For the first time in years of doing Yougov surveys they asked me how I'd vote... so that explains the sudden change...
On the Google Docs spreadsheet posted earlier, is it read only? I can't see a space for current votes in it either way for tracking purposes.
I wonder if YouGov might have tried to test the notion of a Momentum sign-up surge - by going back to those who have not joined their panel in the last say 12 months?
I have been at a family funeral, the second in two weeks. It was a magnificent Brideshead-style Anglo-Irish Catholic service followed by a brilliant, funny and boozy wake that my much beloved cousin would have loved!
I see that @david_herdson has joined me in thinking that Corbyn might just do it.
I don't claim anything approaching your combined predictive value, but I feel similarly.
"A friend of mine who lives here but is registered in Hull had applied for her mum to vote by proxy for her, but her by proxy card never arrived for her mum...is there anything she can do? (I know it's not local but every vote matters and she's super annoyed if she has just lost her vote!!!)"
Any advice?
People can apply for emergency proxy votes on the day of the election even, so electoral services must have staff to handle queries during the day - so I would get in touch with them directly sharpish, and if there truly has been some problems, I'm sure they would try to assist. I'm sure I know of presiding officers at stations being able to help too, and they'll have a hotline to more senior people.
David's comments this evening appear to me to be uncharacteristic and I write as someone who has the highest regard for him and what he writes. I think it's fairly clear that he's currently under a good deal of pressure and whilst I don't doubt that what he reports is entirely accurate I think we need to bear in mind that his conclusions are to a large extent based on one night's canvassing in a small part of just one constituency.
So they faff about for the fortnight prior to the election then change their methodology for the final poll?
I give up. the cynic in me would suggest it's awfully convenient that their methodology showed a narrower lead around the time the campaign in chaos narrative picked up...
If it's NOC, YouGov can point to their mega-wierd-strange-rolling-mass poll and say "hey, we predicted that (within MOE), hire us!"
If it's a Conservative landslide, YouGov can point to this regular poll and say "hey, we predicted that (within MOE), hire us!"
... "For now, YouGov’s final call for the 2017 election is for a seven point Conservative lead, leading to an increased Conservative majority in the Commons." ...
Brilliant! on the one hand, they're forecasting a substantial loss of Conservative seats. On the other they're forecasting an increased Conservative majority . On the same day.
Looks like slight herding to me. YouGov made methodology changes as outlined below:
"As this is our final call before the election we made two minor changes to our method. The first is that rather than asking people which party they’d vote for, we showed respondents a list of the people actually standing in their constituency and asked which one they would vote for. Hopefully this will help pick up any tactical vote considerations and remove any issue of people saying they would vote UKIP or Green in seats where UKIP or the Greens are not actually standing.
Secondly we have reallocated those respondents who say don’t know, but who also say they are very likely to vote (voters who my colleague Adam McDonnell described earlier in the campaign as “true undecided”). We assume uncertain voters who say they “don’t know” at this stage won’t actually vote, but those who say they are 8+/10 certain to vote we have reallocated back to the party they voted for in 2015."
They clearly weren't confident about their own predictions, they're quite keen to state they think there'll be an increased majority:
"For now, YouGov’s final call for the 2017 election is for a seven point Conservative lead, leading to an increased Conservative majority in the Commons."
ALSO: Remember that their model is showing Labour's standing increasing over the last few days. So this doesn't sound that accurate to me, showing a steep 3 point drop. I just think they're bricking it. It's probably more accurate now, though.
It seems to me like data scientists are like criminologists and neuroscientists of yore. Far too many of them with crappy degrees and no real use for them. Experts!
... "For now, YouGov’s final call for the 2017 election is for a seven point Conservative lead, leading to an increased Conservative majority in the Commons." ...
Brilliant! on the one hand, they're forecasting a substantial loss of Conservative seats. On the other they're forecasting an increased Conservative majority . On the same day.
Well done to those from all parties who have put the effort in on the ground. I've no doubt it is a draining experience as well as sometimes rewarding.
On the big picture, despite a poor Tory campaign I think there are too many folk for whom life is ok to create a big enough groundswell for an experiment in Corbynism. I shall be surprised if the Tory overall majority is lower than 40.
Looks like slight herding to me. YouGov made methodology changes as outlined below:
"As this is our final call before the election we made two minor changes to our method. The first is that rather than asking people which party they’d vote for, we showed respondents a list of the people actually standing in their constituency and asked which one they would vote for. Hopefully this will help pick up any tactical vote considerations and remove any issue of people saying they would vote UKIP or Green in seats where UKIP or the Greens are not actually standing.
