TopBadger wrote: Thu Jul 04, 2024 9:17 pm
Hoping they get less than 13 seats in reality... also heard they're on for higher % of national vote than LD's, thats concerning.
Even if exit polls are constituency by constituency, there's still self-selection effects among participants. I imagine/guess/hope that Reform voters are likely to be louder about their support than e.g. Tories or reluctant Labour voters, and more likely to take part.
And I assume that different methods are used for polling overseas, proxy and postal voters. If they are included at all.
Sciolus wrote: Thu Jul 04, 2024 9:16 pm
Presumably the BBC/ITV/Sky exit poll seat predictions are obtained by taking the vote share and plugging into a model similar to the EC or FT ones, so has the same level of uncertainty in terms of mapping vote share to seats? Or do they have consituency-specific data?
It is all constituency based. It's not a % converted into seats.
My avatar was a scientific result that was later found to be 'mistaken' - I rarely claim to be 100% correct
ETA 5/8/20: I've been advised that the result was correct, it was the initial interpretation that needed to be withdrawn
Meta? I'd say so!
Sciolus wrote: Thu Jul 04, 2024 9:32 pm
OK, and it says they polled 130 constituencies -- presumably the ones most likely to be bellwethers -- so likely to be reasonably accurate.
Which leaves a lot of room for error where the are specific issues with specific candidates - e.g. for Reform.
ETA: this was cross-posted with LPM above. It may or may not apply to that candidate.
Don't forget, they're at a sweet spot in the polling where there's a lot of seats very, very close. A little bit of error in the polling will mean the seat numbers change quite a bit.
Sciolus wrote: Thu Jul 04, 2024 9:32 pm
OK, and it says they polled 130 constituencies -- presumably the ones most likely to be bellwethers -- so likely to be reasonably accurate.
Which leaves a lot of room for error where the are specific issues with specific candidates - e.g. for Reform.
ETA: this was cross-posted with LPM above. It may or may not apply to that candidate.
El Pollo Diablo wrote: Thu Jul 04, 2024 1:37 pm
I think that the standard Labour overestimation in the final week is actually already being priced in in these final polls, so I'm not minded to suggest much change. If I was to pick a single poll there which I think represents the most likely outcome, it'd be the BMG one - a 180-200 seat majority.
Pretty chuffed with that pick tbh. The FT prediction from the bmg poll was Bob on
I voted Labour, non-tactically. If I was just voting to deny a Tory win I would have voted Lib Dem, but I genuinely don’t even know the name of the Tory candidate. They can’t have tried very hard. My outgoing MP is William Wragg, who wasn’t standing again.
My Lib Dem candidate was on the news for a gaffe effectively calling Scousers thieves. Even if it was just a silly joke I couldn’t vote for her after that. I was thinking of voting tactically but not any more. The other major reason was that I want this seat to be a Labour/Lib Dem battle and put the Tories into 3rd.
where once I used to scintillate
now I sin till ten past three
lpm wrote: Thu Jun 20, 2024 2:46 pm
23:30 First result from somewhere like Sunderland where they have a silly race to be first
23:31 Embarrassing over-analysis of first result by everyone
Ah, we're at this stage.
A hell of a lot of Reform votes in that Sunderland result.
lpm wrote: Thu Jun 20, 2024 2:46 pm
23:30 First result from somewhere like Sunderland where they have a silly race to be first
23:31 Embarrassing over-analysis of first result by everyone
Ah, we're at this stage.
A hell of a lot of Reform votes in that Sunderland result.
But extrapolating from that, Labour will win 650 seats (this is how it works, isn't it?)
"My interest is in the future, because I'm going to spend the rest of my life there"
JLPartners had 38/23/17/13/5 which isn't bad. Norstat was 37/24/16/11/6 which is pretty good too. They're the best ones I can see. PeoplePolling was 36/16/20/10/9, so a little silly overall.
