Seven YouGov state surveys released on October 7 provide a dated look at Senate preferences in Florida, Kansas, New Hampshire, Michigan, Nebraska, Ohio and Texas. Interviews ran from September 18 through October 5, with different fieldwork dates in each state.
These are individual polls from one organization, not Caldwell averages or election forecasts. The figures below retain every published response category, including undecided voters.
Florida
Fieldwork: 2026-09-28 to 2026-10-05. Survey sample: 3813. Senate-question sample: 3812.
Kansas
Fieldwork: 2026-09-24 to 2026-10-04. Survey sample: 2255. Senate-question sample: 2253.
New Hampshire
Fieldwork: 2026-09-28 to 2026-10-05. Survey sample: 1520. Senate-question sample: 1519.
Michigan
Fieldwork: 2026-09-22 to 2026-10-04. Survey sample: 3949. Senate-question sample: 3947.
Nebraska
Fieldwork: 2026-09-24 to 2026-10-03. Survey sample: 1809. Senate-question sample: 1808.
Ohio
Fieldwork: 2026-09-18 to 2026-09-30. Survey sample: 3780. Senate-question sample: 3777.
Texas
Fieldwork: 2026-09-28 to 2026-10-05. Survey sample: 3625. Senate-question sample: 3623.
Caldwell analysis
Michigan and Nebraska each show a one-point difference between the two leading candidates. Those small differences should not be read as secure leads. Nebraska also has a different party structure: Dan Osborn is an independent, so his support is not labeled as Democratic.
These surveys share a pollster and a likely-voter modeling approach. Their seven state results are not seven independent confirmations of a national trend. The fieldwork windows also differ; later campaign developments are outside these observations.
How the surveys were conducted
YouGov recruited respondents by text from registered-voter lists and conducted interviews online. Likely-voter weights account for past voting, demographics and reported likelihood of voting. Initially undecided Senate respondents were asked their preference and allocated. Percentages can sum to 99% or 101% because of rounding. Each state document explains its weighting and sampling error.
Original YouGov release. Source: YouGov plc, 2026, © All rights reserved.