'Mr Shakespeare said that the figures could change dramatically before June 8: “The data suggests that there is churn on all fronts, with the Conservatives, Labour and the Liberal Democrats both likely to both lose and gain seats. Based on the model’s current estimates, some seats are likely to change hands along EU referendum dividing lines.” '
Giving himself a get-out clause.
I really don't see the point in YouGov publishing this study.
Not sure I buy the YouGov referendum explanation. Just because it picked the winner in a binary choice referendum where both sides were close to 50% doesn't mean it's any good at predicting the outcome of a multi-party FPTP election. Quite possible that their model is rubbish and it picked Leave to win "by accident"
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?
They had this model. Did they say they actually used it for their public figures ? During the post-election review they realised it was correct all along. Remember, the received wisdom was Remain was going to win anyway, like the Tories now.
Yes, they obviously weren't publicly using it at the time.
The problem is you don't know if they're just working backwards, to make the model fit the outcome.
Yougov's public polling certainly didn't show Leave ahead the whole time.
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.
That does sound very similar to the Ashcroft model's approach:
The Ashcroft Model combines the results of a 40,000-sample survey... with detailed census data. This shows us the likelihood that particular kinds of people – in terms of their age, sex, occupation, level of education, and so on – will have a particular opinion. It uses information from people of the same demographic background but who are from substantially different geographic areas to help calculate these probabilities. Crucially, the model then allows us to extrapolate this information to individual constituencies according to the profile of their population."
Similar maybe but with two very constrasting results! One giving the Tories 400 seats and the other questioning if they will get more than 285!
Similar approach but obviously very different assumptions being fed into the respective models. The phrase 'garbage in, garbage out' comes to mind, which must apply to at least one of the models, and possibly both.
Ashcroft introduced his model with a load of caveats, which might be worth revisiting as they no doubt apply equally to Yougov's model:
Just as traditional polling comes with caveats – that figures are subject to a margin of error, and only represent a snapshot in time – the Ashcroft Model comes with stipulations of its own. To adapt my 2015 mantra, we will be dealing with probabilities, not predictions. The combination of survey and census data will show the probability that certain types of people will behave in certain ways, and the data is presented on that basis. This in turn means that we are applying a national model to specific local areas with their own circumstances. Though we will continue to refine the model as we update the findings each week, we should be prepared for a few anomalies. The seat-by-seat estimates – and, by extension, the implied parliamentary majorities – are therefore not by any means hard and fast predictions.
Where's Jason? He'll calm the PB Tories and anto-Corbynites down.
He might have a novel and untried strategy that involves mentioning the IRA.
Could it be that the Tory strategy of concentrating on Jezza is actually helping his brand? and that the more people see him, the more charmed they are?
If you constantly repeat 'It's me or Jeremy Corbyn', people who don't want you know what there choice should be,
Jezza knows how to work a crowd, and gain the big Mo.
(I hate it when you spot a typo too late to correct it... 'Their choice')
On topic, Corbyn has sucked the air out of the other campaigns. The Tories made it a referendum about him, and it doesn't seem to have worked.
Seemed like a safe gamble given the polls, given the locals, given Copeland. Somehow they've mucked it up. Small majority at best, 30-40.
Perhaps the Stoke result should have been more of a warning to them.
May should do the debate tomorrow because it won't cost her anything.
It's a far easier format than C4 yesterday or BBC1 QT on Friday.
Seven people giving one minute soundbites with minimal debate or interaction.
Slow paced, nobody gets the chance to build any momentum as you don't get to talk for long enough.
Agreed.
Not sure, she's so ponderous and flat footed and waffly she might make things worse. Tories should have sent Ruth Davidson. She's quick thinking and an excellent speaker.
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?
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?
They have models for every occasion.
When one is exposed as a failure - see Kellner and the Referendum - then wheel out another and claim it predicted the right result.
'Mr Shakespeare said that the figures could change dramatically before June 8: “The data suggests that there is churn on all fronts, with the Conservatives, Labour and the Liberal Democrats both likely to both lose and gain seats. Based on the model’s current estimates, some seats are likely to change hands along EU referendum dividing lines.” '
Giving himself a get-out clause.
I really don't see the point in YouGov publishing this study.
There is simply no way on God's green earth that remain-leave polarisation of the vote favours Labour. Leave won over 400 seats, the Tories are well up with leave voters.
Almost no-one is voting on Brexit. Only the swivel-eyed on both camps. That is why Labour were so wise in ignoring it. Brexit election , my arse! That was settled last year.
My forecast is still a Tory majority of around 60-100.
That's assuming gimp mask wearing Nick Timothy doesn't escape from the dungeon he's being kept in.
At this rate, Corbyn will be on course for an overall majority this time next week.
First, I agree with the general feeling that too little seems to have changed on the ground to justify the big swings in the polls. Here in Erdington I'm still far from confident that Labour can hang on, so much of the Ukip vote has gone to the Tories - sorry I mean to the Theresa May Party - though obviously the West Midlands is one of the worst areas for Labour so if there's a 5% swing here the national swing might be 2% or less.
Secondly, statistics are a mug's game. It was mentioned today that ICM's numbers are different because they apportion the refuse-to-answers mainly to the Tories. They could be right, but it should hardly be a matter of guesswork, should it. After all these years, the successes and the fiascos, there should be some sort of consensus about how to weight polling data!
I know Tissue Price is an ornament and a grace to this site, and he will no doubt be a credit to his party if elected. But I can't help but look at that picture and think...
...is he trying to knob that sign? His crotch is very close to it...
Okay guys, lets have a reality check on the propensity of young people to vote remembering, of course, that you need to be registered to vote in order to vote.
So consider the following table of votes cast in the local council elections earlier this month. It's a valuable analysis for several reasons: It's real life data, recently found based on actual votes cast and turnouts Norwich South is divided into wards, which are also County Divisions so, in theory, of approximately equal size. Norwich South is a key marginal Norwich South is a University Seat where there are defined concentrations of student voters most notably in UNIVERSITY ward [funny that] and NELSON wards.
You'll see from the table that the average electorate of all wards in Norwich South is 7935. But the total electorate in the UNIVERSITY ward is only 5419, that is 68pc of average - a reduction of over 2500 electors. Nelson ward, immediately to the east and home to the 2nd & 3rd year student houses is also lower than average in electors at 85% of the constituency average.
Division TurnoutPc Electorate BOWTHORPE 28.95% 8,114 102% COSTESSEY 31.88% 11,205 141% EATON 53.80% 7,323 92% LAKENHAM 34.81% 7,133 90% MANCROFT 35.91% 8,109 102% NELSON 54.22% 6,771 85% THORPE HAMLET 37.62% 8,923 112% TOWN CLOSE 46.87% 8,332 105% UNIVERSITY 36.81% 5,420 68% WENSUM 33.40% 8,030 101% Average 7,936
Now, I accept that in the last few weeks more students might have registered to vote... but we can all accept that it's unlikely to have made-up such a backlog in such a short time. Some students may be registered to vote at home rather than at Uni - but that then disperses the concentration effect and makes University-Town-Ness less potent... that is, if they have applied for a postal vote. Oh yes, and, given the tendancy for Universities to go for Semesters means that term ended on 19th May for students that aren't sitting exams in the 'assessment period' falling between 22nd May and June 16th.
What I conclude is that young people can talk all they like about political engagement and protest for Corbyn. They can say they're 'certain to vote'. But if they're not registered, they just don't count. And if they've all gone home.
You'd think that University Students would be more politically engaged than their peers... which would imply that the generality of electoral registration is even lower amongst the youngsters.
If the ONS/ElectionData etc mumbo jumbo hasn't taken this into account then it's just another voodoo poll.
'Mr Shakespeare said that the figures could change dramatically before June 8: “The data suggests that there is churn on all fronts, with the Conservatives, Labour and the Liberal Democrats both likely to both lose and gain seats. Based on the model’s current estimates, some seats are likely to change hands along EU referendum dividing lines.” '
Giving himself a get-out clause.
I really don't see the point in YouGov publishing this study.
Maybe so. I'm not taking the risk - as I say, my vote won't affect the outcome, and I hate to vote negatively, but there it is. A shame, as I want the LDs to do well - unusually they are not working my area hard, either because they were stunned that it went Tory in the locals (despite the local Tories not leafleting half the area), or because they are working target seats.
May should do the debate tomorrow because it won't cost her anything.
It's a far easier format than C4 yesterday or BBC1 QT on Friday.
Seven people giving one minute soundbites with minimal debate or interaction.
Slow paced, nobody gets the chance to build any momentum as you don't get to talk for long enough.
Agreed.
Not sure, she's so ponderous and flat footed and waffly she might make things worse. Tories should have sent Ruth Davidson. She's quick thinking and an excellent speaker.
And not a candidate! It is one thing to send a Cabinet Minister rather than the leader, but that may be stretching credulity.
