Latest polls

Each poll as published, with its pollster, fieldwork dates, sample size and a link to the source.

How to read a poll

Sample size is how many people were asked. More people means less random error, never none.
±3 is the margin of error: the real figure is likely within 3 points either way of the one shown.
Two polls can differ because they were taken on different dates, with different methods and different samples.
est. means the institute did not publish a margin of error, so it is estimated from the sample size.
The average smooths this out: it combines recent polls, so one unusual poll cannot swing the picture.

CountryElectionPolling instituteFromTo

Sort

4 polls · page 1 of 1 · Clear filters

Polls with institute, dates, sample, margin of error, methodology and results
Institute	Date	Fieldwork	Sample	MoE	Method	Voting intention	Source
Zenith Research 🇺🇸 United States · 2026 United States Senate election in South Dakota	30 Sept 2026	24 Sept 2026 – 27 Sept 2026	500	±4.4	—	Rounds (R) 53.0% Bengs (I) 47.0%	Source for Zenith Research poll (opens in a new tab)
Zenith Research 🇺🇸 United States · 2026 Maryland gubernatorial election	12 Aug 2026	27 Jul 2026 – 3 Aug 2026	800	±3.5 est. (Not published: estimated from sample size (95%, simple random sample))	—	Moore (D) 58.0% Cox (R) 31.0%	Source for Zenith Research poll (opens in a new tab)
Zenith Research 🇺🇸 United States · 2026 Michigan Senate Democratic primary	31 Jul 2026	26 Jul 2026 – 30 Jul 2026	536	±4.2 est. (Not published: estimated from sample size (95%, simple random sample))	—	El-Sayed (D) 59.0% Stevens (D) 41.0%	Source for Zenith Research poll (opens in a new tab)
Zenith Research 🇺🇸 United States · 2026 United States Senate election in Michigan	17 Jun 2026	11 Jun 2026 – 14 Jun 2026	602	±4.0 est. (Not published: estimated from sample size (95%, simple random sample))	—	El-Sayed (D) 45.0% Rogers (R) 42.0%	Source for Zenith Research poll (opens in a new tab)

Select two or more polls to compare them side by side.

Latest polls

Each poll as published, with its pollster, fieldwork dates, sample size and a link to the source.

How to read a poll

Sample size is how many people were asked. More people means less random error, never none.
±3 is the margin of error: the real figure is likely within 3 points either way of the one shown.
Two polls can differ because they were taken on different dates, with different methods and different samples.
est. means the institute did not publish a margin of error, so it is estimated from the sample size.
The average smooths this out: it combines recent polls, so one unusual poll cannot swing the picture.