1/4 A difficult defeat for the Greens. Polanski lost 33% to Labour’s 45%: a 3-point gap in May (inc CPA) is now 12, despite standing their leader and flooding the seat with activists. The “22-point rise/swing on 2024” ignores Feinstein’s 19%. Against the combined 29%, it’s about +3.

If you are headed to the Labour Party Conference and would like to catch up, or just want to be briefed on our pre conference research, go here for details: eepurl.com/F2ohRItFCX

The Clacton by-election today is a bit of an odd one - many usual parties missing - the main "challenger" is a unique type of novelty candidate this time and so on. With that in mind, we conducted a mini-experiment yesterday, extending our work beyond the 500+ responses collected mid campaign and conducted 100+ new interviews, but this time, not prompting any candidate or party names, just allowing respondents to tell us whom they would vote for (or not) - without being provided with (read) a full list (our preferred method) - not feasible in this case given 34 candidates/parties to read out. We'll post results later today. Our *prior* research is here: survation.com/clacton-by-ele…

It’s gone full silly season in the Survation press office:

A couple of better polls for Labour from us now, this is the highest Survation has had Labour since 3 June 2025, and the first Labour polling lead since 3 May 2025:

Makerfield result: LAB 54.8%, REF 34.5%, RES 6.8% CON 2.2%. Survation May 26th-1 Jun polling had Lab 49, Reform 39 - Labour lower by 6pts, Reform higher by 5pts. This was the "Burnham blow-out" as flagged on 16 June. The unweighted data recall clustered around Burnham's mayoral figure not the 2024 GE - and weighting back to GE PV suppressed Labour VI and inflated Reform across all pollsters. As warned, the unweighted numbers were closer to the truth!

A shared methodological puzzle across all Makerfield polls: unweighted Labour 2024 recall clusters nearer to the Burnham Mayoral figure (66%, Wigan), not the GE figure (45% for Josh). Weighting corrects to GE PV. But if the recalled vote isn't the GE vote, the correction misfires - suppressing Labour VI and inflating Reform. There's fighting talk from the PM today but what's the plan in the event of a complete Burnham blow-out?

Makerfield by-election briefing. Data tables and all prior materials are on the link below: Points to note. *This is a poll of 505 people in Makerfield this week and should be seen as an indicative early look. *By-election headline voting intention with named candidates as were available is the 3 point Labour lead over Reform from table V5 *Reform (45%) have a 9 point lead over Labour on a generic Westminster polling question *Burnham has strong leads among women and younger groups, Kenyon fares better with older voters in general and leads among men. *How the Green and Restore campaigns pan out could be key. There's also at least a 7% Green Westminster vote share up for grabs. *Do get in touch if you have question or concerns and have a lovely Sunday! mailchi.mp/survation/whats-t…]

Makerfield: We provided an update to media yesterday as forecasts have been made by others in political data using Survation's "What if Andy Burnham Stood" polling - collected after the Gorton & Denton by-election - to show the effect the Greater Manchester Mayor could have on the race in Makerfield. The briefing document, estimates and methodology is linked below. cdn.survation.com/wp-content…

I'm seeing the "Burnham wouldn't win a by-election" line again - as a reminder, we recontacted residents of Gorton & Denton and found that if Burnham had been the Labour candidate, it would not have been close:

We'll be sharing our 2026 Holyrood MRP model that we've been running for some time this evening (time not quite specified yet). There will a final call for constituency and list national vote share, and estimates for all seats. Usual caveats apply, looking forward to the "if this MRP shows this specific result in this specific place, this is a BAD MRP" hot takes.