Trump's approval sits at 32% on the eve of the midterms: what the final Reuters/Ipsos poll really says

What does the last poll before voting day actually say?

On Monday, November 2, 2026, hours before Americans vote in the midterm elections, Reuters published the final Reuters/Ipsos reading of the campaign. Some 32% of respondents said they approved of the job Donald Trump was doing in the White House, a figure that has barely moved in a month: unchanged from a week earlier and one percentage point above the record-low 31% he hit two weeks ago. The number is not a headline in search of a story. It is the closing measurement of a presidency's standing at the moment the verdict is cast.

The poll's secondary numbers are the ones campaigns study. Among registered voters, 46% said they would pick Democrats in congressional elections against 38% for Republicans; the generic-ballot question that, nationally, captures which party voters want running Congress. The eight-point Democratic advantage is down from an 11-point lead a week earlier: the race is tightening even as the president's own number stands still. And the intensity gap is stark: half of likely voters said their Tuesday vote would be against Trump; 13% said it would be in support of him.

The survey itself is the familiar Reuters/Ipsos instrument: a five-day nationwide poll, 1,765 US adults interviewed online from October 29 to November 2, with a margin of error of about two percentage points. It arrived on the same day Trump closed out his campaign with a 54-minute speech at the Lincoln Memorial Reflecting Pool in Washington; the final event of a nearly monthlong campaign blitz aimed at saving the Republican majorities in the House and Senate. The stakes he described were personal: he has repeatedly told supporters to vote as if they are voting for him, and warned that a Democratic Congress could try to impeach him for a third time. For a fuller map of what money has already bought this cycle, read our breakdown of the record-shattering $11.9 billion midterms.

What do 1,765 interviews tell us beyond the headline number?

The headline, 32%%, is only the surface of the poll. Underneath it, the interviews draw a coherent portrait of an electorate that has made up its mind about the economy and is still negotiating its feelings about the president. Three findings carry the weight.

First, the discontent is economic before it is partisan. Some 78%% of voters said the rising cost of living would be very important or the most important issue in deciding how to vote. Asked which party is better on the issue, they chose the Democrats over the Republicans, 43%% to 32%%; an 11-point trust gap on the question the election will be decided on. Reuters connected that pressure to fuel: energy prices have leapt 20%% over the past year, measured by the US Consumer Price Index, since the US-Israel war with Iran began. The war's domestic aftershock, in other words, is arriving at the pump and the checkout line.

Second, the president is the ballot even when his name is not. Trump's approval has drifted downward since the February 28 launch of the US-Israel war on Iran, and the poll suggests voters now treat Tuesday as a confidence vote. That 50%%-to-13%% intensity split, against versus for, is the number strategists in both parties will cite on Tuesday night. It is also the context for Trump's own framing: he told supporters the midterms were being fought over him, and that a Democratic Congress meant impeachment proceedings.

Third, the generic ballot is moving in the Republicans' direction without moving the president's number. Democrats' 46-to-38 edge is real, but the three-point weekly narrowing, from 11 to 8, is the grain of hope the GOP will cling to. Independent readings agree on the direction of travel while differing on the size: an AtlasIntel poll over the weekend gave Democrats a 14-point generic-ballot lead, the multi-poll averages sit between the Reuters/Ipsos 32%% and higher readings, and forecasters give Democrats an 89%% chance of taking the House. The poll, in short, says the same thing the averages say, only louder.

What can a national poll never settle before Tuesday?

A national poll is a thermometer, not a map. The Reuters/Ipsos reading measures the country's mood; Congress is decided district by district, and the generic-ballot question, useful as it is, does not capture the state-by-state and district-by-district contests that determine control of the House and Senate. A voter in a safe seat and a voter in a swing district answer the same question; their answers weigh very differently on Tuesday night.

We also do not know what the final 24 hours did. The poll's field period ended Monday; anything that moved after the interviews, late-breaking news, weather, turnout operations, is invisible to it. The narrowing of the generic-ballot gap from 11 points to 8 could be the start of a late Republican surge, a statistical wobble inside the margin of error, or the normal tightening of a closing campaign. The poll cannot adjudicate between those stories.

There is a deeper unknown the poll's own history raises: the intensity gap. Half of likely voters say they are voting against the president; but anger is not turnout. Midterm electorates are smaller and older than presidential ones, and the gap between telling an interviewer you oppose the president and standing in line on a Tuesday is where elections are actually decided. The poll measures the first; only Tuesday measures the second.

And one political unknown hangs over everything the poll implies: Trump's warning that a Democratic Congress would move to impeach him a third time. Whether that threat mobilizes his base, alienates swing voters, or simply describes what happens next, no survey taken before the votes are counted can answer. The related uncertainty inside the Republican camp is already visible; see how Trump's party is scrambling to stop a Democratic wave in the campaign's final days.

Why is 32%% a five-alarm number for the party in power?

Presidents lose midterms; that is the iron rule of American politics, and Trump acknowledged it himself on Monday: "For some reason, the sitting president doesn't win the midterms. But we gave this a very strong shot." But the depth of the unpopularity matters, because it determines whether Tuesday is a correction or a rout. The New York Times polling tracker, cited in USA Today's coverage, put Trump's approval average at 36%%; the lowest for any president heading into a midterm election in 80 years.

