Has Yougov had some serious Corbynista infiltration.
Seriously it goes against the positioning/canvassing of @HYUFD, Aaron, David Herdson, Big John Owls (NED lost, Lab resources back to Chesterfield), @MarqueeMark (Torbay Lib/Lab -> Con switchers), David Herdson (Yorkshire), Andrea Leadsom (Out to Bolsover), Chuka Ummuna (In Hyndburn), local elections, by-election models, leader ratings, my office...
It goes against EVERYTHING.
What has Aaron being saying about Don Valley? Any chance of a win?
Be fair. Any candidate who says anything other than "it's very close, your vote could make the difference!" is quite simply mad.
He will post again when the election is over; not before.
Meanwhile, THAT Corbyn IRA attack video has now been seen 4.377 million times.....
Perhaps it's actually boosting his popularity!
Those are all paid views. Many if not most will be for 3 seconds or less. 4.3m is a crap number, it could easily be 400,000 people who check Facebook x10 a day.
According the Times article this was how they calculated their results:
YouGov used data from the Office of National Statistics, the British Election Study, and past election results. It then estimated the number of each type of voter in each constituency. Combining the model probabilities and estimated census counts allowed the pollster to produce what it hopes is a fairly accurate estimate of the number of voters in each constituency intending to vote for a party on each day.
Lol. That's it, I've given up on the polls. No one has a f-----g clue do they?
Tbf, most polls show a Conservative majority. YouGov are going against not just ICM but most polls here with this.
And that VI intention doesn't point to a hung parliament either, btw.
Has Yougov had some serious Corbynista infiltration.
You mean they've been massively infiltrated in the period since they were showing the Tories 20 points ahead?
I don't believe infiltrations. But as good a campaign as Labour have had, and as bad the Tories have had, and even assuming Corbyn and Labour are more popular than previous polls and and election results would indicate, this is too big a switch to believe.
Given not a single poll (bar possibly the one 5%) has suggested a hung parliament, and we are 9 days out, isam is right: 1/5 on Con majority is good value.
The poll consensus is a lead of 9% or so, and we know the Labour position is heavily reliant on high turnout from young voters.
Who would be forming the next government in such an outcome?
Conservatives + DUP fall short of 326 but so do Labour + SNP + Green + Lib-Dem?
Conservative minority.
Con minority with Boris and another election in October?
I think if there was yet another election the country would revolt. Brenda from Bristol said it best.
Maybe... But the people will have voted for a Parliament that is literally ungovernable whichever way you cut the numbers and at a time of national crisis (Brexit) so whether the people like it or not if we finish up with a result anywhere close to this we're going back to the polls in six to twelve months...
Has Yougov had some serious Corbynista infiltration.
Seriously it goes against the positioning/canvassing of @HYUFD, Aaron, David Herdson, Big John Owls (NED lost, Lab resources back to Chesterfield), @MarqueeMark (Torbay Lib/Lab -> Con switchers), David Herdson (Yorkshire), Andrea Leadsom (Out to Bolsover), Chuka Ummuna (In Hyndburn), local elections, by-election models, leader ratings, my office...
It goes against EVERYTHING.
Yougov know their members and when they joined, and a good deal about their geography.
It may simply be that voters don't like what they see of May and don't want her manifesto.
According the Times article this was how they calculated their results:
YouGov used data from the Office of National Statistics, the British Election Study, and past election results. It then estimated the number of each type of voter in each constituency. Combining the model probabilities and estimated census counts allowed the pollster to produce what it hopes is a fairly accurate estimate of the number of voters in each constituency intending to vote for a party on each day.
Lol. That's it, I've given up on the polls. No one has a f-----g clue do they?
Tbf, most polls show a Conservative majority. YouGov are going against not just ICM but most polls here with this.
And that VI intention doesn't point to a hung parliament either, btw.
That's, perhaps, the most interesting part.
Maybe Labour's vote share is really efficient this time around?
According the Times article this was how they calculated their results:
YouGov used data from the Office of National Statistics, the British Election Study, and past election results. It then estimated the number of each type of voter in each constituency. Combining the model probabilities and estimated census counts allowed the pollster to produce what it hopes is a fairly accurate estimate of the number of voters in each constituency intending to vote for a party on each day.
