Fred CornellIndependent polling averages

France  ·  2027 presidential election, first round

Marine Le Pen leads the first-round polls with 34.3%.

Jean-Luc Mélenchon is second on 16.8%. The average weights the 40 polls of the last 12 months, the most recent most, with fieldwork to 22 October 2026. Shares are of voters who expressed a choice, as published, with no don't-know line. Vote share only, not a forecast.

Current

Updated 1 Nov 2026Fieldwork to 22 Oct 202640 polls from 9 pollsters in 12 monthsFirst round 18 April 2027How this is made

How the candidates have moved

The daily average for the six candidates with the highest average. The shaded bands show the range of recent polls for the two leading candidates. A dashed line rests on fewer than 3 polls' worth of weight; a line stops where a candidate went more than 21 days without a poll.

Édouard Philippe (14.0% in the table): 3.4 points higher in scenarios without Gabriel Attal.

Gabriel Attal (7.4% in the table): 6.1 points higher in scenarios without Édouard Philippe.

Édouard Philippe (14.0% in the table): 2.1 points higher in scenarios without Raphaël Glucksmann.

Bruno Retailleau (7.1% in the table): 2.8 points higher in scenarios without Édouard Philippe.

Measured within the same polls: in each poll that tests a candidate both with and without the other, the difference between those scenarios, averaged over the polls behind the candidate's figure in the table, with the same weights. Shown when the difference is 2 points or more and at least half the polls of the last 90 days test both ways. Scenarios without a candidate often differ in other names too, so this describes the polls, not what would happen if the candidate did not stand.

fredcornell.comFieldwork to 22 Oct 2026How this average is built

Latest polls

The eight most recent first-round polls in the average, newest first. All polls in the list

  • OpinionWay 21 to 22 Oct 2026 · n=973 · 1 scenario

    Le Pen 36.0Mélenchon 16.0Philippe 12.0Glucksmann 13.0Attal 4.0Retailleau 6.0

  • Ipsos 21 to 22 Oct 2026 · n=1,500 · 4 scenarios

    Le Pen 33.5Mélenchon 17.0Philippe 13.0Glucksmann 13.0Attal 4.0Retailleau 7.5

  • Cluster17 19 to 22 Oct 2026 · n=3,000 · 3 scenarios

    Le Pen 32.0*Mélenchon 19.2*Philippe 16.5*Glucksmann 13.5*Attal 12.0*Retailleau 8.3*

  • Harris Interactive 19 to 21 Oct 2026 · n=2,094 · 5 scenarios

    Le Pen 35.0Mélenchon 17.0Philippe 13.0Glucksmann 12.0Attal 5.0Retailleau 6.0

  • Ifop 17 to 19 Oct 2026 · n=1,928 · 8 scenarios

    Le Pen 35.8Mélenchon 16.4Philippe 12.6Glucksmann 12.6Attal 5.3Retailleau 5.6

  • Elabe 14 to 16 Oct 2026 · n=1,701 · 6 scenarios

    Le Pen 32.5Mélenchon 15.5Philippe 13.5Glucksmann 15.5Attal 4.5Retailleau 7.5

  • YouGov 12 to 15 Oct 2026 · n=1,094 · 5 scenarios

    Le Pen 35.5Mélenchon 14.0Philippe 9.0Glucksmann 12.0Attal 6.0Retailleau 5.5

  • Cluster17 14 to 15 Oct 2026 · n=1,263 · 5 scenarios

    Le Pen 34.4*Mélenchon 19.7*Philippe 18.5*Glucksmann 13.7*Attal 17.0*Retailleau 12.2*

Each figure is the mean of the poll's scenarios that test both Édouard Philippe and Gabriel Attal in which the candidate appears, to one decimal. A star marks a figure from a poll with no such scenario, or a candidate tested only in other scenarios: it is the mean of all the poll's scenarios in which the candidate appears. Scenarios is how many the poll published after those testing another National Rally name were left out. A dash means the candidate was not tested.
PollsterSampleLe PenMélenchonPhilippeGlucksmannAttalRetailleauScenarios
OpinionWay21 to 22 Oct 202697336.016.012.013.04.06.01
Ipsos21 to 22 Oct 20261,50033.517.013.013.04.07.54
Cluster1719 to 22 Oct 20263,00032.0*19.2*16.5*13.5*12.0*8.3*3
Harris Interactive19 to 21 Oct 20262,09435.017.013.012.05.06.05
Ifop17 to 19 Oct 20261,92835.816.412.612.65.35.68
Elabe14 to 16 Oct 20261,70132.515.513.515.54.57.56
YouGov12 to 15 Oct 20261,09435.514.09.012.06.05.55
Cluster1714 to 15 Oct 20261,26334.4*19.7*18.5*13.7*17.0*12.2*5

How these figures are made

61 of the 62 first-round polls on the source page are used; 1 is not.