The practical stakes are the gavels. If Democrats take the House, the second half of Trump's term becomes an exercise in oversight: subpoenas, investigations, and the impeachment threat he has already priced into his rhetoric. If they take the Senate too, forecasters give them roughly two-in-three odds, the legislative agenda of a second Trump term ends on Tuesday night. A Republican hold, conversely, would convert the same 32%% into a footnote: proof that a president can be unpopular and still keep his coalition's machinery intact.

There is also the matter of what the number certifies. The Reuters/Ipsos poll is not an outlier; it is the low end of a range that includes the RealClearPolitics average at 39.3%%, the Times at 37%%, and the Silver Bulletin at 38.4%%; every major average has the president underwater by 20 points or more. When every thermometer agrees on the direction, the argument moves from whether the fever is real to what it does to the patient. That is why this poll, the last before voting, is being read in both parties' war rooms as the closest thing to a final briefing the campaign will get.

How did the economy become the entire campaign?

Strip the poll to its load-bearing wall and you find a single issue: the cost of living. Seventy-eight percent of voters call it decisive; Democrats hold an 11-point trust advantage on it; and the 20% year-on-year rise in energy prices since the US-Israel war with Iran began is the mechanism the poll's own framing identifies. The war that began on February 28 was a foreign-policy event; its domestic afterlife is a gallon of petrol and a grocery receipt. That is the through-line from the war to the ballot box; and it explains why the president's approval has slid steadily since February even as the military picture has shifted week to week. The economic aftershocks of the conflict, already reshaping household budgets, are now reshaping the electorate.

Against that backdrop, the closing imagery of the campaign reads as a study in two different elections. Trump spent his final major speech on Monday at the Lincoln Memorial Reflecting Pool; 54 minutes on Washington's renovations, new sod, a White House helipad, a proposed triumphal arch, and a pool whose months-late, over-budget refurbishment became its own small saga. Reuters' dispatch from the scene noted the dissonance plainly: facing elections that could cost his party Congress, the president talked pipes and paint. Whether that was confidence, defiance, or misreading the room is the argument the poll's numbers have already started.

The counter-argument the White House can still make is the tightening one. An 11-point generic-ballot deficit shrinking to 8 in a week, a forecast model that still has to survive contact with 435 individual races, and a midterm electorate that has surprised pollsters before; these are not nothing. The market's own midterm-eve verdict, with oil plunging on detente hopes, suggests at least some investors are pricing a calmer Wednesday. But the poll's core message survives every caveat: the electorate has rendered its economic judgment, and it is waiting only for the count to make it official.

What does the morning after look like if the House flips?

The second-order consequences of Tuesday begin where the poll's numbers end. A Democratic House majority, the outcome the generic ballot and the forecasters point to, would hand the opposition the investigative machinery of the chamber: committee chairmanships, subpoena power, and the calendar. Trump's pre-emptive warning of a third impeachment was not analysis; it was inoculation, an attempt to make the threat look like partisan reflex before it can be presented as accountability. The warning also tells you what the White House fears most: not losing a policy debate, but losing control of the process.

There is a subtler second-order effect inside the Republican coalition. A defeat of the size the polls imply, the kind that follows a 32%% approval rating into a midterm, historically triggers the blame cycle: the president's allies argue the message was right and the messengers were weak; the party's pragmatists argue the message was the problem. The intensity data in this poll (50%% voting against Trump personally) gives the pragmatists their exhibit. The tightening data (11 to 8) gives the loyalists theirs. The party's internal argument on Wednesday morning is already written; only the winner is unknown.

And there is the polling profession's own second-order reckoning. If Tuesday's results land near the Reuters/Ipsos numbers, the industry's post-2020 repair job, the reason 2018 and 2022 midterm polling was closer to the mark, will be judged a success. If the results land four points to the Republican side, as in 2020, the profession faces a different reckoning: the possibility that Trump-era polling error is structural, not cyclical. Either way, the poll's 32%% will be cited for years as the number that either called the wave or missed it.

What should a reader watch for on Tuesday night?

Three things, in order. First, the early House calls. The generic ballot is national; the House is 435 separate elections. Watch the swing-district results in the first wave of called races; they will tell you faster than any model whether the 46-to-38 national edge is converting into seats. The forecasters' math (Democrats around 89% to take the House, roughly 234 seats projected) is the benchmark; the early calls are the reality check.

Second, the turnout shape. The poll's most lopsided finding is also its most fragile: half of likely voters say they are voting against Trump. Likely-voter screens are models, not measurements, and midterm electorates have a habit of differing from the models. If turnout in the Republican-leaning demographics runs hot, the narrowing from 11 to 8 points could be the real story of the night. If the anti-Trump intensity converts cleanly, the night is the wave the averages predict. The bellwether races to watch include Michigan, where four voters' question, can America still talk to itself?, is also the election's question.

Third, the concession and claim speeches. Trump has already framed the election as a personal referendum and pre-warned of impeachment; his Tuesday-night posture, defiance, delegation of blame, or a pivot to governing with a divided Congress, will set the terms of the next two years before a single committee gavels in. The poll gave both parties their briefing. Tuesday night gives the verdict.