Eh?
I presume it means they're running the demographics of their national polls, and applying them to the demographics of each constituency.
If they're doing that uniformly then, for one, it doesn't take into account how differently marginal seats are fought and contested compared to safe seats.
lol I had planned to not watch election night or only watch a little at the start. But if the polls continue to narrow then it's going to be a must watch.
If the polls continue to disagree so violently, you'll have given up all faith in polling full stop, will listen to the exit poll when in the past you might have thought "oh okay, I can go to bed now"... and be on such tenterhooks over whether it is reliable or not, that you'll be forced to stay up all night.
According the Times article this was how they calculated their results:
YouGov used data from the Office of National Statistics, the British Election Study, and past election results. It then estimated the number of each type of voter in each constituency. Combining the model probabilities and estimated census counts allowed the pollster to produce what it hopes is a fairly accurate estimate of the number of voters in each constituency intending to vote for a party on each day.
Eh?
I presume it means they're running the demographics of their national polls, and applying them to the demographics of each constituency.
People keep saying that the young people who are huge Corbynistas are concentrated in a few safe Labour seats -- but the reverse is also true of the Tory firewall of pensioners, the biggest numbers of them tend to be concentrated in seats which are already safe Tory.
Most polls now are showing that Labour leads with the middle-aged, and it's that age bracket which dominates in a lot of the marginal seats.
Most polls I've seen show the Conservatives leading with voters aged 45 and above.
Given not a single poll (bar possibly the one 5%) has suggested a hung parliament, and we are 9 days out, isam is right: 1/5 on Con majority is good value.
The poll consensus is a lead of 9% or so, and we know the Labour position is heavily reliant on high turnout from young voters.
Has Yougov had some serious Corbynista infiltration.
Seriously it goes against the positioning/canvassing of @HYUFD, Aaron, David Herdson, Big John Owls (NED lost, Lab resources back to Chesterfield), @MarqueeMark (Torbay Lib/Lab -> Con switchers), David Herdson (Yorkshire), Andrea Leadsom (Out to Bolsover), Chuka Ummuna (In Hyndburn), local elections, by-election models, leader ratings, my office...
It goes against EVERYTHING.
Yougov know their members and when they joined, and a good deal about their geography.
It may simply be that voters don't like what they see of May and don't want her manifesto.
To this extent? How often has it been shown that people tend to stick to their original positions, they would switch, in a manner of weeks, from suggesting a massive Tory lead to barely a lead and small to big Tory losses?
A manifesto outright promising to eat the first born of every family wouldn't see such a switch.
So what is YouGov doing that Electoral Calculus, for example, isnt?
Go on, have a guess. What are *all* the polling companies doing in 2017 (well, nearly) that only some did in 2016 EUREF and few in 2015 GE?
Yes, they are using models.
They have recruited some cheeky cherubs with MScs in Data Science or Machine Learning with new laptops that their Mum bought them, with R uploaded and wheeling thru the HighPerformanceComputing and MachineLearning views of CRAN. They're doing Hamiltonian Monte Carlo, they're having a whale of a time, the clients are impressed, and that shit looks really good on your CV, trust me. You can point to stuff on screens. Woo.
There is one teeny problem.
Any model is dependent on assumptions about past behavior, and if those behaviors change (eg bigger than expected youth turnout) then the model is fucked. This isn't an abstract problem, it happened with ComRes and their Voter Turnout Model in 2016 EURef. So they've replaced something which isn't dependent (asking people about their Likelihood To Vote on a 1-10 scale) with something that is dependent (a voter turnout model)
On the specific case of YouGov, they're using another model to deduce seats from votes. Which is poling Pelion upon Ossa if you ask me: why in the name of Almighty God are they producing seat estimates at all?!
According the Times article this was how they calculated their results:
YouGov used data from the Office of National Statistics, the British Election Study, and past election results. It then estimated the number of each type of voter in each constituency. Combining the model probabilities and estimated census counts allowed the pollster to produce what it hopes is a fairly accurate estimate of the number of voters in each constituency intending to vote for a party on each day.
