Ten thousand simulated races for Oklahoma governor, county by county. My model says Mazzei. Change what it assumes to see what Munson would need.
The model splits Oklahoma into the three metro counties (Oklahoma, Tulsa and Cleveland) and the other 74. It first measures how Democratic or Republican each county runs against the rest of its group, averaging the 2024 presidential and 2022 governor results. A county that ran 5 points more Republican than rural Oklahoma in 2024 and 3 points more in 2022 starts about 4 points more Republican than rural Oklahoma. Of the methods I tested on 28 past elections, this one predicted best.
Democrats running for governor in Oklahoma do better than Democrats running for president, so the model estimates how far Munson will run ahead of Harris. It uses two inputs. The first is national mood, measured by the generic congressional ballot, now near D+9. The last four midterms show what each point of national mood is worth in Oklahoma: not much, but something. The second is Oklahoma’s own recent history, the average gap between the Democratic and Republican candidates in the last three governor’s races. The model weights that history more heavily than national mood: 60% to 40% in the metro counties and 70% to 30% in the other 74.
Each of the 10,000 simulations redraws the inputs, their weights and turnout, then adds statewide, regional and county error. Oklahoma has little reliable polling, so the model leaves its own inputs uncertain, and tests on 2010–22 results in Oklahoma, Kansas and Texas say ranges at least this wide are needed.
County returns: Oklahoma State Election Board. Presidential county returns: MIT Election Data and Science Lab; tonmcg county-level results (2024). Historical county panel: Algara and Amlani. Generic ballot: RealClearPolitics final averages.