Dylan Difford
@dylandifford.bsky.social
15K followers 290 following 2.2K posts
YouGov data journalist • Elections, polls, voting systems • "I like people, places and things" • [email protected]
Posts Media Videos Starter Packs
Pinned
dylandifford.bsky.social
A bit behind schedule, but how voters moved in the year since the 2024 election.

Labour facing same splintering of the last govt: a significant bloc crossing floor to primary electoral opposition, with a numerically larger chunk moving to opponents on same side of spectrum, plus many 'don't knows'.
dylandifford.bsky.social
too kind, though I worry that my repeated refrain of "it's basically just the cost of living and public services" might drive some insane
dylandifford.bsky.social
the party of individual responsibility
Reposted by Dylan Difford
dylandifford.bsky.social
Which party underperformed the most in the 2025 local elections?

Arguably Reform UK. A 🧵 on what the locals tell us about a key and neglected dimension in analysis of the current political situation.
dylandifford.bsky.social
Certainly not fully, but there should be some clues in how those that do vote vote. Also think Con DKs are probably an easier bet, given how much they are basically less enthused Con loyalists, while Lab DKs are more distinct.
dylandifford.bsky.social
I think the devolved elections' higher order nature should mean that we get a better view of how don't knows behave when squeezed, even if there are limitations on what lessons can be carried over to England.
Reposted by Dylan Difford
dylandifford.bsky.social
Furthermore, the locals do illustrate an enthusiasm gap with Labour. Those with a Reform VI were more likely to turnout at all (75% vs 68%) and much more likely to vote for ‘their’ party (62% vs 47%), while Labour-originated don’t knows were particularly likely to not bother voting.
dylandifford.bsky.social
A lot of the replies any time this is brought up often have a clear undertone of 'Reform's a blip, Labour will inevitably recover, no need to confront any of the obvious issues'. I think a lot of people want to run away from the fact Labour's programme falls well short of what they'll need by 2029.
dylandifford.bsky.social
The lessons from Autumn 2021 should be:
> Govts should address their severe weaknesses + have a coherent programme
> Pay attention to the cost of living
> Journalists should stop huffing the glue of conference speeches being meaningful

And not:
> Everything's probably fine, election's not tomorrow
dylandifford.bsky.social
Like, the reason the tweet was notable isn't because of 'oh, how things change', it was because it was completely detached from the reality of the moment, which was when multiple "this government is irreparably fucked" alarm bells were beginning to ring.
dylandifford.bsky.social
Continuing to enjoy the irony of the "squatting like a giant toad" being repurposed as an "actually, the govt aren't doing that badly, no need to think about all our weaknesses" comfort blanket.
dylandifford.bsky.social
If anything, the most notable thing about it is the irony in its repurposing as an "actually, the govt aren't doing that badly, no need to think about all our weaknesses" comfort blanket.
dylandifford.bsky.social
It was an appalling call at a time when the tide had already turned against the Tories. They were already falling behind in VI polls, Boris' ratings were in the toilet and govt competency/economy ratings were worse. What it really shows is the flaws in giving credence to conference performances.
dylandifford.bsky.social
(Also, still absolutely insane how few BritPol people were, after May 2023 and 2024, not picking up on Labour having a very efficient vote, but one which clearly had an enthusiasm issue. They were losing a *quarter* of their VI poll in council elections!)
dylandifford.bsky.social
(In fact, if anything, they've become too good at presaging. Before the last election, I was talking about how people were wrong to think the Greens couldn't win multiple MPs or that Leicester E in particular was trending Tory, but always caveated it by saying they were stories for 2029/34!)
dylandifford.bsky.social
Also a useful reminder of why British politics people should spend more 12 hours a year even vaguely interested in local elections.

They've become quite good a presaging trends in wider electoral behaviour and helping understand strengths and weaknesses of VI polls.
dylandifford.bsky.social
Which party underperformed the most in the 2025 local elections?

Arguably Reform UK. A 🧵 on what the locals tell us about a key and neglected dimension in analysis of the current political situation.
dylandifford.bsky.social
No, quite the opposite. Reform UK voters are abnormally engaged and enthusiastic for their party, at least at the moment. It's more a case of them being more likely to 'vote' in polls, because other parties' voters are more unsure atm, while they are 'certain' who they'll vote for.
dylandifford.bsky.social
It's interesting because, as I show downthread, voting intention groups haven't changed their behaviours, it's just that the voting intention groups have changed.

Are we seeing a lag between local election behaviour and general election behaviour, with the latter now catching up?
dylandifford.bsky.social
This means we are still seeing intra-bloc vote switching between Labour, Lib Dem and Green voters. However, the net effect is less tactical, with the number of LD and Grn local election voters returning to Labour for their Westminster VI halved relative to previous years.
dylandifford.bsky.social
What's also notable about VI vs local behaviour this year is, despite significant changes in the sizes of the VI groups, those with a Con/Lab/LD/Grn VI still voted in the locals broadly identically to their counterparts in 2023 and 2024.
dylandifford.bsky.social
Given that around a fifth of Reform UK’s current coalition are previous non voters, 2.5x that of any party at the last election, this is important. If they can’t turn all of them out, that’s a further deflation of their VI figure.
dylandifford.bsky.social
There is also evidence that Reform UK are better – tho still limited – at turning out non voters. 35% of 2024 non voters who say they intend to vote Reform UK actually backed them where they could, relative to 22% for Labour among their ‘gains’ from non voters.
dylandifford.bsky.social
Not all of this is new - Labour-originated DKs were also particularly unlikely to turnout in previous cycles, with this likely part of the reason for Labour's underperformance in the 2024 general election. Starmer's Labour has an issue enthusing those who 'don't dislike' the party.
dylandifford.bsky.social
Furthermore, the locals do illustrate an enthusiasm gap with Labour. Those with a Reform VI were more likely to turnout at all (75% vs 68%) and much more likely to vote for ‘their’ party (62% vs 47%), while Labour-originated don’t knows were particularly likely to not bother voting.
dylandifford.bsky.social
However, none of this changes that a Reform underperformance at present is still a Reform lead. Their lead might be being inflated, but there is no methodologically justifiable way of adjusting it out of existence. Those with stated Vis will have to change their minds.
dylandifford.bsky.social
And there’s little chance of this anti-Reform skew changing, with few Con, Lab or LD-originated DKs having positive views of Nigel Farage’s party.