I never realised Blair had a lower rating that Howard in 2005. And in fact a lower rating than Johnson has now. Blair got a 66 seat majority despite this. I dont think Johnson will beat that.
The Tories, of course, won more votes in England than Labour in 2005.
Labour's 2005 vote was extremely efficient. Raw votes + vote efficiency is the key to any election.
I got bored and have coded up ability to scrape a load of data from the t'interweb, so now can automatically create spreadsheets containing constituency timings, brexit %, vote tallies at last GE. I also have demographic data from the census broken down for each seat.
I currently have a filtered output with Lab held seats that voted for Brexit, ordered by when they will announce so people can follow along. Thinking being if the Tories are going to win a majority we should be seeing them doing well in those constituency, especially Northern / Midland ones.
If people have any thoughts / ideas of what else would be useful, let me know and I will see if I can host them somewhere for election night. Or if somebody wants to put in some excel formulas so it will show swings etc (as I don't really do excel).
What I think would be extremely useful would be to start from the predictions of the YouGov MRP model, and then, as the first results start coming in, compare the YouGov predicted vote shares for each constituency against the actuals. You could then use that to calculate a live update to the YouGov prediction (with error bars).
I've been thinking about the statistics involved in this - obviously you'd have to build some kind of measure as to whether the deviations from the model's predicted vote shares in the early results were statistically significant, but if for example you had the first 10 results and in all of them Labour were doing a couple of percent better than YouGov expected, then you could probably predict with some degree of confidence that the final result would be better for Labour than the starting prediction, and estimate the delta.
As a refinement you could take account of the region and other characteristics of the seat - for example, your first ten seats might be mainly northern safe Labour seats which wouldn't be typical, so ideally you'd want to increase your level of uncertainty accordingly.
We have some expert statisticians here who I'm sure could advise, but I think this might be a very good way to get an early take on any deviation between the exit poll and the final tallies.
Liking this idea....I now have the MRP data into my python setup....
If any of the forecasters on here feel like putting their code up on github, that would be terrific.
Some interesting takes from Ashcroft focus groups:
Perhaps surprisingly, some of our 2017 Labour leave voters had a higher opinion of Boris Johnson than their Conservative counterparts: “He’s actually stood up to people. He won’t be bullied;” “He’s quite ruthless and bloody-minded. If he says he’s going to do it he’ll do it, regardless of the consequences, which is what I think you need;” “For me, the election is about getting the Tories in. I believe Boris will get the deal through. At the end of the day, we’ve had a referendum and the result was the result.” Some still had their doubts, though: “You’d like to think he’d get it done, but you just don’t know, do you? He said we’d come out on 31 October, it would definitely happen.”
Many of those leaning towards the Tories were helped along by their view of the Labour leader: “I think he’s a fantasist. He comes out with things that are just not plausible, that you would never believe;” “We’ve all voted Labour before, but can you imagine him and Diane Abbott running the country?” “He won’t even say what they’re going to campaign for. How can you vote for a party that won’t give you their view of what they’re going to do in the future?” “At first I did like him, but over time I think he’s the same as everybody else.”
Some fantastic quotes in there as usual. “Free broadband. Great, but what about healthcare? At least you can look up what you’re dying of.”
I don’t understand why the next Labour government would give me free broadband yet charge me water rates for the drinking water I get from my tap. Surely clean water is the more basic human right?
I never realised Blair had a lower rating that Howard in 2005. And in fact a lower rating than Johnson has now. Blair got a 66 seat majority despite this. I dont think Johnson will beat that.
The Tories, of course, won more votes in England than Labour in 2005.
Labour's 2005 vote was extremely efficient. Raw votes + vote efficiency is the key to any election.
Yes, 35% of the vote, a 3% lead, and a 66 seat majority.
I got bored and have coded up ability to scrape a load of data from the t'interweb, so now can automatically create spreadsheets containing constituency timings, brexit %, vote tallies at last GE. I also have demographic data from the census broken down for each seat.
I currently have a filtered output with Lab held seats that voted for Brexit, ordered by when they will announce so people can follow along. Thinking being if the Tories are going to win a majority we should be seeing them doing well in those constituency, especially Northern / Midland ones.
If people have any thoughts / ideas of what else would be useful, let me know and I will see if I can host them somewhere for election night. Or if somebody wants to put in some excel formulas so it will show swings etc (as I don't really do excel).
What I think would be extremely useful would be to start from the predictions of the YouGov MRP model, and then, as the first results start coming in, compare the YouGov predicted vote shares for each constituency against the actuals. You could then use that to calculate a live update to the YouGov prediction (with error bars).
I've been thinking about the statistics involved in this - obviously you'd have to build some kind of measure as to whether the deviations from the model's predicted vote shares in the early results were statistically significant, but if for example you had the first 10 results and in all of them Labour were doing a couple of percent better than YouGov expected, then you could probably predict with some degree of confidence that the final result would be better for Labour than the starting prediction, and estimate the delta.
As a refinement you could take account of the region and other characteristics of the seat - for example, your first ten seats might be mainly northern safe Labour seats which wouldn't be typical, so ideally you'd want to increase your level of uncertainty accordingly.
We have some expert statisticians here who I'm sure could advise, but I think this might be a very good way to get an early take on any deviation between the exit poll and the final tallies.
Liking this idea....I now have the MRP data into my python setup....
If any of the forecasters on here feel like putting their code up on github, that would be terrific.
Although I was thinking about scraping automatically the results, It might be easier to have a google doc spreadsheet and PBers can put the results in...and somebody with decent excel type knowledge have the formulas setup to show the differences to the MRP model etc.
Peston needs to work on his vocabulary and stop uttering the largely meaningless adjective "extraordinary". He sounds like somebody who keeps going "Wow" all the time.
