +3.0 |
| Sequoyah | East Central & Kiamichi | 0% | 3,233 | 50.2%1,623 | 49.8%1,609 | +0.4 |
| Okmulgee | Tulsa Metro | 0% | 3,103 | 50.0%1,553 | 50.0%1,551 | +0.1 |
| Custer | Western I-40 & Southwest | 0% | 2,974 | 62.2%1,850 | 37.8%1,125 | +24.4 |
| Garvin | South Central & Southeast | 0% | 2,974 | 51.1%1,519 | 48.9%1,455 | +2.1 |
| Mccurtain | South Central & Southeast | 0% | 2,974 | 48.9%1,456 | 51.1%1,519 | +2.1 |
| Jackson | Western I-40 & Southwest | 0% | 2,457 | 61.7%1,516 | 38.3%941 | +23.4 |
| Kingfisher | Western I-40 & Southwest | 0% | 2,457 | 61.9%1,521 | 38.1%936 | +23.8 |
| Beckham | Western I-40 & Southwest | 0% | 2,328 | 61.2%1,424 | 38.8%904 | +22.4 |
| Caddo | Western I-40 & Southwest | 0% | 2,328 | 62.3%1,451 | 37.7%877 | +24.7 |
| Seminole | South Central & Southeast | 0% | 2,328 | 51.4%1,196 | 48.6%1,131 | +2.8 |
| Woodward | Northwest Plains | 0% | 2,069 | 52.5%1,085 | 47.5%984 | +4.9 |
| Adair | East Central & Kiamichi | 0% | 1,940 | 51.2%993 | 48.8%947 | +2.4 |
| Marshall | South Central & Southeast | 0% | 1,940 | 49.4%958 | 50.6%981 | +1.2 |
| Mcintosh | East Central & Kiamichi | 0% | 1,810 | 49.6%898 | 50.4%912 | +0.8 |
| Murray | South Central & Southeast | 0% | 1,810 | 50.5%915 | 49.5%895 | +1.1 |
| Pawnee | Tulsa Metro | 0% | 1,810 | 50.6%915 | 49.4%895 | +1.1 |
| Washita | Western I-40 & Southwest | 0% | 1,810 | 61.2%1,107 | 38.8%703 | +22.3 |
| Craig | North & Northeast Green Country | 0% | 1,681 | 49.8%838 | 50.2%843 | +0.3 |
| Major | Northwest Plains | 0% | 1,681 | 52.2%878 | 47.8%804 | +4.4 |
| Woods | Northwest Plains | 0% | 1,681 | 50.5%849 | 49.5%832 | +1.0 |
| Atoka | South Central & Southeast | 0% | 1,552 | 51.9%806 | 48.1%746 | +3.8 |
| Haskell | East Central & Kiamichi | 0% | 1,552 | 51.6%801 | 48.4%751 | +3.2 |
| Noble | Northwest Plains | 0% | 1,552 | 52.1%808 | 47.9%743 | +4.2 |
| Nowata | North & Northeast Green Country | 0% | 1,552 | 50.0%776 | 50.0%775 | +0.1 |
| Ottawa | North & Northeast Green Country | 0% | 1,552 | 52.4%814 | 47.6%738 | +4.9 |
| Pushmataha | South Central & Southeast | 0% | 1,552 | 50.1%777 | 49.9%774 | +0.2 |
| Texas | Northwest Plains | 0% | 1,552 | 50.4%782 | 49.6%770 | +0.8 |
| Choctaw | South Central & Southeast | 0% | 1,422 | 50.6%719 | 49.4%703 | +1.2 |
| Johnston | South Central & Southeast | 0% | 1,422 | 51.8%737 | 48.2%686 | +3.6 |
| Blaine | Western I-40 & Southwest | 0% | 1,293 | 62.2%804 | 37.8%489 | +24.4 |
| Love | South Central & Southeast | 0% | 1,164 | 52.9%615 | 47.1%549 | +5.7 |
| Dewey | Northwest Plains | 0% | 905 | 52.4%474 | 47.6%431 | +4.8 |
| Grant | Northwest Plains | 0% | 905 | 52.6%476 | 47.4%429 | +5.2 |
| Hughes | South Central & Southeast | 0% | 905 | 50.4%456 | 49.6%449 | +0.8 |
| Latimer | East Central & Kiamichi | 0% | 905 | 52.1%472 | 47.9%433 | +4.2 |
| Okfuskee | East Central & Kiamichi | 0% | 905 | 52.7%477 | 47.3%429 | +5.3 |
| Alfalfa | Northwest Plains | 0% | 776 | 52.1%404 | 47.9%372 | +4.2 |
| Ellis | Northwest Plains | 0% | 776 | 51.5%399 | 48.5%377 | +2.9 |
| Beaver | Northwest Plains | 0% | 647 | 53.1%343 | 46.9%303 | +6.1 |
| Coal | South Central & Southeast | 0% | 647 | 51.6%333 | 48.4%313 | +3.1 |
| Greer | Western I-40 & Southwest | 0% | 647 | 61.9%400 | 38.1%246 | +23.8 |
| Jefferson | South Central & Southeast | 0% | 647 | 49.9%323 | 50.1%324 | +0.2 |
| Kiowa | Western I-40 & Southwest | 0% | 647 | 61.4%397 | 38.6%249 | +22.8 |
| Roger Mills | Western I-40 & Southwest | 0% | 647 | 62.9%407 | 37.1%240 | +25.8 |
| Cimarron | Northwest Plains | 0% | 388 | 53.6%208 | 46.4%180 | +7.2 |
| Cotton | Western I-40 & Southwest | 0% | 388 | 62.3%242 | 37.7%146 | +24.6 |
| Harper | Northwest Plains | 0% | 388 | 51.8%201 | 48.2%187 | +3.6 |
| Tillman | Western I-40 & Southwest | 0% | 388 | 61.3%238 | 38.7%150 | +22.6 |
| Harmon | Western I-40 & Southwest | 0% | 259 | 61.0%158 | 39.0%101 | +22.0 |
