Republican versus Democratic primary turnout in every U.S. House, Michigan Senate and Michigan House district that has territory in Oakland County — measured against the same August primaries in 2022 and 2018, and translated into a range for November 3, 2026.
County lines are the only geography that does not move between 2018 and 2026 — every district map in Michigan was redrawn at least once in between. So the cleanest like-for-like trend is Oakland County as a whole, measured on the governor's primary, the one contest on every Oakland ballot in all three gubernatorial cycles.
Read 2022 carefully. Gretchen Whitmer was unopposed in that year's Democratic primary while five Republicans fought for their nomination, so 2022's Democratic total is a floor, not a measure of enthusiasm. 2018 and 2026 both had contested primaries on both sides, which makes them the honest pair to compare.
Every district below has at least some Oakland County territory. Michigan's county-level results file splits each district by county, so the Oakland share of a district that crosses county lines is a measured number rather than an estimate. Select a row for the full history, county composition and candidate list.
Ratings restate the projected share as a margin — Safe ≥20 pts, Likely ≥10, Lean ≥4, Toss-up below that — and are not a second opinion. The 95% band is wider than the “Lean” threshold, so a Lean rating means the point estimate leans, not that the seat is decided. Shares are of the two-party vote and exclude minor-party and write-in votes. “Lean vs state” is the district's primary Democratic share minus the statewide primary Democratic share that year — a district-level measure with the year's national mood divided out.
A primary Democratic share means nothing on its own — it moves with whatever else is on that year's ballot. Measured relative to the statewide primary, though, it tracks the November result closely. The model below is fitted on completed cycles and leaves November's statewide number as an input you set, not a number it invents.
Michigan's Bureau of Elections publishes a county-level CENR file for each election — one row per candidate per county. Every figure here is summed from those files; nothing is scraped from a news feed or re-keyed by hand. The parse is checked against the totals the state prints for itself, and the statewide anchors reproduce the published results exactly (2024 U.S. Senate: Slotkin 2,712,686, Rogers 2,693,680).
The model is
district Nov − statewide Nov = a + b × (district primary − statewide primary).
Working in differences removes the year: 2026's Democratic primary surge shows up in the
statewide anchor as well as in each district, so it does not get double-counted into a
forecast. What survives is the district's lean, which is the durable part.
Fitted separately, the intercept is −2.9 in 2018, −8.3 in 2022 and −1.9 in 2024. 2022 is the outlier because Whitmer was unopposed: the statewide Democratic primary vote collapsed to 46%, so every district looks artificially Democratic measured against it. 2026 had contested primaries in both parties for both governor and U.S. Senate — the 2018 and 2024 situation — so those are the default calibration, with 2022 available as a sensitivity test rather than folded into the average.
Redistricting. Congressional lines have been stable since 2022, but the Michigan
Senate was redrawn for 2026 and the Michigan House for 2024. Where a district number
refers to different ground across cycles the table flags it; those comparisons are
between different places that share a number.
Primary structure. A district where one party had a bruising primary and the other
had none will overstate the first party's strength. The residual band absorbs this on
average, not case by case.
A primary electorate is not a general electorate. 390,467 Oakland ballots were
cast in August against 635,275 in the 2022 midterm November and 612,317 in 2018 — the
November electorate is roughly 60% larger and composed differently. The model is a
relationship observed across past cycles, not a poll of the people who will actually
turn out.
2026 figures are unofficial. All 83 counties are reporting, but the Board of State Canvassers has not certified; small movements are still possible. 2018, 2022 and 2024 are certified and final.