Secondly we have reallocated those respondents who say don’t know, but who also say they are very likely to vote (voters who my colleague Adam McDonnell described earlier in the campaign as “true undecided”). We assume uncertain voters who say they “don’t know” at this stage won’t actually vote, but those who say they are 8+/10 certain to vote we have reallocated back to the party they voted for in 2015."
They clearly weren't confident about their own predictions, they're quite keen to state they think there'll be an increased majority:
"For now, YouGov’s final call for the 2017 election is for a seven point Conservative lead, leading to an increased Conservative majority in the Commons."
ALSO: Remember that their model is showing Labour's standing increasing over the last few days. So this doesn't sound that accurate to me, showing a steep 3 point drop. I just think they're bricking it. It's probably more accurate now, though.
Job done though. Waverers well and truly spooked, troops motivated for last minute canvassing. And a much more exciting couple of weeks for the newspapers :-)
Answering my own question, I see YouGov's pre-adjustment data shows a Tory lead of just *2* (33-31, which would found up to 3 when don't knows are eliminated). Tory certainty to vote plus the tweaks appear to account for the entire difference. May's ratings are shown as falling heavily and Corbyn's rising compared with two weeks earlier - but hat's for the full poll before turnout weighting.
Complicated, but important to unerstand if you're betting on it.
For the first time in years of doing Yougov surveys they asked me how I'd vote... so that explains the sudden change...
On the Google Docs spreadsheet posted earlier, is it read only? I can't see a space for current votes in it either way for tracking purposes.
I wonder if YouGov might have tried to test the notion of a Momentum sign-up surge - by going back to those who have not joined their panel in the last say 12 months?
That's an interesting theory. Really interesting theory.
... "For now, YouGov’s final call for the 2017 election is for a seven point Conservative lead, leading to an increased Conservative majority in the Commons." ...
Brilliant! on the one hand, they're forecasting a substantial loss of Conservative seats. On the other they're forecasting an increased Conservative majority . On the same day.
Looks like slight herding to me. YouGov made methodology changes as outlined below:
"As this is our final call before the election we made two minor changes to our method. The first is that rather than asking people which party they’d vote for, we showed respondents a list of the people actually standing in their constituency and asked which one they would vote for. Hopefully this will help pick up any tactical vote considerations and remove any issue of people saying they would vote UKIP or Green in seats where UKIP or the Greens are not actually standing.
Secondly we have reallocated those respondents who say don’t know, but who also say they are very likely to vote (voters who my colleague Adam McDonnell described earlier in the campaign as “true undecided”). We assume uncertain voters who say they “don’t know” at this stage won’t actually vote, but those who say they are 8+/10 certain to vote we have reallocated back to the party they voted for in 2015."
They clearly weren't confident about their own predictions, they're quite keen to state they think there'll be an increased majority:
"For now, YouGov’s final call for the 2017 election is for a seven point Conservative lead, leading to an increased Conservative majority in the Commons."
ALSO: Remember that their model is showing Labour's standing increasing over the last few days. So this doesn't sound that accurate to me, showing a steep 3 point drop. I just think they're bricking it. It's probably more accurate now, though.
Job done though. Waverers well and truly spooked, troops motivated for last minute canvassing. And a much more exciting couple of weeks for the newspapers :-)
And have they forgotten how much Lucas boasted that the Greens have a co-leader system because they do things differently (conveniently not mentioning that they got rid of it the last time she was leader)?
All this election fever is distracting us from the amazing shit going own in America. Comey's written testimony is sensational and the Senate have actually put Obamacare repeal on the fast track and it could be done and dusted by July.
I'd get on Dem gains in the mid-terms right now.
Yep, he found a hammer and a nail on his way in.
On another major news note, given tensions in the Gulf region, that IS attack in Tehran, given live updates via IS' Amaq news agency, is a serious blow to Iran's sense of immunity to such attacks. They blamed Saudi Arabia, partially or totally because they had to. They really would not want to admit IS has got through the gate.
David's comments this evening appear to me to be uncharacteristic and I write as someone who has the highest regard for him and what he writes. I think it's fairly clear that he's currently under a good deal of pressure and whilst I don't doubt that what he reports is entirely accurate I think we need to bear in mind that his conclusions are to a large extent based on one night's canvassing in a small part of just one constituency.
I was with our esteemed Nicholas Soames in 2010 (He's blocked me on twitter over Brexit since...) and he was worried even during the count. I wondered around and told him he'd win comfortably and he did... Some people worry more than others.