There is simply no way on God's green earth that remain-leave polarisation of the vote favours Labour. Leave won over 400 seats, the Tories are well up with leave voters.
Almost no-one is voting on Brexit. Only the swivel-eyed on both camps. That is why Labour were so wise in ignoring it. Brexit election , my arse! That was settled last year.
'Mr Shakespeare said that the figures could change dramatically before June 8: “The data suggests that there is churn on all fronts, with the Conservatives, Labour and the Liberal Democrats both likely to both lose and gain seats. Based on the model’s current estimates, some seats are likely to change hands along EU referendum dividing lines.” '
Giving himself a get-out clause.
I really don't see the point in YouGov publishing this study.
Maybe so. I'm not taking the risk - as I say, my vote won't affect the outcome, and I hate to vote negatively, but there it is. A shame, as I want the LDs to do well - unusually they are not working my area hard, either because they were stunned that it went Tory in the locals (despite the local Tories not leafleting half the area), or because they are working target seats.
I'm not saying you shouldn't vote Tory. I'm saying don't get worked up over this YouGov projection
May should do the debate tomorrow because it won't cost her anything.
It's a far easier format than C4 yesterday or BBC1 QT on Friday.
Seven people giving one minute soundbites with minimal debate or interaction.
Slow paced, nobody gets the chance to build any momentum as you don't get to talk for long enough.
Agreed.
Not sure, she's so ponderous and flat footed and waffly she might make things worse. Tories should have sent Ruth Davidson. She's quick thinking and an excellent speaker.
Surely though with Corbyn gaining in the polls she has to go for it. What is the alternative, giving Corbyn and the also rans prime time publicity? Ok this is only one poll but the trend is clear, and if it continues a Conervative majority is no longer certain.
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?
They have models for every occasion.
When one is exposed as a failure - see Kellner and the Referendum - then wheel out another and claim it predicted the right result.
Kellner during the Referendum wasn't defending a model he was defending a mode: specifically, phone polls vs online polls.
'Mr Shakespeare said that the figures could change dramatically before June 8: “The data suggests that there is churn on all fronts, with the Conservatives, Labour and the Liberal Democrats both likely to both lose and gain seats. Based on the model’s current estimates, some seats are likely to change hands along EU referendum dividing lines.” '
Giving himself a get-out clause.
I really don't see the point in YouGov publishing this study.
Maybe so. I'm not taking the risk - as I say, my vote won't affect the outcome, and I hate to vote negatively, but there it is. A shame, as I want the LDs to do well - unusually they are not working my area hard, either because they were stunned that it went Tory in the locals (despite the local Tories not leafleting half the area), or because they are working target seats.
I'm not saying you shouldn't vote Tory. I'm saying don't get worked up over this YouGov projection
Her [Mrs May] support appears to have plunged after the poor reception of the party manifesto, including plans to make more elderly voters pay for home care.
Plunged from 46% to 44%?
Jesus, these journalists are idiots. (I assume that's a quote from the Times. )
There is simply no way on God's green earth that remain-leave polarisation of the vote favours Labour. Leave won over 400 seats, the Tories are well up with leave voters.
Almost no-one is voting on Brexit. Only the swivel-eyed on both camps. That is why Labour were so wise in ignoring it. Brexit election , my arse! That was settled last year.
Perhaps someone should have mentioned that to the PM.
May should do the debate tomorrow because it won't cost her anything.
It's a far easier format than C4 yesterday or BBC1 QT on Friday.
Seven people giving one minute soundbites with minimal debate or interaction.
Slow paced, nobody gets the chance to build any momentum as you don't get to talk for long enough.
Agreed.
Not sure, she's so ponderous and flat footed and waffly she might make things worse. Tories should have sent Ruth Davidson. She's quick thinking and an excellent speaker.
Surely though with Corbyn gaining in the polls she has to go for it. What is the alternative, giving Corbyn and the also rans prime time publicity? Ok this is only one poll but the trend is clear, and if it continues a Conervative majority is no longer certain.
Where's Jason? He'll calm the PB Tories and anto-Corbynites down.
He might have a novel and untried strategy that involves mentioning the IRA.
Could it be that the Tory strategy of concentrating on Jezza is actually helping his brand? and that the more people see him, the more charmed they are?
If you constantly repeat 'It's me or Jeremy Corbyn', people who don't want you know what there choice should be,
Jezza knows how to work a crowd, and gain the big Mo.
(I hate it when you spot a typo too late to correct it... 'Their choice')
On topic, Corbyn has sucked the air out of the other campaigns. The Tories made it a referendum about him, and it doesn't seem to have worked.
Seemed like a safe gamble given the polls, given the locals, given Copeland. Somehow they've mucked it up. Small majority at best, 30-40.
Perhaps the Stoke result should have been more of a warning to them.
Stoke Central for Labour is 1/2, and Copeland 10/3.
I see Gill Troughton is running a broader based anti austerity campaign, particularly on education.
The NUT video is now on 3.3 million views BTW, and Facebook is a mums media.
Her [Mrs May] support appears to have plunged after the poor reception of the party manifesto, including plans to make more elderly voters pay for home care.
Plunged from 46% to 44%?
Jesus, these journalists are idiots. (I assume that's a quote from the Times. )
'Why are YouGov trying to put themselves out of business?'
So their competitors are laughing at them then. Still can't believe people are in meltdown over a single projection.
Days ago this lot had the Tory lead up to 7%.
Hold on ! The Tory lead is still 4%. The seats projection is not coming from UNS but their constituency model.
Yes.
Their constituency model might be quite good - even if the VI is overstating Lab.
The lab vote could potentially be nice and efficient.
Tories piling up votes where they either a) don't need them (blue heartlands) and/or b) where they're not strong enough to challenge (lab heartlands)
I want to play with their constituency model. Hope it goes online.
I just don't get why the Tories would be inefficient. They are bringing in a social care charge which will disproportionately hit larger homes in the south so presumably will hold back their vote in seats they already hold
Does Iain Dale still contribute here ? I remember when he was a regular when he was a Tory candidate in East Anglia somewhere- Norwich perhaps ?
I can't recall reading any posts of his on here.
Are you Paul Mason? PaulM = Paul Mason?
No. I used to post here regularly years ago, and then found that when the vanilla forums came in I couldn't get them to work on my macbook, so lapsed for a few years.
Retaining 50 seats would be a stonkingly good figure for them really.
That is also wrong as even on their own figures on Sunday Yougov had SNP 40%, Tories 30% in Scotland which would see 10 SNP seats go to the Tories leaving the SNP on 46 seats, so it looks like Yougov haven't even bothered to include their own Scottish figures properly in their overall calculation
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?
They have models for every occasion.
When one is exposed as a failure - see Kellner and the Referendum - then wheel out another and claim it predicted the right result.
Kellner during the Referendum wasn't defending a model he was defending a mode: specifically, phone polls vs online polls.
'Why are YouGov trying to put themselves out of business?'
So their competitors are laughing at them then. Still can't believe people are in meltdown over a single projection.
Days ago this lot had the Tory lead up to 7%.
Hold on ! The Tory lead is still 4%. The seats projection is not coming from UNS but their constituency model.
Yes.
Their constituency model might be quite good - even if the VI is overstating Lab.
The lab vote could potentially be nice and efficient.
Tories piling up votes where they either a) don't need them (blue heartlands) and/or b) where they're not strong enough to challenge (lab heartlands)
I want to play with their constituency model. Hope it goes online.
If that were the case, surely private constituency polling would show this and we'd hear murmurings?
In fairness, if the private constituency polling is being done by the tories, TM would know the state of play and nobody can accuse her of not saying it straight with her warnings about losing the majority. We all read it as spin, but maybe she's seeing how the votes are falling and the reality is petrifying?
What's happening in scotland (decent con vote, but terrible conversion into seats) might be happening in labour seats in E&W too.
And the anti-corbyn stuff is effective, but it's piling up massive majorities in the shires.
The new national Theresa May party could be bloody hopeless under FPTP.
'Why are YouGov trying to put themselves out of business?'
So their competitors are laughing at them then. Still can't believe people are in meltdown over a single projection.
Days ago this lot had the Tory lead up to 7%.
Hold on ! The Tory lead is still 4%. The seats projection is not coming from UNS but their constituency model.
Yes.
Their constituency model might be quite good - even if the VI is overstating Lab.
The lab vote could potentially be nice and efficient.
Tories piling up votes where they either a) don't need them (blue heartlands) and/or b) where they're not strong enough to challenge (lab heartlands)
I want to play with their constituency model. Hope it goes online.
I just don't get why the Tories would be inefficient. They are bringing in a social care charge which will disproportionately hit larger homes in the south so presumably will hold back their vote in seats they already hold
Even after the social care clusterfuck, the polls are still showing that the Tories' gains on 2015 are still being drawn heavily from the 65+ age bracket - who are almost as inefficiently-distributed under FPTP as students are.