Lol. That's it, I've given up on the polls. No one has a f-----g clue do they?
Tbf, most polls show a Conservative majority. YouGov are going against not just ICM but most polls here with this.
And that VI intention doesn't point to a hung parliament either, btw.
It gives the Tories 338 seats, although Yougov must be projecting the Tories get votes where they don't benefit them.
The model is based on 50,000 interviews over the course of a week, with voters from a panel brought together by YouGov. This allows the pollster to assess the intention of every type of voter, from where they live to how they voted in the EU referendum, their age and social background, to weight the results.
Even putting the YouGov model to one side - this is based on a 4% Con lead which is lower than every single other published poll.
Yet we are now living in times in which a 4% Tory lead doesn't sound out of the realms. The Tory lead has been continuing to shrink week on week, just at different rates depending on the pollster.
According the Times article this was how they calculated their results:
YouGov used data from the Office of National Statistics, the British Election Study, and past election results. It then estimated the number of each type of voter in each constituency. Combining the model probabilities and estimated census counts allowed the pollster to produce what it hopes is a fairly accurate estimate of the number of voters in each constituency intending to vote for a party on each day.
Lol. That's it, I've given up on the polls. No one has a f-----g clue do they?
If this model always had Leave ahead last year, why was Peter Kellner so confident that Remain would win, even after the first set of results had been declared?
So what is YouGov doing that Electoral Calculus, for example, isnt?
On the specific case of YouGov, they're using another model to deduce seats from votes. Which is poling Pelion upon Ossa if you ask me: why in the name of Almighty God are they producing seat estimates at all?!
Taking a gamble on people the only ones who will be right.
I still don't get why the same people would flock to Labour, even with all the gimmes and the positive vision. Not in such mass proportions, without any prior indication they were interested in doing so.
Economy, economy, economy. Earlier, we had people bemoaning why JC was never attacked on this as in Woman's Hour. It is because the Cons no they have no good answers. Neither does Labour, but folk can see more of the same is not working. They will still win easily.
Meanwhile, THAT Corbyn IRA attack video has now been seen 4.377 million times.....
Perhaps it's actually boosting his popularity!
Those are all paid views. Many if not most will be for 3 seconds or less. 4.3m is a crap number, it could easily be 400,000 people who check Facebook x10 a day.
Also Momentum is encouraging people to click on Tory ads. They need to pay for each view.
According the Times article this was how they calculated their results:
YouGov used data from the Office of National Statistics, the British Election Study, and past election results. It then estimated the number of each type of voter in each constituency. Combining the model probabilities and estimated census counts allowed the pollster to produce what it hopes is a fairly accurate estimate of the number of voters in each constituency intending to vote for a party on each day.
Lol. That's it, I've given up on the polls. No one has a f-----g clue do they?
If this model always had Leave ahead last year, why was Peter Kellner so confident that Remain would win, even after the first set of results had been declared?
Such a model might work well for a one-off high-turnout national referendum, with a binary choice, but not for a differential turnout constituency-based General Election where both past voting intentions and loyalties are key.
Someone at YouGov is either very courageous or on very good drugs. If I was the last port of call in the building before they flogged that to the Times, I'd ring them up and say the dog had eaten it.
Either that, or Corbyn has benefitted from a Brexit/Trump tidal wave of People Who Don't Do Politics. Almost certainly not the same one as 11 months ago. Mad.
According the Times article this was how they calculated their results:
YouGov used data from the Office of National Statistics, the British Election Study, and past election results. It then estimated the number of each type of voter in each constituency. Combining the model probabilities and estimated census counts allowed the pollster to produce what it hopes is a fairly accurate estimate of the number of voters in each constituency intending to vote for a party on each day.
Lol. That's it, I've given up on the polls. No one has a f-----g clue do they?
If this model always had Leave ahead last year, why was Peter Kellner so confident that Remain would win, even after the first set of results had been declared?
Why was the genie not even close to visible a few weeks ago though!