Being pedantic, how can two people both describes themselves as THE Conservative Party Candidate? Surely A Conservative Party Candidate would be better?
So they faff about for the fortnight prior to the election then change their methodology for the final poll?
I give up. the cynic in me would suggest it's awfully convenient that their methodology showed a narrower lead around the time the campaign in chaos narrative picked up...
These mid-campaign changes in methodology are completely pathetic and make a mockery of the polling industry's claims to be in any way scientific.
If they're going to make changes at this stage, they need to re-state the previous poll's figures so as to be comparing apples with apples. As it is, comparisons with the previous polls are meaningless.
Possibly the first time ever I've managed to post on my mobile. For 2 years I thought this day would never come on Vanilla Forums! The yougov doesn't make me any more confident as it does appear to be herding. I'm hopeful DH has had a long day and maybe come tomorrow after a bit of sleep he might offer a morsal of optimism to support my 50-175 majority spread which is where all my money has gone. I would be surprised of any last minute swing towards Labour, it does feel like if there is any momentum now it's with the Blues, but I'm not in a Northern marginal and I respect DH's words enough to have a jitter or two.
Looking at the Wikipedia overview, it really does look as though the London terrorism produced a 4-point swing from Lab to Con. Obviously it could be a coincidence with so much else going on, but I wonder if the single "Enough is Enough" phrase cut through to a small but siginificant slice of waverers.
David Herdson has clearly had a bad session but we all have those sometimes - take a break for a few days once the election's over, you must have earned it.
Don't think so Nick.
When Sunil publishes his final ELBOW I think it's going to show the position almost unchanged from last week.
The last 3 days has brought stabilisation (after several weeks of the Con lead falling).
What caused this - who knows - maybe Fri QT helped, maybe London attacks were also a factor, maybe a few people having a rethink at the last minute.
Serious? question. The conventional wisdom is that if the election results in a basically standstill position then it will be a disaster for the Tories. But thinking totally objectively, is this an over-reaction based on what might have been. Because all other things being equal surely a breakeven position is still a 'gain' however slight. Negotiating Brexit is going to be a nightmare and the simple fact of extending the period before another election is required would be helpful in itself. Even without nothing else.
Looks like slight herding to me. YouGov made methodology changes as outlined below:
"As this is our final call before the election we made two minor changes to our method. The first is that rather than asking people which party they’d vote for, we showed respondents a list of the people actually standing in their constituency and asked which one they would vote for. Hopefully this will help pick up any tactical vote considerations and remove any issue of people saying they would vote UKIP or Green in seats where UKIP or the Greens are not actually standing.
Secondly we have reallocated those respondents who say don’t know, but who also say they are very likely to vote (voters who my colleague Adam McDonnell described earlier in the campaign as “true undecided”). We assume uncertain voters who say they “don’t know” at this stage won’t actually vote, but those who say they are 8+/10 certain to vote we have reallocated back to the party they voted for in 2015."
They clearly weren't confident about their own predictions, they're quite keen to state they think there'll be an increased majority:
"For now, YouGov’s final call for the 2017 election is for a seven point Conservative lead, leading to an increased Conservative majority in the Commons."
ALSO: Remember that their model is showing Labour's standing increasing over the last few days. So this doesn't sound that accurate to me, showing a steep 3 point drop. I just think they're bricking it. It's probably more accurate now, though.
That's helpful, thanks - what they've done is similar to part of ICM's tweaks - ICM also (in my view rather contrroversially) override don't knows, though ICM go further and override people who won't say how they voted last time either (by *guessing* that they'll behave like *other* people who've said they didn't know in past elections).
Do we have a figure for raw "party preferred before turnout adjustment" from any of these?
ICM had Con 43.5 Lab 39.5 before turnout adjustments , Table 3 after eliminating don't knows etc
Answering my own question, I see YouGov's pre-adjustment data shows a Tory lead of just *2* (33-31, which would found up to 3 when don't knows are eliminated). Tory certainty to vote plus the tweaks appear to account for the entire difference. May's ratings are shown as falling heavily and Corbyn's rising compared with two weeks earlier - but hat's for the full poll before turnout weighting.
Complicated, but important to unerstand if you're betting on it.
People get more partisan as polling day approaches. So, people who are now going to vote Labour have decided that Corbyn is best PM, Labour have best policies etc., whereas previously they'd say they were undecided.