'Why are YouGov trying to put themselves out of business?'
So their competitors are laughing at them then. Still can't believe people are in meltdown over a single projection.
Days ago this lot had the Tory lead up to 7%.
Hold on ! The Tory lead is still 4%. The seats projection is not coming from UNS but their constituency model.
Yes.
Their constituency model might be quite good - even if the VI is overstating Lab.
The lab vote could potentially be nice and efficient.
Tories piling up votes where they either a) don't need them (blue heartlands) and/or b) where they're not strong enough to challenge (lab heartlands)
I want to play with their constituency model. Hope it goes online.
I just don't get why the Tories would be inefficient. They are bringing in a social care charge which will disproportionately hit larger homes in the south so presumably will hold back their vote in seats they already hold
Typically social care costs would be in the tens of thousands, not hundreds. It is a bigger hit on a £180 000 house percentagewise than a million pound one.
'Why are YouGov trying to put themselves out of business?'
So their competitors are laughing at them then. Still can't believe people are in meltdown over a single projection.
Days ago this lot had the Tory lead up to 7%.
Hold on ! The Tory lead is still 4%. The seats projection is not coming from UNS but their constituency model.
Yes.
Their constituency model might be quite good - even if the VI is overstating Lab.
The lab vote could potentially be nice and efficient.
Tories piling up votes where they either a) don't need them (blue heartlands) and/or b) where they're not strong enough to challenge (lab heartlands)
I want to play with their constituency model. Hope it goes online.
If that were the case, surely private constituency polling would show this and we'd hear murmurings?
In fairness, if the private constituency polling is being done by the tories, TM would know the state of play and nobody can accuse her of not saying it straight with her warnings about losing the majority. We all read it as spin, but maybe she's seeing how the votes are falling and the reality is petrifying?
What's happening in scotland (decent con vote, but terrible conversion into seats) might be happening in labour seats in E&W too.
And the anti-corbyn stuff is mostly piling up massive majorities in the shires.
The new national Theresa May party could be bloody hopeless under FPTP.
Interesting times.
Tories aren't doing constituency polling, that would break election expenses.
They will undoubtedly being doing regional poling.
'Why are YouGov trying to put themselves out of business?'
So their competitors are laughing at them then. Still can't believe people are in meltdown over a single projection.
Days ago this lot had the Tory lead up to 7%.
Hold on ! The Tory lead is still 4%. The seats projection is not coming from UNS but their constituency model.
Yes.
Their constituency model might be quite good - even if the VI is overstating Lab.
The lab vote could potentially be nice and efficient.
Tories piling up votes where they either a) don't need them (blue heartlands) and/or b) where they're not strong enough to challenge (lab heartlands)
I want to play with their constituency model. Hope it goes online.
I just don't get why the Tories would be inefficient. They are bringing in a social care charge which will disproportionately hit larger homes in the south so presumably will hold back their vote in seats they already hold
Even after the social care clusterfuck, the polls are still showing that the Tories' gains on 2015 are still being drawn heavily from the 65+ age bracket - who are almost as inefficiently-distributed under FPTP as students are.
Pensioners are distributed throughout the country - its young people who are more concentrated.
May should do the debate tomorrow because it won't cost her anything.
It's a far easier format than C4 yesterday or BBC1 QT on Friday.
Seven people giving one minute soundbites with minimal debate or interaction.
Slow paced, nobody gets the chance to build any momentum as you don't get to talk for long enough.
Agreed.
Not sure, she's so ponderous and flat footed and waffly she might make things worse. Tories should have sent Ruth Davidson. She's quick thinking and an excellent speaker.
Surely though with Corbyn gaining in the polls she has to go for it. What is the alternative, giving Corbyn and the also rans prime time publicity? Ok this is only one poll but the trend is clear, and if it continues a Conervative majority is no longer certain.
'Why are YouGov trying to put themselves out of business?'
So their competitors are laughing at them then. Still can't believe people are in meltdown over a single projection.
Days ago this lot had the Tory lead up to 7%.
Hold on ! The Tory lead is still 4%. The seats projection is not coming from UNS but their constituency model.
Yes.
Their constituency model might be quite good - even if the VI is overstating Lab.
The lab vote could potentially be nice and efficient.
Tories piling up votes where they either a) don't need them (blue heartlands) and/or b) where they're not strong enough to challenge (lab heartlands)
I want to play with their constituency model. Hope it goes online.
If that were the case, surely private constituency polling would show this and we'd hear murmurings?
In fairness, if the private constituency polling is being done by the tories, TM would know the state of play and nobody can accuse her of not saying it straight with her warnings about losing the majority. We all read it as spin, but maybe she's seeing how the votes are falling and the reality is petrifying?
What's happening in scotland (decent con vote, but terrible conversion into seats) might be happening in labour seats in E&W too.
And the anti-corbyn stuff is mostly piling up massive majorities in the shires.
The new national Theresa May party could be bloody hopeless under FPTP.
Interesting times.
Tories aren't doing constituency polling, that would break election expenses.
'Why are YouGov trying to put themselves out of business?'
So their competitors are laughing at them then. Still can't believe people are in meltdown over a single projection.
Days ago this lot had the Tory lead up to 7%.
Hold on ! The Tory lead is still 4%. The seats projection is not coming from UNS but their constituency model.
Yes.
Their constituency model might be quite good - even if the VI is overstating Lab.
The lab vote could potentially be nice and efficient.
Tories piling up votes where they either a) don't need them (blue heartlands) and/or b) where they're not strong enough to challenge (lab heartlands)
I want to play with their constituency model. Hope it goes online.
I just don't get why the Tories would be inefficient. They are bringing in a social care charge which will disproportionately hit larger homes in the south so presumably will hold back their vote in seats they already hold
Typically social care costs would be in the tens of thousands, not hundreds. It is a bigger hit on a £180 000 house percentagewise than a million pound one.
Also, if you have a million quid, you are probably able to give 50k to your kids / grandkids easily to help them get on the property ladder. So you feel safe seeing them get on, the way you did.
If on the other hand 180k is all you've got, you're unable to help your kids, and are far more worried about them never owning property and having a lower standard of living than you ever did.
As I've said on here many times, the dementia tax looks disproportionately bad to lower middle class voters in marginal seats and that is why it was a terrible, terrible policy.
May should do the debate tomorrow because it won't cost her anything.
It's a far easier format than C4 yesterday or BBC1 QT on Friday.
Seven people giving one minute soundbites with minimal debate or interaction.
Slow paced, nobody gets the chance to build any momentum as you don't get to talk for long enough.
Agreed.
Not sure, she's so ponderous and flat footed and waffly she might make things worse. Tories should have sent Ruth Davidson. She's quick thinking and an excellent speaker.
Surely though with Corbyn gaining in the polls she has to go for it. What is the alternative, giving Corbyn and the also rans prime time publicity? Ok this is only one poll but the trend is clear, and if it continues a Conervative majority is no longer certain.
Do we actually know for certain that Corbyn is doing this show?
'Why are YouGov trying to put themselves out of business?'
So their competitors are laughing at them then. Still can't believe people are in meltdown over a single projection.
Days ago this lot had the Tory lead up to 7%.
Hold on ! The Tory lead is still 4%. The seats projection is not coming from UNS but their constituency model.
Yes.
Their constituency model might be quite good - even if the VI is overstating Lab.
The lab vote could potentially be nice and efficient.
Tories piling up votes where they either a) don't need them (blue heartlands) and/or b) where they're not strong enough to challenge (lab heartlands)
I want to play with their constituency model. Hope it goes online.
I just don't get why the Tories would be inefficient. They are bringing in a social care charge which will disproportionately hit larger homes in the south so presumably will hold back their vote in seats they already hold
Even after the social care clusterfuck, the polls are still showing that the Tories' gains on 2015 are still being drawn heavily from the 65+ age bracket - who are almost as inefficiently-distributed under FPTP as students are.
Pensioners are distributed throughout the country - its young people who are more concentrated.
The seats with the highest % of pensioners tend to already be held by the Tories -- very few marginal seats are pensioner hotspots (though an interesting exception to that is the Wirral seats, which have been trending Labour despite the demographics suggesting they should be doing the reverse).
'Why are YouGov trying to put themselves out of business?'
So their competitors are laughing at them then. Still can't believe people are in meltdown over a single projection.
Days ago this lot had the Tory lead up to 7%.
Hold on ! The Tory lead is still 4%. The seats projection is not coming from UNS but their constituency model.
Yes.
Their constituency model might be quite good - even if the VI is overstating Lab.
The lab vote could potentially be nice and efficient.
Tories piling up votes where they either a) don't need them (blue heartlands) and/or b) where they're not strong enough to challenge (lab heartlands)
I want to play with their constituency model. Hope it goes online.
