24JUL

Pollsters split on the Burnham bounce

2 min read

12:51UTC

Deltapoll put Labour up eight points to 28%, level with Reform, days after Burnham took office; YouGov's near-simultaneous poll left Reform clear on 23% to Labour's 20%.

← Back to England's council map redrawn as five sue Jump to analysis ↓

PoliticsDeveloping

Key takeaway

Deltapoll ties Labour with Reform while YouGov leaves Reform clear, with fieldwork timing explaining the divergence.

Deltapoll put Labour up eight points on 28%, level with Reform UK, in its first survey since Burnham entered Downing Street, with fieldwork running 17 to 20 July 2026⁠¹. YouGov, polling 19 to 20 July, recorded far less movement: Reform 23%, the Conservatives 21%, Labour 20%⁠². Both cannot be right about the size of the shift.

Fieldwork timing explains most of the gap. Deltapoll's sample was gathered fully after Keir Starmer's resignation had been absorbed by voters; YouGov's window opened while it was still settling. One survey shows Burnham erasing a Reform lead in a fortnight, the other a ripple.

Reform's five-week polling plateau had already begun to crack before Burnham arrived, so the direction of travel is shared even where the magnitude is disputed. A second poll confirming Deltapoll's reading would turn a fieldwork artefact into a genuine bounce; until one lands, the scale of Labour's recovery stays an open question.

Deep Analysis

In plain English

Two polling companies asked people who they would vote for in the days after Andy Burnham became prime minister, and got very different answers. Deltapoll found Labour jumped eight points to 28%, level with Reform UK. YouGov found barely any change, with Reform still ahead on 23% and Labour on 20%. This kind of gap between pollsters is normal in the days right after a big political change. It usually takes several more polls, over several weeks, before a clear picture of any genuine shift emerges.

Source Landscape

This story draws on neutral-leaning sources

Primary parallel: Theresa May's post-coronation polling in July 2016 also showed a sharp but short-lived Conservative bounce, recorded at different magnitudes by different pollsters before it faded within two months, the same caution Curtice is applying here.

Counter-parallel: Gordon Brown's 2007 'Brown bounce' was larger and more consistent across pollsters than the divergent Deltapoll and YouGov readings of Burnham, and it held for several months before collapsing, showing new-leader bounces do not always fade or diverge quickly.

Consensus view: Sir John Curtice has argued that a single week's post-leadership-change polling should be read cautiously, because honeymoon bounces for new party leaders often halve within a month once the novelty of a leadership change wears off.

Counter-view: The Electoral Reform Society has argued the more important story in the Deltapoll and YouGov gap is not the size of Labour's bounce but the fact that two reputable pollsters can diverge by eight points on the same event, a sign of how volatile a four-party field has become under first-past-the-post.

Key tension: Whether the Deltapoll-YouGov gap reflects genuine uncertainty about Burnham's bounce, or simply normal house-effect variation between the two pollsters that would exist regardless of who leads Labour.

Sources:Deltapoll·YouGov

Mentions:Deltapoll →YouGov →Andy Burnham →Reform UK →Keir Starmer →Gordon Brown →Electoral Reform Society →Downing Street →Labour Party →Conservative Party →John Curtice →

First Reported In

Update #13 · England's council map redrawn as five sue

Deltapoll· 24 Jul 2026

Read original →

Causes and effects

Caused by

Reform plateau breaks after five weeks

Reform's plateau breaking apart in mid-July set up the divergent Deltapoll and YouGov readings of Burnham's bounce.

Occurred 13 Jul 2026

Read story →

This Event

Pollsters split on the Burnham bounce

Deltapoll and YouGov disagree on whether Burnham has erased Reform's lead or barely moved it.

Led to

A second pollster puts Labour ahead of Reform

The unresolved split between pollsters on the Burnham bounce is settled in direction, not size, by More in Common's new lead.

Occurred 27 Jul 2026

Read story →

Different Perspectives

Scottish National Party / John Swinney

Swinney told the SNP conference in Aberdeen on 17 October that Scotland is ready for independence, but set no referendum date, bill or Section 30 request. He launched a new campaign instead and extended the £2 bus fare cap to the north east from April 2027.

Liberal Democrats / Ed Davey

Davey ruled out a national partnership with the Greens on 19 October after their Brighton conference passed a "Zionism is racism" motion, and wrote to two Jewish community bodies. He left untouched the councils the two parties already run together.

Reform UK / Richard Tice

Tice alleged ITV's reporter used false details to get a press-office job and said ITV had "likely broken the law", telling members not to speak to it until it apologises. Reform also opened investigations into members, with some suspensions, so the party is contesting the broadcaster and disciplining councillors at once.

Rhun ap Iorwerth / Welsh Government (Plaid Cymru)

Ap Iorwerth presented six priorities to the Senedd on 22 September without numeric targets or a named partner party, promising to work across the chamber where there is common ground. He says his government will be judged on results.

Michael Marra / Scottish Labour

Marra won the leadership on 19 September, and three days later his 17 MSPs backed Swinney's right-to-decide motion once their child-poverty amendment was written in. He says Scottish Labour's task is to tell voters what it stands for.

Andy Burnham / Labour Party

Burnham told conference in Liverpool on 29 September that Labour's next manifesto will commit to changing the electoral system, and invited every party into a National Commission on Electoral Reform. He has not named a replacement system, and no chair or remit has been published.

Written with AI from Lowdown's published reporting. Not individually reviewed by an editor. Editorial standards.