David's comments this evening appear to me to be uncharacteristic and I write as someone who has the highest regard for him and what he writes. I think it's fairly clear that he's currently under a good deal of pressure and whilst I don't doubt that what he reports is entirely accurate I think we need to bear in mind that his conclusions are to a large extent based on one night's canvassing in a small part of just one constituency.
So they faff about for the fortnight prior to the election then change their methodology for the final poll?
I give up. the cynic in me would suggest it's awfully convenient that their methodology showed a narrower lead around the time the campaign in chaos narrative picked up...
These mid-campaign changes in methodology are completely pathetic and make a mockery of the polling industry's claims to be in any way scientific.
If they're going to make changes at this stage, they need to re-state the previous poll's figures so as to be comparing apples with apples. As it is, comparisons with the previous polls are meaningless.
The real laugh being they described the changes as "minor", despite totally changing how they prompt for voting intention!
The Conservatives have launched a barrage of targeted Facebook adverts ahead of tomorrow's general election that make the most of a previously unnoticed loophole in electoral law to enable the party to spend a near-unlimited amount on localised messages in key target constituencies, BuzzFeed News has learned.
The party's superior online advertising efforts have already attracted substantial attention in recent weeks, with hundreds of thousands of pounds being spent pushing anti-Labour attack adverts into Facebook feeds. The majority have focused on the supposed deficiencies in Jeremy Corbyn's policies and his team.
Localised online political advertising has previously been constrained by a UK law that requires any material mentioning a local party candidate to be counted as part of strict local campaign spending limits, which can be as low as £12,000 per candidate per constituency.
However, the Conservatives have now spotted a previously underused loophole that means even adverts that mention a specific constituency – but not the local candidate or local policies – can be paid out of the national campaign budget, which can be up to £19 million.
"If it's promoting the national party and national policies, then regardless of whether it mentions the constituency name or not then it's national spend," an Electoral Commission spokesperson told BuzzFeed News, adding that the situation is different if it talks "about a candidate's views and a candidate's policies".
The end result is that Conservative paid-for Facebook ads that only appear to voters in a single targeted constituency and look remarkably like localised adverts can be legally counted as national spending.
In one example from Westminster North, obtained by BuzzFeed News, the advert repeatedly mentions "your constituency", features the words "Westminster North" in the graphic, and emphasises that local voters in the constituency could determine who gets into Downing Street.
Underlining the extent to which the advert is specifically targeted at local voters in that constituency, it also quotes an Evening Standard headline that reads "Shock surge for Labour in London".
David's comments this evening appear to me to be uncharacteristic and I write as someone who has the highest regard for him and what he writes. I think it's fairly clear that he's currently under a good deal of pressure and whilst I don't doubt that what he reports is entirely accurate I think we need to bear in mind that his conclusions are to a large extent based on one night's canvassing in a small part of just one constituency.
I was with our esteemed Nicholas Soames in 2010 (He's blocked me on twitter over Brexit since...) and he was worried even during the count. I wondered around and told him he'd win comfortably and he did... Some people worry more than others.
The big man does like the block button. I got blocked too and I have absolutely no idea why.
Being pedantic, how can two people both describes themselves as THE Conservative Party Candidate? Surely A Conservative Party Candidate would be better?
They are both candidates for THE Conservative party, not any old Conservative party
Possibly the first time ever I've managed to post on my mobile. For 2 years I thought this day would never come on Vanilla Forums! The yougov doesn't make me any more confident as it does appear to be herding. I'm hopeful DH has had a long day and maybe come tomorrow after a bit of sleep he might offer a morsal of optimism to support my 50-175 majority spread which is where all my money has gone. I would be surprised of any last minute swing towards Labour, it does feel like if there is any momentum now it's with the Blues, but I'm not in a Northern marginal and I respect DH's words enough to have a jitter or two.
Being pedantic, how can two people both describes themselves as THE Conservative Party Candidate? Surely A Conservative Party Candidate would be better?
Should really be 'The Conservative Party' candidate.
Had one in the recent locals where a candidate was on the nominations list twice - once withdrawn as they'd forgotten to include party nomination papers it seemed, so they would have been listed as blank rather than a party candidate.
Looking at the Wikipedia overview, it really does look as though the London terrorism produced a 4-point swing from Lab to Con. Obviously it could be a coincidence with so much else going on, but I wonder if the single "Enough is Enough" phrase cut through to a small but siginificant slice of waverers.
David Herdson has clearly had a bad session but we all have those sometimes - take a break for a few days once the election's over, you must have earned it.