If that were the case, surely private constituency polling would show this and we'd hear murmurings?
In fairness, if the private constituency polling is being done by the tories, TM would know the state of play and nobody can accuse her of not saying it straight with her warnings about losing the majority. We all read it as spin, but maybe she's seeing how the votes are falling and the reality is petrifying?
What's happening in scotland (decent con vote, but terrible conversion into seats) might be happening in labour seats in E&W too.
And the anti-corbyn stuff is mostly piling up massive majorities in the shires.
The new national Theresa May party could be bloody hopeless under FPTP.
Interesting times.
Tories aren't doing constituency polling, that would break election expenses.
May should do the debate tomorrow because it won't cost her anything.
It's a far easier format than C4 yesterday or BBC1 QT on Friday.
Seven people giving one minute soundbites with minimal debate or interaction.
Slow paced, nobody gets the chance to build any momentum as you don't get to talk for long enough.
Agreed.
Not sure, she's so ponderous and flat footed and waffly she might make things worse. Tories should have sent Ruth Davidson. She's quick thinking and an excellent speaker.
Surely though with Corbyn gaining in the polls she has to go for it. What is the alternative, giving Corbyn and the also rans prime time publicity? Ok this is only one poll but the trend is clear, and if it continues a Conervative majority is no longer certain.
I see what you're thinking but if she performed anything like as woodenly as she did on the channel 4 thing she'd struggle to get a word in edgeways and flounder if she did. It's not her strength. Her best strategy is to keep campaigning and hope the anti Corbyn vote saves her bacon. It's disappointing to be dragged down to that level but she has nothing else to offer.
Shame really it was nice being able to vote Conservative for positive reasons at the last 2 elections.
Retaining 50 seats would be a stonkingly good figure for them really.
That is also wrong as even on their own figures on Sunday Yougov had SNP 40%, Tories 30% in Scotland which would see 10 SNP seats go to the Tories leaving the SNP on 46 seats, so it looks like Yougov haven't even bothered to include their own Scottish figures properly in their overall calculation
A lot depends on whether the Scottish Tactical vote is anti Con or anti SNP. The Tories assume the latter, but could it be the former? Left wing unionists may not be that bothered voting against SNP now that Indyref2 has been kicked into touch by Nicola.
There is simply no way on God's green earth that remain-leave polarisation of the vote favours Labour. Leave won over 400 seats, the Tories are well up with leave voters.
Almost no-one is voting on Brexit. Only the swivel-eyed on both camps. That is why Labour were so wise in ignoring it. Brexit election , my arse! That was settled last year.
The seats with the highest % of pensioners tend to already be held by the Tories -- very few marginal seats are pensioner hotspots (though an interesting exception to that is the Wirral seats, which have been trending Labour despite the demographics suggesting they should be doing the reverse).
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.
That does sound very similar to the Ashcroft model's approach:
The Ashcroft Model combines the results of a 40,000-sample survey... with detailed census data. This shows us the likelihood that particular kinds of people – in terms of their age, sex, occupation, level of education, and so on – will have a particular opinion. It uses information from people of the same demographic background but who are from substantially different geographic areas to help calculate these probabilities. Crucially, the model then allows us to extrapolate this information to individual constituencies according to the profile of their population."
Similar maybe but with two very constrasting results! One giving the Tories 400 seats and the other questioning if they will get more than 285!
You say that but then in Ashcroft's explanation he points out a high turnout like the EU ref would drastically reduce the majority;
There is simply no way on God's green earth that remain-leave polarisation of the vote favours Labour. Leave won over 400 seats, the Tories are well up with leave voters.
Almost no-one is voting on Brexit. Only the swivel-eyed on both camps. That is why Labour were so wise in ignoring it. Brexit election , my arse! That was settled last year.
Perhaps someone should have mentioned that to the PM.
It was quite clear when she called the election that they believed quite rightly that success was predicated on it being The Brexit Election. Hence the way it was launched, the slogans, the focus on the leader and so on. They believed/assumed that Brexit was such a big deal to everyone that maintaining that focus would be straightforward. How wrong they were. Having said that, the Brexit lie will still lead to a Tory victory, on the basis of 50% of Ukip voters switching to the Tories, but I think the party will soon rue the cost of taking Ukip's clothes.
'Why are YouGov trying to put themselves out of business?'
So their competitors are laughing at them then. Still can't believe people are in meltdown over a single projection.
Days ago this lot had the Tory lead up to 7%.
Hold on ! The Tory lead is still 4%. The seats projection is not coming from UNS but their constituency model.
Yes.
Their constituency model might be quite good - even if the VI is overstating Lab.
The lab vote could potentially be nice and efficient.
Tories piling up votes where they either a) don't need them (blue heartlands) and/or b) where they're not strong enough to challenge (lab heartlands)
I want to play with their constituency model. Hope it goes online.
I just don't get why the Tories would be inefficient. They are bringing in a social care charge which will disproportionately hit larger homes in the south so presumably will hold back their vote in seats they already hold
Typically social care costs would be in the tens of thousands, not hundreds. It is a bigger hit on a £180 000 house percentagewise than a million pound one.
Not sure about that. Mrs BJs home care is set to cost £20k per year. She is 57 so probably in excess of £400k That is typical of spinal injury costs. Dementia tax is not only about dementia.
There is simply no way on God's green earth that remain-leave polarisation of the vote favours Labour. Leave won over 400 seats, the Tories are well up with leave voters.
Almost no-one is voting on Brexit. Only the swivel-eyed on both camps. That is why Labour were so wise in ignoring it. Brexit election , my arse! That was settled last year.
No. It has nothing to do with Brexit - but LOOK AT THE REMAIN-LEAVE split in any of the subsamples.
The Tories are clearly doing better with leave and Labour with remain voters. FFSake
A plausible explanation is that the CDE voters that went Leave as an anti status quo option are now going Corbyn Labour as an anti status quo option. Plausible doesn't mean right, however.
'Why are YouGov trying to put themselves out of business?'
So their competitors are laughing at them then. Still can't believe people are in meltdown over a single projection.
Days ago this lot had the Tory lead up to 7%.
Hold on ! The Tory lead is still 4%. The seats projection is not coming from UNS but their constituency model.
Yes.
Their constituency model might be quite good - even if the VI is overstating Lab.
The lab vote could potentially be nice and efficient.
Tories piling up votes where they either a) don't need them (blue heartlands) and/or b) where they're not strong enough to challenge (lab heartlands)
I want to play with their constituency model. Hope it goes online.
I just don't get why the Tories would be inefficient. They are bringing in a social care charge which will disproportionately hit larger homes in the south so presumably will hold back their vote in seats they already hold
Typically social care costs would be in the tens of thousands, not hundreds. It is a bigger hit on a £180 000 house percentagewise than a million pound one.
Also, if you have a million quid, you are probably able to give 50k to your kids / grandkids easily to help them get on the property ladder. So you feel safe seeing them get on, the way you did.
If on the other hand 180k is all you've got, you're unable to help your kids, and are far more worried about them never owning property and having a lower standard of living than you ever did.
As I've said on here many times, the dementia tax looks disproportionately bad to lower middle class voters in marginal seats and that is why it was a terrible, terrible policy.
Weren't you talking about houses in northern marginal constituencies costing £450k last week ?
Now you casually talk about £180k as if its a pittance.
(Irrelevant anecdote and name dropping alert....) I have just got back from a dinner at Balliol. One of the more amusing conversations was how long it would be before Dave Cameron was recognised as the worst Prime Minister since Anthony Eden, and probably worse even than him.
May should do the debate tomorrow because it won't cost her anything.
It's a far easier format than C4 yesterday or BBC1 QT on Friday.
Seven people giving one minute soundbites with minimal debate or interaction.
Slow paced, nobody gets the chance to build any momentum as you don't get to talk for long enough.
Agreed.
Not sure, she's so ponderous and flat footed and waffly she might make things worse. Tories should have sent Ruth Davidson. She's quick thinking and an excellent speaker.
Surely though with Corbyn gaining in the polls she has to go for it. What is the alternative, giving Corbyn and the also rans prime time publicity? Ok this is only one poll but the trend is clear, and if it continues a Conervative majority is no longer certain.
Do we actually know for certain that Corbyn is doing this show?
I don't think he has said for certain but with polls like this he would be stupid not to. If it goes well then he has an evening of good news stories.
'Why are YouGov trying to put themselves out of business?'
So their competitors are laughing at them then. Still can't believe people are in meltdown over a single projection.
Days ago this lot had the Tory lead up to 7%.
Hold on ! The Tory lead is still 4%. The seats projection is not coming from UNS but their constituency model.
Yes.
Their constituency model might be quite good - even if the VI is overstating Lab.
The lab vote could potentially be nice and efficient.