It was, I pointed out Mrs May was crap.
But in polls and actual elections, there was no hint people were desperate for a Labour surge. May's crap, but nobody is that crap.
Rump Lab support is one thing. But a surge?
Well why not? It's only going to affect the rich, right?
The only explanation is like the Scot Ref, 5-10% of the voters who are always in the electoral register but never vote [ "they are all the same" ] may come out this time.
The first and only indication will be the turnout. If the turnout is 70-75%, anything is possible.
In GE2015 , not only 66% voted but it had the lowest turnout of 18-24 year olds.
Nobody - I mean NOBODY from ANY party - has suggested that that this matches anything they are seeing on the doorsteps.
No I agree.
Social Care house theft definitely not a game changer.
Agreed. The only way this sort of result could actually happen is if the oldies had shifted. They haven't on the basis of all the available evidence.
I do not know if the C4 News audience tonight was representative or not. But the oldies did not like Glumbucket one jot.
I watched some of that, the old chestnut comment from one old dear "I used to vote Conservative but now after x or y" is so obviously coming from someone who was never a Tory makes programmes like that of little value. I have heard people say words similar to those for the last 25 years. The point being people who change vote like that are not normally so vocal and sometimes bitter. Its usually someone from an anti-tory party venting their distaste.
Incidentally, Momentum are bringing out the big guns - just got an email from Noam Chomsky, inviting to go out knocking doors for Labour. Though given the email text seems pretty much the same as all the other ones they're emailing me, I think it's a safe bet Chomsky may only have contributed the picture of him that you can "Click the image above to pledge some time on June 8th."
According the Times article this was how they calculated their results:
YouGov used data from the Office of National Statistics, the British Election Study, and past election results. It then estimated the number of each type of voter in each constituency. Combining the model probabilities and estimated census counts allowed the pollster to produce what it hopes is a fairly accurate estimate of the number of voters in each constituency intending to vote for a party on each day.
Lol. That's it, I've given up on the polls. No one has a f-----g clue do they?
If this model always had Leave ahead last year, why was Peter Kellner so confident that Remain would win, even after the first set of results had been declared?
According the Times article this was how they calculated their results:
YouGov used data from the Office of National Statistics, the British Election Study, and past election results. It then estimated the number of each type of voter in each constituency. Combining the model probabilities and estimated census counts allowed the pollster to produce what it hopes is a fairly accurate estimate of the number of voters in each constituency intending to vote for a party on each day.
Lol. That's it, I've given up on the polls. No one has a f-----g clue do they?
If this model always had Leave ahead last year, why was Peter Kellner so confident that Remain would win, even after the first set of results had been declared?
Because he was no longer working for YouGov then, so he wasn't seeing those figures
Still, in all this scatter-gun polling, there is one constant - the LibDems are fucked.
Maybe that'll be the shock of the election - all polls, be they labour surge or tory landslide, were bollocks as they missed the LDs actually doing well.
Why the hell not, in a situation where YouGov are saying this outcome is going to happen even though no Labour figure has seriously suggested such an outcome is possible?
According the Times article this was how they calculated their results:
YouGov used data from the Office of National Statistics, the British Election Study, and past election results. It then estimated the number of each type of voter in each constituency. Combining the model probabilities and estimated census counts allowed the pollster to produce what it hopes is a fairly accurate estimate of the number of voters in each constituency intending to vote for a party on each day.
Lol. That's it, I've given up on the polls. No one has a f-----g clue do they?
If this model always had Leave ahead last year, why was Peter Kellner so confident that Remain would win, even after the first set of results had been declared?
Peter Kellner wasn't working for YouGov at the time of the Twitter spat: he'd retired at that point. This is why Stephan Shakespear went apeshit: Kellner was pissing in the tent from outside.
According the Times article this was how they calculated their results:
YouGov used data from the Office of National Statistics, the British Election Study, and past election results. It then estimated the number of each type of voter in each constituency. Combining the model probabilities and estimated census counts allowed the pollster to produce what it hopes is a fairly accurate estimate of the number of voters in each constituency intending to vote for a party on each day.