Let's see what happens tomorrow, but also:
For about a fortnight my social media feed was "Dementia Tax! Vote Labour!" then "Police Cuts! Vote Labour!"
For the last 48 hours it has been "Stop the Racists and Misogynists by showing your solidarity with Diane Abbott! Vote Labour!"
I'm going out on a limb but I think one of these messages may be less effective with floating voters than the other two.
By which I mean, I think Labour got played by a dead cat. Again. Jack Monroe's article called "We Need To Talk About Diane Abbott" has gone megaviral.
You know who actually needs to talk about Diane Abbott right now? Tories. But go ahead and do it for them, Jack.
So they faff about for the fortnight prior to the election then change their methodology for the final poll?
I give up. the cynic in me would suggest it's awfully convenient that their methodology showed a narrower lead around the time the campaign in chaos narrative picked up...
These mid-campaign changes in methodology are completely pathetic and make a mockery of the polling industry's claims to be in any way scientific.
If they're going to make changes at this stage, they need to re-state the previous poll's figures so as to be comparing apples with apples. As it is, comparisons with the previous polls are meaningless.
The real laugh being they described the changes as "minor", despite totally changing how they prompt for voting intention!
Quite right. I'd say it's almost tantamount to intellectual fraud. Fine to change your model but only after it's been discredited by events. Hacking around with the algorithms because they don't 'feel right' (what other motivation can there be?) reduces the whole exercise to mysticism.
Being pedantic, how can two people both describes themselves as THE Conservative Party Candidate? Surely A Conservative Party Candidate would be better?
They are both candidates for THE Conservative party, not any old Conservative party
All this election fever is distracting us from the amazing shit going own in America. Comey's written testimony is sensational and the Senate have actually put Obamacare repeal on the fast track and it could be done and dusted by July.
I'd get on Dem gains in the mid-terms right now.
Yep, he found a hammer and a nail on his way in.
On another major news note, given tensions in the Gulf region, that IS attack in Tehran, given live updates via IS' Amaq news agency, is a serious blow to Iran's sense of immunity to such attacks. They blamed Saudi Arabia, partially or totally because they had to. They really would not want to admit IS has got through the gate.
I honestly believe this is happening with pollsters. They get their samples, do their usual weightings and then say "this cannot be right !"
Then they start playing around with the figures to arrive at what they would like to arrive at.
I've said it before and I'll say it again, the BPC should outlaw this for their members. The pollsters should if they are serious determine their methodology in advance and then do their predetermined calculations and publish or be damned. If this means they're wrong fine, it means their methodology or sampling was wrong. If they tweak the results in order to be right then that just means their untweaked results were wrong but that's being masked.
If somehow YouGov are "most accurate pollster" this year then that could be highlighted time and again for the next five years with their generic polls despite their generic polls being all wrong until the tweak.
I can only get 7 parties: Con, Lab, LD, Grn, UKIP, PC & SNP. Who is the 8th?
Womens' Equality Party
Oh come on, if they're included there have to be others who deserve it just as much.
Christian People's Alliance, The Yorkshire Party, Alliance, SDLP, Sinn Fein, DUP, UUP, OMRLP, BNP, Pirates, English Democrats and Green Party in NI have more or as many candidates.
Even if you exclude NI, that's plenty just as deserving.
David Herdson (who I regard as a top level poster on here) is concerned we (CON) won't win Wakefield.
He is correct in this analysis. We won't win Wakefield. But that was always the case. We were never going to win a seat that LAB has held since 1931. But that doesn't justify a projection that we will get 325 max.
JackW and JohnO are top level analysts here. They have both projected a CON maj 100. But that won't happen because again it is dependent on CON winning lots of seats which we haven't won since 1931. That won't happen.
BUT we are 10% clear. We may struggle in London and South although I am feeling more comfortable now. We will do ok in Midlands and North eg will win Wolverhampton SW, Barrow, Chester maybe Copeland. We will also win at least 10 seats in Scotland.
Ave it has always projected a 10% lead and around a 40 maj - it might be slightly better than that.
David's comments this evening appear to me to be uncharacteristic and I write as someone who has the highest regard for him and what he writes. I think it's fairly clear that he's currently under a good deal of pressure and whilst I don't doubt that what he reports is entirely accurate I think we need to bear in mind that his conclusions are to a large extent based on one night's canvassing in a small part of just one constituency.
I was with our esteemed Nicholas Soames in 2010 (He's blocked me on twitter over Brexit since...) and he was worried even during the count. I wondered around and told him he'd win comfortably and he did... Some people worry more than others.