Tories piling up votes where they either a) don't need them (blue heartlands) and/or b) where they're not strong enough to challenge (lab heartlands)
I want to play with their constituency model. Hope it goes online.
I just don't get why the Tories would be inefficient. They are bringing in a social care charge which will disproportionately hit larger homes in the south so presumably will hold back their vote in seats they already hold
Typically social care costs would be in the tens of thousands, not hundreds. It is a bigger hit on a £180 000 house percentagewise than a million pound one.
Also, if you have a million quid, you are probably able to give 50k to your kids / grandkids easily to help them get on the property ladder. So you feel safe seeing them get on, the way you did.
If on the other hand 180k is all you've got, you're unable to help your kids, and are far more worried about them never owning property and having a lower standard of living than you ever did.
As I've said on here many times, the dementia tax looks disproportionately bad to lower middle class voters in marginal seats and that is why it was a terrible, terrible policy.
The demenatia 'tax' was actually most unpopular in London in the polling
'Why are YouGov trying to put themselves out of business?'
So their competitors are laughing at them then. Still can't believe people are in meltdown over a single projection.
Days ago this lot had the Tory lead up to 7%.
Hold on ! The Tory lead is still 4%. The seats projection is not coming from UNS but their constituency model.
Yes.
Their constituency model might be quite good - even if the VI is overstating Lab.
The lab vote could potentially be nice and efficient.
Tories piling up votes where they either a) don't need them (blue heartlands) and/or b) where they're not strong enough to challenge (lab heartlands)
I want to play with their constituency model. Hope it goes online.
If that were the case, surely private constituency polling would show this and we'd hear murmurings?
In fairness, if the private constituency polling is being done by the tories, TM would know the state of play and nobody can accuse her of not saying it straight with her warnings about losing the majority. We all read it as spin, but maybe she's seeing how the votes are falling and the reality is petrifying?
What's happening in scotland (decent con vote, but terrible conversion into seats) might be happening in labour seats in E&W too.
And the anti-corbyn stuff is mostly piling up massive majorities in the shires.
The new national Theresa May party could be bloody hopeless under FPTP.
Interesting times.
Tories aren't doing constituency polling, that would break election expenses.
They will undoubtedly being doing regional poling.
Can we assume you've not received any panicky requests to help out in Pudsey again ?
There is simply no way on God's green earth that remain-leave polarisation of the vote favours Labour. Leave won over 400 seats, the Tories are well up with leave voters.
Almost no-one is voting on Brexit. Only the swivel-eyed on both camps. That is why Labour were so wise in ignoring it. Brexit election , my arse! That was settled last year.
No. It has nothing to do with Brexit - but LOOK AT THE REMAIN-LEAVE split in any of the subsamples.
The Tories are clearly doing better with leave and Labour with remain voters. FFSake
I didn't say that. I said few people are voting on Brexit.
Oh I know. But there is very clear evidence that remain and leave voters are swinging in wildly different directions. Its not really about the EU ref vote, but what it shows behind the constituency - it explains very well why Ealing Central and Acton may well look gloomy whereas Torbay is looking like an increased Tory Maj. Also see Bolsover/Carshalton (LD Loss probable)/Twickers (LD Gain being talked up)
'Why are YouGov trying to put themselves out of business?'
So their competitors are laughing at them then. Still can't believe people are in meltdown over a single projection.
Days ago this lot had the Tory lead up to 7%.
Hold on ! The Tory lead is still 4%. The seats projection is not coming from UNS but their constituency model.
Yes.
Their constituency model might be quite good - even if the VI is overstating Lab.
The lab vote could potentially be nice and efficient.
Tories piling up votes where they either a) don't need them (blue heartlands) and/or b) where they're not strong enough to challenge (lab heartlands)
I want to play with their constituency model. Hope it goes online.
If that were the case, surely private constituency polling would show this and we'd hear murmurings?
In fairness, if the private constituency polling is being done by the tories, TM would know the state of play and nobody can accuse her of not saying it straight with her warnings about losing the majority. We all read it as spin, but maybe she's seeing how the votes are falling and the reality is petrifying?
What's happening in scotland (decent con vote, but terrible conversion into seats) might be happening in labour seats in E&W too.
And the anti-corbyn stuff is mostly piling up massive majorities in the shires.
The new national Theresa May party could be bloody hopeless under FPTP.
Interesting times.
Tories aren't doing constituency polling, that would break election expenses.
They will undoubtedly being doing regional poling.
Can we assume you've not received any panicky requests to help out in Pudsey again ?
I did get an invite, but it wasn't a panicky, more, we haven't seen you in ages, and we miss you.
'Why are YouGov trying to put themselves out of business?'
So their competitors are laughing at them then. Still can't believe people are in meltdown over a single projection.
Days ago this lot had the Tory lead up to 7%.
Hold on ! The Tory lead is still 4%. The seats projection is not coming from UNS but their constituency model.
Yes.
Their constituency model might be quite good - even if the VI is overstating Lab.
The lab vote could potentially be nice and efficient.
Tories piling up votes where they either a) don't need them (blue heartlands) and/or b) where they're not strong enough to challenge (lab heartlands)
I want to play with their constituency model. Hope it goes online.
I just don't get why the Tories would be inefficient. They are bringing in a social care charge which will disproportionately hit larger homes in the south so presumably will hold back their vote in seats they already hold
Typically social care costs would be in the tens of thousands, not hundreds. It is a bigger hit on a £180 000 house percentagewise than a million pound one.
Not sure about that. Mrs BJs home care is set to cost £20k per year. She is 57 so probably in excess of £400k That is typical of spinal injury costs. Dementia tax is not only about dementia.
There certainly are exceptions. Most people aren't doing sums though, they don't like the vibe.
Retaining 50 seats would be a stonkingly good figure for them really.
That is also wrong as even on their own figures on Sunday Yougov had SNP 40%, Tories 30% in Scotland which would see 10 SNP seats go to the Tories leaving the SNP on 46 seats, so it looks like Yougov haven't even bothered to include their own Scottish figures properly in their overall calculation
A lot depends on whether the Scottish Tactical vote is anti Con or anti SNP. The Tories assume the latter, but could it be the former? Left wing unionists may not be that bothered voting against SNP now that Indyref2 has been kicked into touch by Nicola.
Indyref2 has certainly not been 'kicked into touch' by Sturgeon she has confirmed today they will push for it at the end of the Brexit process and there will be clear anti SNP tactical voting in rural Scotland and Edinburgh which is where all the Tory target seats are, anti Tory tactical voting will be in the central belt and Glasgow
I don't really understand about this YouGov model. There seem to be two issues: (1) YouGov is showing a much smaller Tory lead than ICM in projected national vote share, and (2) YouGov's constituency model is showing the Tories doing worse than uniform swing would indicate.
But which aspect of this is meant to have been validated by the referendum? Surely it can't be the second, because the geographical distribution of the votes doesn't matter in a referendum. So are they saying that they've switched to a different turnout model for the national vote share, which was validated by the referendum?
(Irrelevant anecdote and name dropping alert....) I have just got back from a dinner at Balliol. One of the more amusing conversations was how long it would be before Dave Cameron was recognised as the worst Prime Minister since Anthony Eden, and probably worse even than him.
May should do the debate tomorrow because it won't cost her anything.
It's a far easier format than C4 yesterday or BBC1 QT on Friday.
Seven people giving one minute soundbites with minimal debate or interaction.
Slow paced, nobody gets the chance to build any momentum as you don't get to talk for long enough.
Agreed.
Not sure, she's so ponderous and flat footed and waffly she might make things worse. Tories should have sent Ruth Davidson. She's quick thinking and an excellent speaker.
Surely though with Corbyn gaining in the polls she has to go for it. What is the alternative, giving Corbyn and the also rans prime time publicity? Ok this is only one poll but the trend is clear, and if it continues a Conervative majority is no longer certain.
I see what you're thinking but if she performed anything like as woodenly as she did on the channel 4 thing she'd struggle to get a word in edgeways and flounder if she did. It's not her strength. Her best strategy is to keep campaigning and hope the anti Corbyn vote saves her bacon. It's disappointing to be dragged down to that level but she has nothing else to offer.
Shame really it was nice being able to vote Conservative for positive reasons at the last 2 elections.
This. We can bang on about everything but to my mind it boils down to this: the Tory manifesto got precisely no one excited. There was nothing in it to motivate the Tory vote and bring others over. You can't really polish a turd and Tezza doing more TV ain't going to make a difference at this stage. She just needs to sail on above the fray hoping she looks PM-like and that's what voters will be thinking about when their pencil's hovering over the ballot, and put her trust in Crosby that hammering away on Corbyn's past will have some impact. The course is set: suddenly turning up at TV shows will declare PANIC in capital letters and another u-turn. One more u-turn from May would be fatal.
Retaining 50 seats would be a stonkingly good figure for them really.