Lol. That's it, I've given up on the polls. No one has a f-----g clue do they?
If this model always had Leave ahead last year, why was Peter Kellner so confident that Remain would win, even after the first set of results had been declared?
Because he's a prize tit. Remember Ed Miliband PM back in 2015 ?
Has Yougov had some serious Corbynista infiltration.
Seriously it goes against the positioning/canvassing of @HYUFD, Aaron, David Herdson, Big John Owls (NED lost, Lab resources back to Chesterfield), @MarqueeMark (Torbay Lib/Lab -> Con switchers), David Herdson (Yorkshire), Andrea Leadsom (Out to Bolsover), Chuka Ummuna (In Hyndburn), local elections, by-election models, leader ratings, my office...
It goes against EVERYTHING.
I believe each area is different NE Derbyshire has different issues building on green belt and fracking and Engel has played both issues very badly and I think will lose by at least 3000. Derby North canvass apparently much better than 2015. Chris Williamson could be a gain. Chesterfield increased majority. So in 2 of 3 areas YG could be right.
I don't believe YouGov for a minute, but I do believe it increases the chance of (as someone put it upthread) a black swan. If Corbyn turns up to the debate and monsters Rudd, while Farron holds firm, the narrative could swing.
Incidentally, Momentum are bringing out the big guns - just got an email from Noam Chomsky, inviting to go out knocking doors for Labour. Though given the email text seems pretty much the same as all the other ones they're emailing me, I think it's a safe bet Chomsky may only have contributed the picture of him that you can "Click the image above to pledge some time on June 8th."
Maybe he has a domino theory about seats toppling one by one for Labour, like his theory on how communism would sweep the world (It must have happened when I wasn't looking!).
Has Yougov had some serious Corbynista infiltration.
Seriously it goes against the positioning/canvassing of @HYUFD, Aaron, David Herdson, Big John Owls (NED lost, Lab resources back to Chesterfield), @MarqueeMark (Torbay Lib/Lab -> Con switchers), David Herdson (Yorkshire), Andrea Leadsom (Out to Bolsover), Chuka Ummuna (In Hyndburn), local elections, by-election models, leader ratings, my office...
It goes against EVERYTHING.
What has Aaron being saying about Don Valley? Any chance of a win?
Seems to be mostly on focus Tory messages, with the occasional 'great response on doorstep' stuff. He had a tough ask anyway. Even discounting this projection, it's impossible now.
YouGov. Cough. Does anyone believe this stuff? Really?
It's true that the young are in the tank for Corbyn (my office is unbelievable, the bizarre strength of Corbynism among posh, privately educated, twentysomething women) but there simply aren't enough of them in the right places, and many don't vote.
Indeed. Hence why even you Corbynistas haven't been this optimistic. This would have to mean the middle aged or even the old Tory vote is hemorrhaging and no one else has noticed.
Incidentally, Momentum are bringing out the big guns - just got an email from Noam Chomsky, inviting to go out knocking doors for Labour. Though given the email text seems pretty much the same as all the other ones they're emailing me, I think it's a safe bet Chomsky may only have contributed the picture of him that you can "Click the image above to pledge some time on June 8th."
Maybe he has a domino theory about seats toppling one by one for Labour, like his theory on how communism would sweep the world (It must have happened when I wasn't looking!).
"The model had been publicly tested during the EU referendum campaign last year when it always had Leave ahead"
If it was publicly tested where are results published?
How many models did they test? If someone tested (say) 10 models chances are that one of them would have predicted the result very well. But that wouldn't necessarily mean the model has good predictive power - if you throw enough darts one of them will hit the bullseye.
I'd be surprised if YouGov were still forecasting a Hung Parliament on the eve of the election. In the meantime they've produced a humdinger of a headline for tomorrow's Times.
Meanwhile, THAT Corbyn IRA attack video has now been seen 4.377 million times.....
Perhaps it's actually boosting his popularity!
Those are all paid views. Many if not most will be for 3 seconds or less. 4.3m is a crap number, it could easily be 400,000 people who check Facebook x10 a day.
Also Momentum is encouraging people to click on Tory ads. They need to pay for each view.