South Carolina · August 25•Statewide forecast · no county model
U.S. Senate Special Republican Runoff
TPSI gives Graham a 58% chance of winning the nomination. There is no TPSI poll of this runoff. The starting probability was a desk judgement from the first round, the endorsements since, and how runoff electorates have behaved here before — not a survey. The margin below is derived from the probability, not measured.
Win probability
58%Graham
Reported results
Polls close at 7:00 PM ET, in CLOSED. Candidates appear as soon as the feed carries them.
Forecast
Projected turnout279,046
First round465,076
Votes counted0
Est. reporting0%
Projected marginGraham +1.7
Margin SD±8.5
County results · 0 of 46 reporting
No county has reported yet.
This map shows reported votes only. There is no county forecast for South Carolina, because there is no poll to build one from.
GrahamNormanreported votes only
Turnout assumes 60% of the 465,076 ballots cast in the first round — runoffs draw a fraction of the electorate that produced them. Everything on this panel other than the reported votes rests on that assumption and on the win probability. First-round result, endorsement movement and historical runoff dropoff. No TPSI poll of the runoff was fielded.
Run it yourself
The scenario engine behind this county baseline is public. Set candidate performance by region and demographic, move turnout, and watch all 77 counties recompute against the same model this board projects from.
Method
Reported vote is solid. Projected share is muted and dashed, on the same 0 to 100% scale, and is the sum of all 77 county projections. Estimated reporting is the share of expected vote, not precincts. Race calls come from AP through CivicAPI; TPSI projects independently once the leader’s margin clears three standard deviations of the outstanding vote and at least 35% is counted, and those projections are labeled as ours.
The statewide probability is not built from the county intervals. County errors share one dominant statewide swing — if the model has Drummond wrong, it has him wrong in all 77 counties at once — so adding county variances independently would understate the real spread by roughly an order of magnitude. Probability comes from the statewide model, whose margin carries a standard deviation of 7.9 points against a projected margin of 1.4. That standard deviation contracts with the outstanding vote as counties report. Field dates August 21–23, 2026, n=460, ±4.7.
South Carolina is a different kind of number. No TPSI survey of that runoff exists. Its starting 58% figure was a desk judgement, and the 1.7-point opening margin was derived from it by assuming a 8.5-point standard deviation — wider than Oklahoma’s precisely because there is no poll underneath it. As votes are counted that prior gives way to the returns in proportion to how much has reported, so the probability shown is now still entirely the prior. There is no South Carolina county forecast, and the map on that panel paints reported votes only.
All candidate estimates on this page are built from pre-election day data. The Oklahoma survey closed before polls opened and has not been updated since; nothing a voter did today is in the shares. Turnout is a registration-and-history prior of 375,000 ballots against 403,099 cast in the June first round and 1,282,277 registered Republicans.
A turnout figure lower than a general election, or lower than the primary that produced this runoff, is not a forecast that fewer people will vote. It defines which electorate the shares describe: a runoff is a small, self-selected slice of a closed primary electorate, and every percentage here is a share of that slice and of nothing wider. Read these numbers as a description of the people who actually turned out, not as a prediction about the state.
© 2026 The Public Sentiment InstitutePowered by CivicAPI