That is also wrong as even on their own figures on Sunday Yougov had SNP 40%, Tories 30% in Scotland which would see 10 SNP seats go to the Tories leaving the SNP on 46 seats, so it looks like Yougov haven't even bothered to include their own Scottish figures properly in their overall calculation
A lot depends on whether the Scottish Tactical vote is anti Con or anti SNP. The Tories assume the latter, but could it be the former? Left wing unionists may not be that bothered voting against SNP now that Indyref2 has been kicked into touch by Nicola.
Labour needs independence to stop being the salient issue in Scotland. It can then feasibly win back support that was lost to the SNP. In the long run this is more useful for them than scrapping with the. Tories for Unionist votes.
I don't really understand about this YouGov model. There seem to be two issues: (1) YouGov is showing a much smaller Tory lead than ICM in projected national vote share, and (2) YouGov's constituency model is showing the Tories doing worse than uniform swing would indicate.
But which aspect of this is meant to have been validated by the referendum? Surely it can't be the second, because the geographical distribution of the votes doesn't matter in a referendum. So are they saying that they've switched to a different turnout model for the national vote share, which was validated by the referendum?
Well we got constituency-by-constituency results for the referendum. Even if they didn't actually matter, they could be compared against YouGov's model.
'Why are YouGov trying to put themselves out of business?'
So their competitors are laughing at them then. Still can't believe people are in meltdown over a single projection.
Days ago this lot had the Tory lead up to 7%.
Hold on ! The Tory lead is still 4%. The seats projection is not coming from UNS but their constituency model.
Yes.
Their constituency model might be quite good - even if the VI is overstating Lab.
The lab vote could potentially be nice and efficient.
Tories piling up votes where they either a) don't need them (blue heartlands) and/or b) where they're not strong enough to challenge (lab heartlands)
I want to play with their constituency model. Hope it goes online.
I just don't get why the Tories would be inefficient. They are bringing in a social care charge which will disproportionately hit larger homes in the south so presumably will hold back their vote in seats they already hold
Typically social care costs would be in the tens of thousands, not hundreds. It is a bigger hit on a £180 000 house percentagewise than a million pound one.
Not sure about that. Mrs BJs home care is set to cost £20k per year. She is 57 so probably in excess of £400k That is typical of spinal injury costs. Dementia tax is not only about dementia.
There certainly are exceptions. Most people aren't doing sums though, they don't like the vibe.
Not sure that's exceptional 2 people 75 mins each per day costs that under new system. Old contribution £2k per annum
ICM gives the Tories a lead of 44% to 39% in Labour marginals with a Labour majority of less than 15% and the Tories lead in Tory seats with a majority of less than 10% by 39% to 38%. Labour lead in Labour seats with a majority of more than 15% by 50% to 33% and the Tories in Tory seats with a majority of more than 10% by 59% to 25% so the Yougov figures look rather dubious to me https://www.icmunlimited.com/wp-content/uploads/2017/05/2017_guardian_poll8_may26-29.pdf
I don't really understand about this YouGov model. There seem to be two issues: (1) YouGov is showing a much smaller Tory lead than ICM in projected national vote share, and (2) YouGov's constituency model is showing the Tories doing worse than uniform swing would indicate.
But which aspect of this is meant to have been validated by the referendum? Surely it can't be the second, because the geographical distribution of the votes doesn't matter in a referendum. So are they saying that they've switched to a different turnout model for the national vote share, which was validated by the referendum?
Well we got constituency-by-constituency results for the referendum. Even if they didn't actually matter, they could be compared against YouGov's model.
I did wonder about that. But they seemed to be emphasising was that _something_ had been validated by its ability to predict that Leave was ahead. So apparently the _something_ was a model of national vote share, not the way votes were distributed geographically?
'Why are YouGov trying to put themselves out of business?'
So their competitors are laughing at them then. Still can't believe people are in meltdown over a single projection.
Days ago this lot had the Tory lead up to 7%.
Hold on ! The Tory lead is still 4%. The seats projection is not coming from UNS but their constituency model.
Yes.
Their constituency model might be quite good - even if the VI is overstating Lab.
The lab vote could potentially be nice and efficient.
Tories piling up votes where they either a) don't need them (blue heartlands) and/or b) where they're not strong enough to challenge (lab heartlands)
I want to play with their constituency model. Hope it goes online.
I just don't get why the Tories would be inefficient. They are bringing in a social care charge which will disproportionately hit larger homes in the south so presumably will hold back their vote in seats they already hold
Typically social care costs would be in the tens of thousands, not hundreds. It is a bigger hit on a £180 000 house percentagewise than a million pound one.
Also, if you have a million quid, you are probably able to give 50k to your kids / grandkids easily to help them get on the property ladder. So you feel safe seeing them get on, the way you did.
If on the other hand 180k is all you've got, you're unable to help your kids, and are far more worried about them never owning property and having a lower standard of living than you ever did.
As I've said on here many times, the dementia tax looks disproportionately bad to lower middle class voters in marginal seats and that is why it was a terrible, terrible policy.
The demenatia 'tax' was actually most unpopular in London in the polling
Lots of renters in London hoping to inherit mummy and daddy's house in the shires to help them get on the property ladder?
May should do the debate tomorrow because it won't cost her anything.
It's a far easier format than C4 yesterday or BBC1 QT on Friday.
Seven people giving one minute soundbites with minimal debate or interaction.
Slow paced, nobody gets the chance to build any momentum as you don't get to talk for long enough.
Agreed.
Not sure, she's so ponderous and flat footed and waffly she might make things worse. Tories should have sent Ruth Davidson. She's quick thinking and an excellent speaker.
Surely though with Corbyn gaining in the polls she has to go for it. What is the alternative, giving Corbyn and the also rans prime time publicity? Ok this is only one poll but the trend is clear, and if it continues a Conervative majority is no longer certain.
I see what you're thinking but if she performed anything like as woodenly as she did on the channel 4 thing she'd struggle to get a word in edgeways and flounder if she did. It's not her strength. Her best strategy is to keep campaigning and hope the anti Corbyn vote saves her bacon. It's disappointing to be dragged down to that level but she has nothing else to offer.
Shame really it was nice being able to vote Conservative for positive reasons at the last 2 elections.
This. We can bang on about everything but to my mind it boils down to this: the Tory manifesto got precisely no one excited. There was nothing in it to motivate the Tory vote and bring others over. You can't really polish a turd and Tezza doing more TV ain't going to make a difference at this stage. She just needs to sail on above the fray hoping she looks PM-like and that's what voters will be thinking about when their pencil's hovering over the ballot, and put her trust in Crosby that hammering away on Corbyn's past will have some impact. The course is set: suddenly turning up at TV shows will declare PANIC in capital letters and another u-turn. One more u-turn from May would be fatal.
Retaining 50 seats would be a stonkingly good figure for them really.
That is also wrong as even on their own figures on Sunday Yougov had SNP 40%, Tories 30% in Scotland which would see 10 SNP seats go to the Tories leaving the SNP on 46 seats, so it looks like Yougov haven't even bothered to include their own Scottish figures properly in their overall calculation
A lot depends on whether the Scottish Tactical vote is anti Con or anti SNP. The Tories assume the latter, but could it be the former? Left wing unionists may not be that bothered voting against SNP now that Indyref2 has been kicked into touch by Nicola.
Um Nicola has certainly not kicked Indref2 into touch,she was demanding it is held before Brexit is enacted in her speech earlier today at their manifesto launch. As for the Anti-SNP vote it is pretty strong from what I can gather but whilst there are 50% of Scots who won't vote for the Nats there are probably 60% of Scots would never vote for the Tories. The SNP are therefore only likely to lose seats where they start with less than 45% of the vote or are being challenged by Labour or the LDs.
(Irrelevant anecdote and name dropping alert....) I have just got back from a dinner at Balliol. One of the more amusing conversations was how long it would be before Dave Cameron was recognised as the worst Prime Minister since Anthony Eden, and probably worse even than him.
And did the discussion come to an agreement ?
Basically, in about 6 weeks. Mrs May wins, everyone realises that the Brexit negotiations are a series of circles that cannot be squared, the Nats kick off again, and the Historians, economists, political commentators and assorted academics dust off their word processors and knock out a 15,000 word paper showing that Dave really was crap. I paraphrase, but the view is that he is going to have to carry the can for destroying both the European and domestic Unions. There is no-one else to blame (or praise, if that's your particular bent.)
'Why are YouGov trying to put themselves out of business?'
So their competitors are laughing at them then. Still can't believe people are in meltdown over a single projection.
Days ago this lot had the Tory lead up to 7%.
Hold on ! The Tory lead is still 4%. The seats projection is not coming from UNS but their constituency model.
Yes.
Their constituency model might be quite good - even if the VI is overstating Lab.
The lab vote could potentially be nice and efficient.
