Final-Week Senate Polls: Maine, Michigan, Ohio and Texas Compared
Dated editorial analysis; figures in this article are not recalculated automatically. Visit the linked race pages for current data.
With election day close, the final published polls in the most closely watched Senate contests are arriving in quick succession. This piece collects the verified polls in our tracker that ended in the 14 days before our November 2 data date for four states: Maine, Michigan, Ohio and Texas. We report the numbers, show a simple average where there are two or more polls, and explain what the figures can and cannot say. We do not predict who will win.
How the averages are made
For each race we take verified polls only and average the Democratic and Republican shares across polls that ended in the 14-day window. This is a plain mean. It does not weight by sample size or pollster quality, and it ignores differences in method, so treat it as a summary of what was published rather than a forecast. Where only one poll qualifies, we report it as a single result.
Maine
Two verified polls ended in the window in the Maine Senate race, where Sen. Susan Collins (R), in office since 1997, faces Troy Jackson (D), a former president of the state Senate:
- Data for Progress (online, 652 likely voters, ending October 27): Jackson 50%, Collins 45%.
- Marist Poll (multi-mode, 1,449 likely voters, ending October 25): Jackson 49%, Collins 47%.
The simple average is Jackson 49.5% and Collins 46.0%, a gap of 3.5 points. The two polls use different methods and sample sizes, and the gap in one is five points while the other shows two, which is the kind of spread normal sampling variation produces. Collins's earlier races are discussed in our Maine polling history piece.
Michigan
The Michigan Senate seat is open because Sen. Gary Peters (D) is retiring. Abdul El-Sayed (D), a physician and former Detroit health director, faces Mike Rogers (R), a former U.S. Representative. Two verified polls ended in the window:
- Mitchell Research (online panel and text-to-web, 636 likely voters, ending October 27): El-Sayed 51%, Rogers 44%.
- Glengariff Group (live-interviewer phone, 600 likely voters, ending October 22): El-Sayed 51.4%, Rogers 45.1%.
The average is El-Sayed 51.2% and Rogers 44.6%, a gap of about 6.6 points. The two polls agree closely even though one is online and the other is by telephone. In the Michigan governor's race, the Glengariff poll alone had Jocelyn Benson (D) at 51.4% and John James (R) at 44.5%.
Ohio
Only one verified poll ended in the window for the Ohio Senate special election: Suffolk University (500 likely voters, ending October 26). It had Sherrod Brown (D), a former U.S. Senator, at 47% and Sen. Jon Husted (R), who was appointed to the seat, at 44%. With 500 respondents, the margin of error on each share is roughly four to five points, so a three-point gap in a single survey is well within it. The same poll found Amy Acton (D) at 48% and Vivek Ramaswamy (R) at 44% in the Ohio governor's race.
Texas
Three verified polls ended in the window in the Texas Senate race, between James Talarico (D), a state representative, and Ken Paxton (R), the state Attorney General:
- Quantus Insights (1,221 likely voters, ending October 28): Talarico 47.6%, Paxton 48.7%.
- Texas Southern University (text-to-web, 2,000 likely voters, ending October 25): Talarico 49%, Paxton 47%.
- Emerson (1,000 likely voters, ending October 19): Talarico 48%, Paxton 46%.
The average is Talarico 48.2% and Paxton 47.2%, a one-point gap. Individual polls point in both directions, and each gap is smaller than its margin of error. In the Texas governor's race, the two polls that covered it averaged Gina Hinojosa (D) at 46% and Gov. Greg Abbott (R) at 49.5%. Our earlier Texas poll article covers the polls from the previous window.
Reading the four together
- Averages are not margins of safety. A 6-point average gap with two polls is more informative than a 1-point gap, but neither tells us about turnout or late changes.
- Few polls, wide uncertainty. Ohio has one poll in the window, so no average exists.
- Method differences matter. Phone, online and text-based polls reach different people and weight them differently.
- Polls are snapshots. Interviews ended between October 19 and 28, and none of them measure what happens afterward.
To see how a Senate majority is decided, read our Senate control math explainer.
Data date: November 2, 2026. Averages are simple means of verified polls ending within 14 days of this date. This article describes published polls and does not predict outcomes or endorse any candidate.
Relevant race pages: Maine Senate · Michigan Senate · Ohio Senate · Texas Senate · Maine Governor · Michigan Governor · Ohio Governor · Texas Governor