But we know that putting all the parties available in the question means you overstate minor parties and so a more accurate question gets a less accurate result. So it's possible this does the same the other way round if the default of Con/Lab etc isn't there. Anyhow, glad I'm not doing it anymore!

I don't work on voting intention any more but when we went through this with CUK/TIG we ran parallel polls for a few weeks, version with them and a version without. Then they flamed out before we had to make a decision.

I’m curious what thresholds Your Party needs to reach for pollsters to start asking/reporting about them routinely

Consistent with the belief that you win reelection based on how well you fine-tune your messaging on X, Y, Z issues, rather than whether people feel better off

Bluesky's LLM hate is kind of bizarre. ChatGPT isn't perfect (it's bad at certain types of task and lots of people use it badly) but it can be super useful. Why is it so popular on here to signal that you refuse to use it?

A meta-analysis in Psychological Bulletin suggests short form video use is associated with poorer cognition (especially attention and inhibitory control) and mental health (esp stress and anxiety) psycnet.apa.org/fulltext/202...

Danish right-wing populists are pretty much back to the same level as before Mette Frederiksen came to power. Soc Dems are competitive because they lead a left-wing bloc in a PR system. Doesn't transfer neatly to the UK, where losing liberal voters to Greens/etc would cost Labour seats to Reform

Column on the obvious. Loveeeee the cartoon

Off-year elections shouldn't be thought of as predictions of midterms *because things can change a lot in a year.* What they are useful for is a reality check for the other evidence of what's happening now, and the results were consistent with the polls and with the huge protests.

What sort of midterms do the Democrats' New Jersey and Virginia midterms point to? I had a look using historic results and wound up with a central estimate of D+6 and a confidence interval from R+5 to D+16. Not exactly conclusive! owenwinter.co.uk/2025/11/10/d...

Thread.

What happened in Tuesday's elections? Here's our anatomy of a blue wave: www.economist.com/graphic-deta...

NYC election by largest ethnic group, by precinct: Asian - 54% Harris, 48% Mamdani (6pt difference) Hispanic - 69% Harris, 59% Mamdani (11pt) White - 62% Harris, 44% Mamdani (19pt) Black - 87% Harris, 64% Mamdani (23pt) www.economist.com/graphic-deta...

Andrew Cuomo won precincts with over two-thirds car ownership by a large margin, Zohran Mamdani won parts of the city where fewer people own a car

Alright but a broader level the governments line on NY elections being "the words on the red boxes say United Kingdom, not United States" is...kinda valid

Curious how Republican candidates in Virginia and New Jersey are feeling about the government shutdown. It's clear now that Trump's approval has taken a hit (across multiple pollsters) right before the first major elections of his second term

In our polling with YouGov, since the start of his second term Donald Trump's net approval has fallen 17pts among white Americans (+17 to -1), 28pts among Hispanic Americans (-9 to -37) and 38pts among black Americans (-36 to -74)

Got to wonder what effect this sort of coverage will have on Reform’s brand over the next few years before a general election

It's not fashionable, I know, but sometimes when events in the world are truly awful, terrible, horrific it seems best to watch the news reflectively - as professionals do their best to report for us, as details emerge - without attempting to use those events to validate one's owns views.

My only take on the Dem moderation debate is that data people are far too confident you can find the secret recipe for electability with polls or election results

Our cover this week “Europe should take heart and recognise its own strength. Its military budget is already four times larger than Russia’s; its economy is ten times larger. Far from shying away from a financial contest with the Kremlin, Europe should embrace it—and win the war”

We haven't had a poll of London recently but based on MRPs + nationwide moves I would guess VI here is something like Lab 26, Green 19, Reform 19. Would be a huge swing from the 2022 local elections if it holds until May

Wrote about the dominance of "victims" in our political discourse, which included this fab chart @owenwntr.bsky.social www.economist.com/britain/2025...