Tories piling up votes where they either a) don't need them (blue heartlands) and/or b) where they're not strong enough to challenge (lab heartlands)
I want to play with their constituency model. Hope it goes online.
I just don't get why the Tories would be inefficient. They are bringing in a social care charge which will disproportionately hit larger homes in the south so presumably will hold back their vote in seats they already hold
Typically social care costs would be in the tens of thousands, not hundreds. It is a bigger hit on a £180 000 house percentagewise than a million pound one.
Also, if you have a million quid, you are probably able to give 50k to your kids / grandkids easily to help them get on the property ladder. So you feel safe seeing them get on, the way you did.
If on the other hand 180k is all you've got, you're unable to help your kids, and are far more worried about them never owning property and having a lower standard of living than you ever did.
As I've said on here many times, the dementia tax looks disproportionately bad to lower middle class voters in marginal seats and that is why it was a terrible, terrible policy.
The demenatia 'tax' was actually most unpopular in London in the polling
Lots of renters in London hoping to inherit mummy and daddy's house in the shires to help them get on the property ladder?
Almost certainly and whether the cap helps them much remains to be seen
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?
They have models for every occasion.
When one is exposed as a failure - see Kellner and the Referendum - then wheel out another and claim it predicted the right result.
Kellner during the Referendum wasn't defending a model he was defending a mode: specifically, phone polls vs online polls.
'Why are YouGov trying to put themselves out of business?'
So their competitors are laughing at them then. Still can't believe people are in meltdown over a single projection.
Days ago this lot had the Tory lead up to 7%.
Hold on ! The Tory lead is still 4%. The seats projection is not coming from UNS but their constituency model.
Yes.
Their constituency model might be quite good - even if the VI is overstating Lab.
The lab vote could potentially be nice and efficient.
Tories piling up votes where they either a) don't need them (blue heartlands) and/or b) where they're not strong enough to challenge (lab heartlands)
I want to play with their constituency model. Hope it goes online.
I just don't get why the Tories would be inefficient. They are bringing in a social care charge which will disproportionately hit larger homes in the south so presumably will hold back their vote in seats they already hold
Typically social care costs would be in the tens of thousands, not hundreds. It is a bigger hit on a £180 000 house percentagewise than a million pound one.
Also, if you have a million quid, you are probably able to give 50k to your kids / grandkids easily to help them get on the property ladder. So you feel safe seeing them get on, the way you did.
If on the other hand 180k is all you've got, you're unable to help your kids, and are far more worried about them never owning property and having a lower standard of living than you ever did.
As I've said on here many times, the dementia tax looks disproportionately bad to lower middle class voters in marginal seats and that is why it was a terrible, terrible policy.
The demenatia 'tax' was actually most unpopular in London in the polling
Lots of renters in London hoping to inherit mummy and daddy's house in the shires to help them get on the property ladder?
'Why are YouGov trying to put themselves out of business?'
So their competitors are laughing at them then. Still can't believe people are in meltdown over a single projection.
Days ago this lot had the Tory lead up to 7%.
Hold on ! The Tory lead is still 4%. The seats projection is not coming from UNS but their constituency model.
Yes.
Their constituency model might be quite good - even if the VI is overstating Lab.
The lab vote could potentially be nice and efficient.
Tories piling up votes where they either a) don't need them (blue heartlands) and/or b) where they're not strong enough to challenge (lab heartlands)
I want to play with their constituency model. Hope it goes online.
I just don't get why the Tories would be inefficient. They are bringing in a social care charge which will disproportionately hit larger homes in the south so presumably will hold back their vote in seats they already hold
Even after the social care clusterfuck, the polls are still showing that the Tories' gains on 2015 are still being drawn heavily from the 65+ age bracket - who are almost as inefficiently-distributed under FPTP as students are.
Pensioners are distributed throughout the country - its young people who are more concentrated.
The seats with the highest % of pensioners tend to already be held by the Tories -- very few marginal seats are pensioner hotspots (though an interesting exception to that is the Wirral seats, which have been trending Labour despite the demographics suggesting they should be doing the reverse).
Its the marginals you need to concentrate on.
And if you list the seats with the highest numbers of young voters I predict they'll have far more young voters than the seats with the highest number of retired voters have retired voters.
May should do the debate tomorrow because it won't cost her anything.
It's a far easier format than C4 yesterday or BBC1 QT on Friday.
Seven people giving one minute soundbites with minimal debate or interaction.
Slow paced, nobody gets the chance to build any momentum as you don't get to talk for long enough.
Agreed.
Not sure, she's so ponderous and flat footed and waffly she might make things worse. Tories should have sent Ruth Davidson. She's quick thinking and an excellent speaker.
Surely though with Corbyn gaining in the polls she has to go for it. What is the alternative, giving Corbyn and the also rans prime time publicity? Ok this is only one poll but the trend is clear, and if it continues a Conervative majority is no longer certain.
I see what you're thinking but if she performed anything like as woodenly as she did on the channel 4 thing she'd struggle to get a word in edgeways and flounder if she did. It's not her strength. Her best strategy is to keep campaigning and hope the anti Corbyn vote saves her bacon. It's disappointing to be dragged down to that level but she has nothing else to offer.
Shame really it was nice being able to vote Conservative for positive reasons at the last 2 elections.
This. We can bang on about everything but to my mind it boils down to this: the Tory manifesto got precisely no one excited. There was nothing in it to motivate the Tory vote and bring others over. You can't really polish a turd and Tezza doing more TV ain't going to make a difference at this stage. She just needs to sail on above the fray hoping she looks PM-like and that's what voters will be thinking about when their pencil's hovering over the ballot, and put her trust in Crosby that hammering away on Corbyn's past will have some impact. The course is set: suddenly turning up at TV shows will declare PANIC in capital letters and another u-turn. One more u-turn from May would be fatal.
Crosby hasn't worked any magic yet.
Whether he has or not will only become clear June 9th onward.
Weren't you talking about houses in northern marginal constituencies costing £450k last week ?
Now you casually talk about £180k as if its a pittance.
Do you live in London by any chance ?
Lol.
The example I used last week, referred to a best case scenario of a working class family who had started on the property ladder in the early 80s and traded up to the point they were now living in a prime property.
Something like this 4 bed semi in Newcastle at 425k would do the trick
My theory was that with such a family, other than their pension it would be their only major asset and therefore the ony thing they were able to pass on to their kids / grandkids. They would feel disproportionately hit by the tax as they would not have other assets they were able to liquidate in their old age they could use to help their kids.
I'm well aware of how much houses cost. And I certainly don't see 180k as a pittance.
My theory is that the dementia tax disproportionately worries those people whose major asset is their home and who have few liquid assets they can easily dispose of in their old age to help their kids.
On that basis, the dementia tax is a terrible own goal because it worries precisely the sort of hard working middle class voter in marginal constituencies the Tories should be targeting.
(Irrelevant anecdote and name dropping alert....) I have just got back from a dinner at Balliol. One of the more amusing conversations was how long it would be before Dave Cameron was recognised as the worst Prime Minister since Anthony Eden, and probably worse even than him.
And did the discussion come to an agreement ?
Basically, in about 6 weeks. Mrs May wins, everyone realises that the Brexit negotiations are a series of circles that cannot be squared, the Nats kick off again, and the Historians, economists, political commentators and assorted academics dust off their word processors and knock out a 15,000 word paper showing that Dave really was crap. I paraphrase, but the view is that he is going to have to carry the can for destroying both the European and domestic Unions. There is no-one else to blame (or praise, if that's your particular bent.)
Dave is IIRC a Brasenose man.
Is there a little historical rivalry between Balliol and Brasenose ?
There is simply no way on God's green earth that remain-leave polarisation of the vote favours Labour. Leave won over 400 seats, the Tories are well up with leave voters.
Almost no-one is voting on Brexit. Only the swivel-eyed on both camps. That is why Labour were so wise in ignoring it. Brexit election , my arse! That was settled last year.
No. It has nothing to do with Brexit - but LOOK AT THE REMAIN-LEAVE split in any of the subsamples.
The Tories are clearly doing better with leave and Labour with remain voters. FFSake
I didn't say that. I said few people are voting on Brexit.
Oh I know. But there is very clear evidence that remain and leave voters are swinging in wildly different directions. Its not really about the EU ref vote, but what it shows behind the constituency - it explains very well why Ealing Central and Acton may well look gloomy whereas Torbay is looking like an increased Tory Maj. Also see Bolsover/Carshalton (LD Loss probable)/Twickers (LD Gain being talked up)
Agree. But Brexit has made this election binary. If the 48 % all voted Labour, we would have the most flaccid of Brexits. My point is that many of the 400 seats had wafer-thin Leave majorities. If Labour win 80% of Remainers and Tories win only 75% of Leavers, that is a Labour win.
Basically, there is not much between a Leave or Remain constituency (apart from the extreme few).