The big man does like the block button. I got blocked too and I have absolutely no idea why.
He's no longer the big man, he really has lost a lot of weight.
I'm not a whisky drinker but I have two miniatures in the pantry: a 10 year old Laphroaig and a 18 year old Glenfiddich. Which one for the exit poll tomorrow?
Both.
Glenfiddich first then savour the Laphroaig.
Very sound advice, if I may say so.
I agree Glenfiddich first. But you can't savour a Laphroaig. It has no complexity. You just experience it.
David Herdson (who I regard as a top level poster on here) is concerned we (CON) won't win Wakefield.
He is correct in this analysis. We won't win Wakefield. But that was always the case. We were never going to win a seat that LAB has held since 1931. But that doesn't justify a projection that we will get 325 max.
JackW and JohnO are top level analysts here. They have both projected a CON maj 100. But that won't happen because again it is dependent on CON winning lots of seats which we haven't won since 1931. That won't happen.
BUT we are 10% clear. We may struggle in London and South although I am feeling more comfortable now. We will do ok in Midlands and North eg will win Wolverhampton SW, Barrow, Chester maybe Copeland. We will also win at least 10 seats in Scotland.
Ave it has always projected a 10% lead and around a 40 maj - it might be slightly better than that.
BUT we are 10% clear. We may struggle in London and South although I am feeling more comfortable now. We will do ok in Midlands and North eg will win Wolverhampton SW, Barrow, Chester maybe Copeland. We will also win at least 10 seats in Scotland.
Struggle in the south? I presume you mean to make any further gains there, as there's not many non-Tories to take!
Looks like slight herding to me. YouGov made methodology changes as outlined below:
"As this is our final call before the election we made two minor changes to our method. The first is that rather than asking people which party they’d vote for, we showed respondents a list of the people actually standing in their constituency and asked which one they would vote for. Hopefully this will help pick up any tactical vote considerations and remove any issue of people saying they would vote UKIP or Green in seats where UKIP or the Greens are not actually standing.
Secondly we have reallocated those respondents who say don’t know, but who also say they are very likely to vote (voters who my colleague Adam McDonnell described earlier in the campaign as “true undecided”). We assume uncertain voters who say they “don’t know” at this stage won’t actually vote, but those who say they are 8+/10 certain to vote we have reallocated back to the party they voted for in 2015."
They clearly weren't confident about their own predictions, they're quite keen to state they think there'll be an increased majority:
"For now, YouGov’s final call for the 2017 election is for a seven point Conservative lead, leading to an increased Conservative majority in the Commons."
ALSO: Remember that their model is showing Labour's standing increasing over the last few days. So this doesn't sound that accurate to me, showing a steep 3 point drop. I just think they're bricking it. It's probably more accurate now, though.
It seems to me like data scientists are like criminologists and neuroscientists of yore. Far too many of them with crappy degrees and no real use for them. Experts!
I honestly believe this is happening with pollsters. They get their samples, do their usual weightings and then say "this cannot be right !"
Then they start playing around with the figures to arrive at what they would like to arrive at.
I've said it before and I'll say it again, the BPC should outlaw this for their members. The pollsters should if they are serious determine their methodology in advance and then do their predetermined calculations and publish or be damned. If this means they're wrong fine, it means their methodology or sampling was wrong. If they tweak the results in order to be right then that just means their untweaked results were wrong but that's being masked.
If somehow YouGov are "most accurate pollster" this year then that could be highlighted time and again for the next five years with their generic polls despite their generic polls being all wrong until the tweak.
Despite being wrong the La Times/USC poll was utterly amazing for making its data and model completely transparent. That should be the gold standard of polling.
Possibly the first time ever I've managed to post on my mobile. For 2 years I thought this day would never come on Vanilla Forums! The yougov doesn't make me any more confident as it does appear to be herding. I'm hopeful DH has had a long day and maybe come tomorrow after a bit of sleep he might offer a morsal of optimism to support my 50-175 majority spread which is where all my money has gone. I would be surprised of any last minute swing towards Labour, it does feel like if there is any momentum now it's with the Blues, but I'm not in a Northern marginal and I respect DH's words enough to have a jitter or two.
Does it work now? What phone do you use?
I'm using an iPhone 6s. I swear it has never worked up until my attempt today. God, I'll never be off the forum now
David Herdson (who I regard as a top level poster on here) is concerned we (CON) won't win Wakefield.