The rumour I hear is that the Tories have 'turned the taps on' so to speak on paying to put this ad out. That means they are paying lots of money to put it in every facebook feed, at the start of every youtube clip, and so on, they can possibly buy.
Far from the view count rising being a sign of strength, if this rumour were true then it could be a sign of weakness, in fact a sign of outright panic, at CCHQ.
To me the Tories are ramping up their ad spend in a way that looks panicky, although not yet outright desperate. My personal interpretation of the placement and frequency of attack ads suggests that CCHQ think (or at least worry) that this is going to be close.
YouGov. Cough. Does anyone believe this stuff? Really?
It's true that the young are in the tank for Corbyn (my office is unbelievable, the bizarre strength of Corbynism among posh, privately educated, twentysomething women) but there simply aren't enough of them in the right places, and many don't vote.
Indeed. Hence why even you Corbynistas haven't been this optimistic. This would have to mean the middle aged or even the old Tory vote is hemorrhaging and no one else has noticed.
Me a Corbynista? Er, no.
Sorry, this prediction has scrambled my wits.
Although if this thing is right, hundreds of thousands have turned into Corbynistas and aren't telling anyone else.
Labour gaining seats? Where exactly are they going to gain seats....
Good point. Let's play name the Labour seat gain game.
Okay, Leeds NW, Cambridge.
Err...
Gower, Sheffield Hallam, Morley, Croydon central off the top of my head.
I'm sticking with 4 Labour gains from Tories: Bury North, Cardiff North, Bristol North West, Brighton Kemptown.
Plus Leeds NW, Brighton Pavilion, East Lothian and Edinburgh North from other parties.
Yeah. My point is there are plenty of seats it's not impossible for Labour to gain . People seem to be very inert in accepting the Tories aren't doing as well as they should be. Rather reminiscent of Labours fabled ground game of 2015. We believe what we want to.
Labour gaining seats? Where exactly are they going to gain seats....
Good point. Let's play name the Labour seat gain game.
Okay, Leeds NW, Cambridge.
Err...
Gower, Sheffield Hallam, Morley, Croydon central off the top of my head.
I'm sticking with 4 Labour gains from Tories: Bury North, Cardiff North, Bristol North West, Brighton Kemptown.
Plus Leeds NW, Brighton Pavilion, East Lothian and Edinburgh North from other parties.
Can't see Lucas losing. What's the point ?
Some polls have had the Greens as low as 1%. IF they really do lose three-quarters of their support (and tbf there have been some polls showing them doing better, including ICM today), then it's hard to see that Brighton Pavilion would be immune to that.
I'd love to see a YouGov poll restricted to people who have been on their panel at least 2 years - just in case.....
I have been on their panel for 5 years. I have been polled only once in the last year about four weeks ago. I understand they have 800k on the panel but even so.
YouGov. Cough. Does anyone believe this stuff? Really?
It's true that the young are in the tank for Corbyn (my office is unbelievable, the bizarre strength of Corbynism among posh, privately educated, twentysomething women) but there simply aren't enough of them in the right places, and many don't vote.
Indeed. Hence why even you Corbynistas haven't been this optimistic. This would have to mean the middle aged or even the old Tory vote is hemorrhaging and no one else has noticed.
Me a Corbynista? Er, no.
Sorry, this prediction has scrambled my wits.
Although if this thing is right, hundreds of thousands have turned into Corbynistas and aren't telling anyone else.
Like the Trump supporters who wouldn't even admit it to their spouses?
Seriously though, Labour have had a couple of rough patches in the last few days, as compared to their otherwise faultless campaign (barring Abbott's messes), if that has no impact in a day or so, start to panic.
Seriously though, Labour have had a couple of rough patches in the last few days, as compared to their otherwise faultless campaign (barring Abbott's messes), if that has no impact in a day or so, start to panic.
Comments
He will post again when the election is over; not before.
And that VI intention doesn't point to a hung parliament either, btw.
The poll consensus is a lead of 9% or so, and we know the Labour position is heavily reliant on high turnout from young voters.