Retaining 50 seats would be a stonkingly good figure for them really.
That is also wrong as even on their own figures on Sunday Yougov had SNP 40%, Tories 30% in Scotland which would see 10 SNP seats go to the Tories leaving the SNP on 46 seats, so it looks like Yougov haven't even bothered to include their own Scottish figures properly in their overall calculation
A lot depends on whether the Scottish Tactical vote is anti Con or anti SNP. The Tories assume the latter, but could it be the former? Left wing unionists may not be that bothered voting against SNP now that Indyref2 has been kicked into touch by Nicola.
Um Nicola has certainly not kicked Indref2 into touch,she was demanding it is held before Brexit is enacted in her speech earlier today at their manifesto launch. As for the Anti-SNP vote it is pretty strong from what I can gather but whilst there are 50% of Scots who won't vote for the Nats there are probably 60% of Scots would never vote for the Tories. The SNP are therefore only likely to lose seats where they start with less than 45% of the vote or are being challenged by Labour or the LDs.
I asked this question some two weeks back. I cannot imagine a left person [ even a unionist ] voting Tory. I would rather be dead than vote Tory.
Comments
Giving himself a get-out clause.
I really don't see the point in YouGov publishing this study.
Yougov's public polling certainly didn't show Leave ahead the whole time.
Ashcroft introduced his model with a load of caveats, which might be worth revisiting as they no doubt apply equally to Yougov's model:
Just as traditional polling comes with caveats – that figures are subject to a margin of error, and only represent a snapshot in time – the Ashcroft Model comes with stipulations of its own. To adapt my 2015 mantra, we will be dealing with probabilities, not predictions. The combination of survey and census data will show the probability that certain types of people will behave in certain ways, and the data is presented on that basis. This in turn means that we are applying a national model to specific local areas with their own circumstances. Though we will continue to refine the model as we update the findings each week, we should be prepared for a few anomalies. The seat-by-seat estimates – and, by extension, the implied parliamentary majorities – are therefore not by any means hard and fast predictions.
CON 360
LAB 207
SNP 49
LD 11
Others 23
TMICIPM
When one is exposed as a failure - see Kellner and the Referendum - then wheel out another and claim it predicted the right result.
YouGov are a bunch of jokers.
Secondly, statistics are a mug's game. It was mentioned today that ICM's numbers are different because they apportion the refuse-to-answers mainly to the Tories. They could be right, but it should hardly be a matter of guesswork, should it. After all these years, the successes and the fiascos, there should be some sort of consensus about how to weight polling data!
So consider the following table of votes cast in the local council elections earlier this month. It's a valuable analysis for several reasons:
It's real life data, recently found based on actual votes cast and turnouts
Norwich South is divided into wards, which are also County Divisions so, in theory, of approximately equal size.
Norwich South is a key marginal
Norwich South is a University Seat where there are defined concentrations of student voters most notably in UNIVERSITY ward [funny that] and NELSON wards.
You'll see from the table that the average electorate of all wards in Norwich South is 7935. But the total electorate in the UNIVERSITY ward is only 5419, that is 68pc of average - a reduction of over 2500 electors. Nelson ward, immediately to the east and home to the 2nd & 3rd year student houses is also lower than average in electors at 85% of the constituency average.
Division TurnoutPc Electorate
BOWTHORPE 28.95% 8,114 102%
COSTESSEY 31.88% 11,205 141%
EATON 53.80% 7,323 92%
LAKENHAM 34.81% 7,133 90%
MANCROFT 35.91% 8,109 102%
NELSON 54.22% 6,771 85%
THORPE HAMLET 37.62% 8,923 112%
TOWN CLOSE 46.87% 8,332 105%
UNIVERSITY 36.81% 5,420 68%
WENSUM 33.40% 8,030 101%
Average 7,936
Now, I accept that in the last few weeks more students might have registered to vote... but we can all accept that it's unlikely to have made-up such a backlog in such a short time.
Some students may be registered to vote at home rather than at Uni - but that then disperses the concentration effect and makes University-Town-Ness less potent... that is, if they have applied for a postal vote.
Oh yes, and, given the tendancy for Universities to go for Semesters means that term ended on 19th May for students that aren't sitting exams in the 'assessment period' falling between 22nd May and June 16th.
What I conclude is that young people can talk all they like about political engagement and protest for Corbyn. They can say they're 'certain to vote'. But if they're not registered, they just don't count. And if they've all gone home.
You'd think that University Students would be more politically engaged than their peers... which would imply that the generality of electoral registration is even lower amongst the youngsters.
If the ONS/ElectionData etc mumbo jumbo hasn't taken this into account then it's just another voodoo poll.
Bunnco - Your Man on the Spot
You're supposed to predict something extreme and get the BEJSUS publicity.
No. It has nothing to do with Brexit - but LOOK AT THE REMAIN-LEAVE split in any of the subsamples.
The Tories are clearly doing better with leave and Labour with remain voters. FFSake
Well on a constituency level LEAVE won by a landslide? So that should be good news for Con?
Jesus, these journalists are idiots. (I assume that's a quote from the Times. )
Are you Paul Mason? PaulM = Paul Mason?
I see Gill Troughton is running a broader based anti austerity campaign, particularly on education.
The NUT video is now on 3.3 million views BTW, and Facebook is a mums media.
https://m.facebook.com/story.php?story_fbid=856455754511897&id=514209758736500
What's happening in scotland (decent con vote, but terrible conversion into seats) might be happening in labour seats in E&W too.
And the anti-corbyn stuff is effective, but it's piling up massive majorities in the shires.
The new national Theresa May party could be bloody hopeless under FPTP.
Interesting times.
They will undoubtedly being doing regional poling.
If on the other hand 180k is all you've got, you're unable to help your kids, and are far more worried about them never owning property and having a lower standard of living than you ever did.
As I've said on here many times, the dementia tax looks disproportionately bad to lower middle class voters in marginal seats and that is why it was a terrible, terrible policy.
Her best strategy is to keep campaigning and hope the anti Corbyn vote saves her bacon.
It's disappointing to be dragged down to that level but she has nothing else to offer.
Shame really it was nice being able to vote Conservative for positive reasons at the last 2 elections.
http://lordashcroftpolls.com/2017/05/ashcroft-model-update-estimated-vote-shares-seat-new-potential-majorities/
Now you casually talk about £180k as if its a pittance.
Do you live in London by any chance ?
I have just got back from a dinner at Balliol. One of the more amusing conversations was how long it would be before Dave Cameron was recognised as the worst Prime Minister since Anthony Eden, and probably worse even than him.
Paddy needs to un-eat his hat!
(1) YouGov is showing a much smaller Tory lead than ICM in projected national vote share, and
(2) YouGov's constituency model is showing the Tories doing worse than uniform swing would indicate.
But which aspect of this is meant to have been validated by the referendum? Surely it can't be the second, because the geographical distribution of the votes doesn't matter in a referendum. So are they saying that they've switched to a different turnout model for the national vote share, which was validated by the referendum?
Labour lead in Labour seats with a majority of more than 15% by 50% to 33% and the Tories in Tory seats with a majority of more than 10% by 59% to 25% so the Yougov figures look rather dubious to me
https://www.icmunlimited.com/wp-content/uploads/2017/05/2017_guardian_poll8_may26-29.pdf
As for the Anti-SNP vote it is pretty strong from what I can gather but whilst there are 50% of Scots who won't vote for the Nats there are probably 60% of Scots would never vote for the Tories. The SNP are therefore only likely to lose seats where they start with less than 45% of the vote or are being challenged by Labour or the LDs.
And if you list the seats with the highest numbers of young voters I predict they'll have far more young voters than the seats with the highest number of retired voters have retired voters.
The example I used last week, referred to a best case scenario of a working class family who had started on the property ladder in the early 80s and traded up to the point they were now living in a prime property.
Something like this 4 bed semi in Newcastle at 425k would do the trick
http://www.rightmove.co.uk/property-for-sale/property-66250469.html
My theory was that with such a family, other than their pension it would be their only major asset and therefore the ony thing they were able to pass on to their kids / grandkids. They would feel disproportionately hit by the tax as they would not have other assets they were able to liquidate in their old age they could use to help their kids.
I'm well aware of how much houses cost. And I certainly don't see 180k as a pittance.
My theory is that the dementia tax disproportionately worries those people whose major asset is their home and who have few liquid assets they can easily dispose of in their old age to help their kids.
On that basis, the dementia tax is a terrible own goal because it worries precisely the sort of hard working middle class voter in marginal constituencies the Tories should be targeting.
That was my point then and remains my point now.
Is there a little historical rivalry between Balliol and Brasenose ?
No matter what you think it is desperately, desperately unhealthy.
Basically, there is not much between a Leave or Remain constituency (apart from the extreme few).
BoZo would be top of my list