He is correct in this analysis. We won't win Wakefield. But that was always the case. We were never going to win a seat that LAB has held since 1931. But that doesn't justify a projection that we will get 325 max.
JackW and JohnO are top level analysts here. They have both projected a CON maj 100. But that won't happen because again it is dependent on CON winning lots of seats which we haven't won since 1931. That won't happen.
BUT we are 10% clear. We may struggle in London and South although I am feeling more comfortable now. We will do ok in Midlands and North eg will win Wolverhampton SW, Barrow, Chester maybe Copeland. We will also win at least 10 seats in Scotland.
Ave it has always projected a 10% lead and around a 40 maj - it might be slightly better than that.
But, but - what about Bootle?
I stick to my prediction of a 1979-type result, if you add SNP to Labour.
The Conservatives have launched a barrage of targeted Facebook adverts ahead of tomorrow's general election that make the most of a previously unnoticed loophole in electoral law to enable the party to spend a near-unlimited amount on localised messages in key target constituencies, BuzzFeed News has learned.
The party's superior online advertising efforts have already attracted substantial attention in recent weeks, with hundreds of thousands of pounds being spent pushing anti-Labour attack adverts into Facebook feeds. The majority have focused on the supposed deficiencies in Jeremy Corbyn's policies and his team.
I heard a rumour that Conservative spend on YouTube was mind blowing. Saturation bombing. Spend was ramped up heavily around the time of the Corbyn is a friend of the IRA attack ad.
At the time I saw the increase in spend as a huge weakness, a sign of panic even. But the second terrorist attack has turned the conversation in the final week heavily towards security and this has worked out very well for the Tories.
Last week I had a lot of people telling me that attack ad was unfair and it made them feel sorry for Corbyn, this week if it's been talked about at all it has been 'oh, maybe the Tories have a point after all'.
YouGov may be herding but I do think there has been substantial swingback to the Tories after the London terrorist attack and the 'enough is enough' line has definitely cut through.
If 3 different posters on this site have never heard of WEP, I am going to guess they are not going to get many votes tomorrow...
Oh I've heard of the party, but not the leader - they're standing fewer people than the Pirates, the Loonies or the BNP, why should anyone know who the leader is to the point of highlighting on a front page? Very very odd.
Comments
If it's a Conservative landslide, YouGov can point to this regular poll and say "hey, we predicted that (within MOE), hire us!"
Covering all the bases.
WillS.
You know what? I think they'll be right.
https://twitter.com/ShippersUnbound/status/872446020118073344
ELBOW for the SEVEN polls so far this week:
Con 43.43
Lab 35.86
LD 7.71
UKIP 4.71
Tory lead 7.57
On the big picture, despite a poor Tory campaign I think there are too many folk for whom life is ok to create a big enough groundswell for an experiment in Corbynism. I shall be surprised if the Tory overall majority is lower than 40.
Perhaps he voted but for someone else as he didn't want to actually win ! Why he'd think he'd win as TUSC though I don't know.
From Twitter, this remake of Dallas looks shit
Complicated, but important to unerstand if you're betting on it.
And have they forgotten how much Lucas boasted that the Greens have a co-leader system because they do things differently (conveniently not mentioning that they got rid of it the last time she was leader)?
On another major news note, given tensions in the Gulf region, that IS attack in Tehran, given live updates via IS' Amaq news agency, is a serious blow to Iran's sense of immunity to such attacks. They blamed Saudi Arabia, partially or totally because they had to. They really would not want to admit IS has got through the gate.
The attackers notably did not speak Farsi.
If they're going to make changes at this stage, they need to re-state the previous poll's figures so as to be comparing apples with apples. As it is, comparisons with the previous polls are meaningless.
Then they start playing around with the figures to arrive at what they would like to arrive at.
The yougov doesn't make me any more confident as it does appear to be herding.
I'm hopeful DH has had a long day and maybe come tomorrow after a bit of sleep he might offer a morsal of optimism to support my 50-175 majority spread which is where all my money has gone.
I would be surprised of any last minute swing towards Labour, it does feel like if there is any momentum now it's with the Blues, but I'm not in a Northern marginal and I respect DH's words enough to have a jitter or two.
When Sunil publishes his final ELBOW I think it's going to show the position almost unchanged from last week.
The last 3 days has brought stabilisation (after several weeks of the Con lead falling).
What caused this - who knows - maybe Fri QT helped, maybe London attacks were also a factor, maybe a few people having a rethink at the last minute.
He needs a long rest after all his work.