It may simply be that voters don't like what they see of May and don't want her manifesto.
Maybe Labour's vote share is really efficient this time around?
Con 42 (-1) Lab 38 (+2) LD 9 (nc) UKIP 4 (nc)
Changes from their Sunday Times poll.
I am posting on mobile and now in my PC in my £1m house paid for purely by myself!
Why is this poll rubbish posted on here!
We are 12% clear on ICM and 12% clear on ComRes! Game over Party like its 1987!!!!
A manifesto outright promising to eat the first born of every family wouldn't see such a switch.
That's assuming gimp mask wearing Nick Timothy doesn't escape from the dungeon he's being kept in.
Yes, they are using models.
They have recruited some cheeky cherubs with MScs in Data Science or Machine Learning with new laptops that their Mum bought them, with R uploaded and wheeling thru the HighPerformanceComputing and MachineLearning views of CRAN. They're doing Hamiltonian Monte Carlo, they're having a whale of a time, the clients are impressed, and that shit looks really good on your CV, trust me. You can point to stuff on screens. Woo.
There is one teeny problem.
Any model is dependent on assumptions about past behavior, and if those behaviors change (eg bigger than expected youth turnout) then the model is fucked. This isn't an abstract problem, it happened with ComRes and their Voter Turnout Model in 2016 EURef. So they've replaced something which isn't dependent (asking people about their Likelihood To Vote on a 1-10 scale) with something that is dependent (a voter turnout model)
On the specific case of YouGov, they're using another model to deduce seats from votes. Which is poling Pelion upon Ossa if you ask me: why in the name of Almighty God are they producing seat estimates at all?!
Pause.
I get a bit intense about this...
But, it wouldn't explain why polls 4-6 weeks ago showed huge Tory leads.
I have to say YouGov is taking a big risk.
Don't forget that the man who came up with it was the son of a second hand horse dealer...
The error could be huge.
https://capx.co/land-value-tax-is-a-great-idea-but-itll-never-happen/
Either that, or Corbyn has benefitted from a Brexit/Trump tidal wave of People Who Don't Do Politics. Almost certainly not the same one as 11 months ago. Mad.
This is the first time it's shown the Conservatives behind the winning post, and up until a few weeks ago showed very solid leads.
An awful lot of Yougov panellists have either changed their minds, or firmed up, very very rapidly.
The first and only indication will be the turnout. If the turnout is 70-75%, anything is possible.
In GE2015 , not only 66% voted but it had the lowest turnout of 18-24 year olds.
We are both on the same side let's enjoy it then
Why the hell not, in a situation where YouGov are saying this outcome is going to happen even though no Labour figure has seriously suggested such an outcome is possible?
Plus Leeds NW, Brighton Pavilion, East Lothian and Edinburgh North from other parties.
My gut says ICM is right though and TMICIPM
I don't believe YouGov for a minute, but I do believe it increases the chance of (as someone put it upthread) a black swan. If Corbyn turns up to the debate and monsters Rudd, while Farron holds firm, the narrative could swing.
https://twitter.com/Aaron4DonValley/status/868832704997076992
but Caroline Flint has the team out times 3, and looks pretty chipper:
https://twitter.com/CarolineFlintMP/status/868929154170986503
If it was publicly tested where are results published?
How many models did they test? If someone tested (say) 10 models chances are that one of them would have predicted the result very well. But that wouldn't necessarily mean the model has good predictive power - if you throw enough darts one of them will hit the bullseye.
Far from the view count rising being a sign of strength, if this rumour were true then it could be a sign of weakness, in fact a sign of outright panic, at CCHQ.
To me the Tories are ramping up their ad spend in a way that looks panicky, although not yet outright desperate. My personal interpretation of the placement and frequency of attack ads suggests that CCHQ think (or at least worry) that this is going to be close.
Although if this thing is right, hundreds of thousands have turned into Corbynistas and aren't telling anyone else.
Mike
* the YouGov Voter Intention poll report
* the YouGov Seat Model
The former has been around for months. The latter went live (I think) last week.
So I suspect its wrong simply because a lot of people aren't going to get out and vote...