  • One observation per poll, one field. French polls test several scenarios, each with a different set of candidates, and a candidate's figure depends on who else is in it. Each poll counts once. The figures are read from the scenarios that test both Édouard Philippe and Gabriel Attal (our reference field, version RF-1, declared 11 October 2026): for each candidate, the mean of their figures across those scenarios. Of the 27 polls of the last 90 days, 5 have no such scenario (Cluster17, 31 Aug to 1 Sep 2026; Cluster17, 15 to 16 Sep 2026; Odoxa, 24 Sep to 7 Oct 2026; Cluster17, 14 to 15 Oct 2026; Cluster17, 19 to 22 Oct 2026); for them the mean of all scenarios is used, and the figures are starred in the latest polls. A candidate tested in a poll only outside those scenarios (Hollande in 8 of the 27 polls; Lisnard and Villepin in 3 each; Faure and Ruffin in 2 each; Arthaud, Guedj, Maire, Maurel, Polony and Royal in 1 each) is also starred: for that poll the mean of all the poll's scenarios in which they appear is used. The field changes only when a candidate declares or withdraws, as set out in advance in our method.
  • The average. The UK method: each poll is weighted by recency (the weight halves every 14 days) and by sample size (square root of the sample over 1,000, capped at 3 times). Pollsters are not rated: France has no pollster scores yet, so no poll is weighted up or down for who conducted it. The range of recent polls is the average plus or minus 1.645 times the larger of two figures: the weighted standard deviation of the poll figures, and 1.5 points. So it is never narrower than plus or minus 2.5 points. It is a spread of the polls, showing how much they disagree; it is not a confidence interval or a margin of error. Every listed candidate has an average.
  • Second place. The line under the headline compares second and third, poll by poll, when the order changes if every scenario is counted instead of the reference field, or when the gap between them, measured poll by poll, could be zero: 0 lies within its weighted mean plus or minus 1.645 weighted standard deviations. Counting every scenario instead keeps Jean-Luc Mélenchon above Édouard Philippe. Poll by poll, over the 39 polls of the last 12 months that test both, Jean-Luc Mélenchon leads Édouard Philippe by 2.8 points on average; allowing for how much that gap varies, it runs from 0.4 to 5.2 points.
  • Who paid for the poll. Each poll of the last 90 days is checked against its notice at the Commission des sondages, which names who paid for it. A poll paid for by a party, a candidate or a campaign team is left out of the average and listed below with its sponsor. Of the 27 polls of the last 90 days, 26 have been checked (last on 28 October 2026); for 1 the notice did not settle it, and it stays in. None of those checked was paid for by a party or candidate.
  • Who is shown. A candidate is listed when tested in at least 3 of the polls of the last 90 days, and charted when also tested in at least half of them; the chart shows the six highest. Not listed, tested in fewer polls: Bruno Le Maire (1 of 27), Emmanuel Maurel (1 of 27), Jérôme Guedj (1 of 27), Natacha Polony (2 of 27), Ségolène Royal (2 of 27) and Xavier Bertrand (1 of 27).
  • Limits. The figures describe the contest in which both Édouard Philippe and Gabriel Attal stand; the lines under the table show how much a figure moves when one of a pair is not tested. Averages are comparable over time only while pollsters' scenario mixes stay stable. Pollsters are not adjusted for house effects. A cell published as <1% is counted as 0.5 (7 such cells in the polls used). Wikipedia shows 3 of them as <0.5%, where the Commission notices print <1%. Between polls the average carries the last polls forward; where a candidate went more than 21 days without a poll, no line is drawn across that stretch (a lone poll shows as a dot), and a dashed line means the average rests on fewer than 3 polls' worth of weight. The longest stretch with no first-round poll in the last year runs from 31 October 2025 to 27 February 2026. Wikipedia is a lagging source that people edit. This page does not forecast the first round, the run-off or the winner.

Marine Le Pen is tested in 27 of the 27 first-round polls with fieldwork in the last 90 days. Jordan Bardella is tested in 0 of them; in our source, he was last tested in a first-round poll ending 9 December 2024. The page charts Marine Le Pen only: a scenario that tests another National Rally name is left out of the averages. This is a choice about what is charted and says nothing about who is eligible or will stand.

Left out: 7 scenarios and 1 poll
Scenarios left out of the poll's figures. A poll that still has another scenario stays in the average.
FieldworkPollsterScenarioReason
6 to 9 Dec 2024Ifop4tests Jordan Bardella, not Marine Le Pen
11 to 12 Sep 2024OpinionWay5tests Jordan Bardella, not Marine Le Pen
16 to 18 Apr 2024Ifop2tests Jordan Bardella, not Marine Le Pen
16 to 18 Apr 2024Ifop1tests Jordan Bardella, not Marine Le Pen
9 to 10 Jan 2024Ifop2tests Jordan Bardella, not Marine Le Pen
9 to 10 Jan 2024Ifop1tests Jordan Bardella, not Marine Le Pen
12 to 13 Apr 2023OpinionWay1tests Jordan Bardella, not Marine Le Pen
Polls not used at all.
FieldworkPollsterReasonScenarios
12 to 13 Apr 2023OpinionWayevery scenario tests a different National Rally candidate1

Source: Wikipedia, "Opinion polling for the 2027 French presidential election", licensed under CC BY-SA 4.0, retrieved 1 November 2026. We read the poll tables and do not alter the figures.