The party's superior online advertising efforts have already attracted substantial attention in recent weeks, with hundreds of thousands of pounds being spent pushing anti-Labour attack adverts into Facebook feeds. The majority have focused on the supposed deficiencies in Jeremy Corbyn's policies and his team.
Localised online political advertising has previously been constrained by a UK law that requires any material mentioning a local party candidate to be counted as part of strict local campaign spending limits, which can be as low as £12,000 per candidate per constituency.
However, the Conservatives have now spotted a previously underused loophole that means even adverts that mention a specific constituency – but not the local candidate or local policies – can be paid out of the national campaign budget, which can be up to £19 million.
"If it's promoting the national party and national policies, then regardless of whether it mentions the constituency name or not then it's national spend," an Electoral Commission spokesperson told BuzzFeed News, adding that the situation is different if it talks "about a candidate's views and a candidate's policies".
The end result is that Conservative paid-for Facebook ads that only appear to voters in a single targeted constituency and look remarkably like localised adverts can be legally counted as national spending.
In one example from Westminster North, obtained by BuzzFeed News, the advert repeatedly mentions "your constituency", features the words "Westminster North" in the graphic, and emphasises that local voters in the constituency could determine who gets into Downing Street.
Underlining the extent to which the advert is specifically targeted at local voters in that constituency, it also quotes an Evening Standard headline that reads "Shock surge for Labour in London".
https://www.buzzfeed.com/jimwaterson/the-tories-are-exploiting-a-new-loophole-to-launch-a-last?utm_term=.ehJy17wM59#.jtlQDzMnbN
https://twitter.com/hendopolis/status/872569441682432000
I struggle to see how anyone could find such a groundswell of opposition given whose doors they have been asked to knock on.
.
.
.
.
I'll get my coat.
Had one in the recent locals where a candidate was on the nominations list twice - once withdrawn as they'd forgotten to include party nomination papers it seemed, so they would have been listed as blank rather than a party candidate.
Mrs Fleet is a presiding officer... I am trying to persuade her to put a bottle of tippex on her desk at the polling station.
DH may well be on the zeitgeist.
For about a fortnight my social media feed was "Dementia Tax! Vote Labour!" then "Police Cuts! Vote Labour!"
For the last 48 hours it has been "Stop the Racists and Misogynists by showing your solidarity with Diane Abbott! Vote Labour!"
I'm going out on a limb but I think one of these messages may be less effective with floating voters than the other two.
By which I mean, I think Labour got played by a dead cat. Again. Jack Monroe's article called "We Need To Talk About Diane Abbott" has gone megaviral.
You know who actually needs to talk about Diane Abbott right now? Tories. But go ahead and do it for them, Jack.
If somehow YouGov are "most accurate pollster" this year then that could be highlighted time and again for the next five years with their generic polls despite their generic polls being all wrong until the tweak.
Christian People's Alliance, The Yorkshire Party, Alliance, SDLP, Sinn Fein, DUP, UUP, OMRLP, BNP, Pirates, English Democrats and Green Party in NI have more or as many candidates.
Even if you exclude NI, that's plenty just as deserving.
David Herdson (who I regard as a top level poster on here) is concerned we (CON) won't win Wakefield.
He is correct in this analysis. We won't win Wakefield. But that was always the case. We were never going to win a seat that LAB has held since 1931. But that doesn't justify a projection that we will get 325 max.
JackW and JohnO are top level analysts here. They have both projected a CON maj 100. But that won't happen because again it is dependent on CON winning lots of seats which we haven't won since 1931. That won't happen.
BUT we are 10% clear. We may struggle in London and South although I am feeling more comfortable now. We will do ok in Midlands and North eg will win Wolverhampton SW, Barrow, Chester maybe Copeland. We will also win at least 10 seats in Scotland.
Ave it has always projected a 10% lead and around a 40 maj - it might be slightly better than that.
In English, normally abbreviated as The Conservative Party candidate
And yes I agree.
I don't think they have any councillors but hey they have celebrity links.
The Conservative Party - candidate.
On that note, night night all.
At the time I saw the increase in spend as a huge weakness, a sign of panic even. But the second terrorist attack has turned the conversation in the final week heavily towards security and this has worked out very well for the Tories.
Last week I had a lot of people telling me that attack ad was unfair and it made them feel sorry for Corbyn, this week if it's been talked about at all it has been 'oh, maybe the Tories have a point after all'.
YouGov may be herding but I do think there has been substantial swingback to the Tories after the London terrorist attack and the 'enough is enough' line has definitely cut through.