With the official prognosis giving one additional seat to PVV and one to FvD compared to the ext poll, the narrative should really not be that the Netherlands/Jetten has beaten the radical right. The far right bloc remains remarkably stable.

7% van de D66 stemmers kwamen van de PVV. Dat is ongeveer 130,000 mensen.

Wat kunnen we in de exit polls zien? Centrum-rechts (CDA, VVD, CU, SGP, 50PLUS) was nog nooit zo klein. De linkse partijen (inclusief D66) is het grootste blok. Radicaal rechts handhaaft zich.

Worth saying (if the exit poll is right) that the far-right in the Netherlands hasn't really shrunk, just shifted from PVV to JA21 and FvD

Final poll from Peil shows PVV in steep decline. Final average PVV 25 (-12) GL-PVDA 24 (-1) D66 23 (+14) CDA 19 (+14) VVD 18 (-6) JA21 9 (+8)

Looks like the far-right PVV is stumbling a bit in the final polls for tomorrow's election in the Netherlands. Liberal D66 surging

theme_minimal() instant 10x trustworthiness of results

Trump's approval with white people has been hovering around 0 since July, but his ratings with black, Hispanic and people of another race are collapsing

Brutal for Trump with young people

Donald Trump's approval rating in our tracker has fallen to -18, the lowest it has ever been (lower than any point in his first term) www.economist.com/interactive/...

'there are too many black people on my television these days' is about as far from legitimate as a concern can be

Coming back to this. The amount of attention Reform gave Caerphilly (and lack of expectation management) suggests their data showed them winning. Worth remembering when we hear similar stories about Reform's advanced data analysis in future!

Mr Caerphilly wins (and Labour collapses to 11% of the vote in a seat it has held for a century) www.economist.com/britain/2025...

Three Northern regions since 1950. Labour's overperformance in the North West is a relatively recent phenomenon, whereas in the North East it peaked in 1997

The MIDLANDS on the other hand

As someone who grew up in the rural north it really cracks me up, usually a weird combo of people who have literally never been to the north and London journalists who last lived there decades ago and need to self mythologise. Oh hang on.

It's funny how much we talk about "the North" in political discourse, given that the three northern regions aren't particularly correlated in election swings

Average of recent polls for the Dutch election on 29th October: PVV — 30 seats (-7 vs 2023) GL-PvdA — 24 (-1) CDA — 23 (+18) D66 — 15 (+6) VVD — 14 (-10) JA21 — 13 (+12) FvD — 5 (+2) BBB — 4 (-3) SP — 4 (-1) PvdD — 3 (-) Volt — 3 (+1) CU — 3 (-) Denk — 3 (-) SGP — 3 (-) 50+ — 2 (+2) NSC — 0 (-20)

Fairly confident that this is the only Senedd by-election (let alone a normal Senedd contest) to have ever featured in the hallowed columns of The Economist… www.economist.com/britain/2025... #devolution #senedd #caerphilly

Super obituary for the now-deceased Labour Safe Seat (1906-2024), not least for the mention of Alfred Onions MP.

that's weird, election twitter told me he was *obviously* more electable than Mills

Caerphilly is on the long list of former Labour safe seats www.economist.com/britain/2025...

I'm always quite sceptical of claims like this. The main barrier to data-led campaigns in Britain is the availability of data on registered voters, not the tech to analyse it

The Plaid candidate in Caerphilly has stood for election to be an MP or MS sixteen times since 1983

Bar chart in the wild!

Labour's polling deficit is the largest for any government at this stage of the parliament since John Major was PM 1993 Major -14 2025 Starmer -10 1971 Heath -10 1980 Thatcher -7 2011 Cameron -6 ~ 1988 Thatcher +7 1966 Wilson +8 2016 May +9 2002 Blair +11 1998 Blair +22

I know statistics are poorly understood and are misused. But facts matter Between the 1991, 2001, 2011 and 2021 Census *every* ethnic group in the UK has become *less* geographically segregated and *all* groups, majority and minorities, are more likely to interact with people not like them