Based on data from 1979 the Mori net leadership satisfaction data point to a Tory lead of 8.8pp rather than the 12pp of their headline poll. In the ten previous elections in the sample the actual result was in line with the prediction from the leadership ratings (within about 1pp) 7 times. Of the other two elections, the result was in the same direction as the headline poll (ie better for the government in this case) twice (1979 and 2005) and in the opposite direction once (1992). A similar error to 1992 would put us at around 6pp, hung parliament territory.
I was hoping someone had done this. Thanks. So that's a 1-in-10 chance of the Tory vote share lead being held down to 6pp, which might be a hung Parliament if the SNP/Lib Dems can win enough seats from the Tories.
That would make the 4/11 from Ladbrokes on a Conservative majority the value bet, wouldn't it?
For International Women's Day someone highlighted Empress Irene. I politely reminded the person that she'd had her own son so brutally mutilated he died of his wounds. At this point the individual blocked me, deleted the original tweet so my reply would be very difficult to find then posted a new one claiming they'd made a 'grammatical error'.
It doesn't matter whether Boris said talent or colour. Nor does it matter that C4 has apologised. People will believe what they want to believe irrespective of the truth.
Here’s a parallel universe scenario for you. At the weekend the Labour and LibDem (/Remain Alliance) parties jointly announce that the risks of giving Bozo a majority for the next five years are so great that they have agreed to which party in each seat voters should give their support. And posted the recommendations on all of their websites. How much difference if any do we think this would make?
For Labour I’d have thought there isn’t really any downside (yet I am sure they would oppose it the most vigorously). For the LibDems there is the risk that being tainted with Corbynism would lose them some Tory remainers, whilst maximising the chance of pulling Labour voters across in key seats like Esher against Raab.
At this stage in the cycle, serious risk of enraging lots of committed voters into backing the Tories. Earlier in the cycle, would have been less damaging as more time for people to calm down. Longer term, terminal damage to the Lib Dems; less so to Labour, as you say.
Short term: meh. Might swing ~20-30 seats, possibly but not definitely enough to deprive the Tories of a majority. Might plausibly be a net positive for the Tories, seeing as most voters aren't strongly affiliated enough to a party to care what they think you should do (or are so affiliated that they'll vote for them even if told not to).
It's a trick that only works once, though. I wouldn't want to be trying to defend that decision as a PPC in 2024.
Yes, you are mostly right. The problem for the centre-left is that so long as Labour remains burrowed down its Corbynite rabbit hole, we are destined to live under Bozo and his ilk for the foreseeable. Labour needs to do us a favour and either return to sanity or spiral off into further insanity to clear the decks for an electable centre party alternative.
I never realised Blair had a lower rating that Howard in 2005. And in fact a lower rating than Johnson has now. Blair got a 66 seat majority despite this. I dont think Johnson will beat that.
The Tories, of course, won more votes in England than Labour in 2005.
Labour's 2005 vote was extremely efficient. Raw votes + vote efficiency is the key to any election.
That's an excellent point - Labour won 41 Scottish seats in 2005, wth just 1 predicted now. That performance would have put Labour on 314...
Note that Boris Johnson is by far the lowest ranked of all the higher of the pairs. There are huge numbers of voters hostile to both. How they resolve their distaste for both is going to decide this election.
Unfortunately, despite 'other alternatives are available', its likely they'd just hold their nose and vote for the other one.
And therefore, Johnson will get his majority.
I was thinking a few weeks ago about the start of 2017. I think a few posters on here remarked how campaigns very rarely changed polling number and broadly where you started was where you ended up. Of course, 2017 then became the exception that proved the rule.
And I do wonder if thats were we are here. Polling gap started about 10% four/five weeks ago, and despite a squeeze for the LD and BXP, that gap has remained steady. Short of a miracle of tactical voting, I can't see anything other than a Con majority next Thursday.
Some interesting takes from Ashcroft focus groups:
Perhaps surprisingly, some of our 2017 Labour leave voters had a higher opinion of Boris Johnson than their Conservative counterparts: “He’s actually stood up to people. He won’t be bullied;” “He’s quite ruthless and bloody-minded. If he says he’s going to do it he’ll do it, regardless of the consequences, which is what I think you need;” “For me, the election is about getting the Tories in. I believe Boris will get the deal through. At the end of the day, we’ve had a referendum and the result was the result.” Some still had their doubts, though: “You’d like to think he’d get it done, but you just don’t know, do you? He said we’d come out on 31 October, it would definitely happen.”
Many of those leaning towards the Tories were helped along by their view of the Labour leader: “I think he’s a fantasist. He comes out with things that are just not plausible, that you would never believe;” “We’ve all voted Labour before, but can you imagine him and Diane Abbott running the country?” “He won’t even say what they’re going to campaign for. How can you vote for a party that won’t give you their view of what they’re going to do in the future?” “At first I did like him, but over time I think he’s the same as everybody else.”
Some fantastic quotes in there as usual. “Free broadband. Great, but what about healthcare? At least you can look up what you’re dying of.”
I don’t understand why the next Labour government would give me free broadband yet charge me water rates for the drinking water I get from my tap. Surely clean water is the more basic human right?
Yes but not so popular with the yoof vote that so love Jezza, apparently
Wow looks like I missed an entire scandal's life cycle in the space of a lunch break.
Fwiw I just listened to the clip and it's definitely talent, you can hear the t at the end. In any case it's too unsubtle for Boris to say something like that, indeed when I first saw someone quote the phrase I presumed it was said by Pat Mountain or equivalent.
There are multiple clips though. In this one it certainly does sound like talent. In the other one that was posted, it's definitely colour, with no hint of a 't' at the start or end. It's all very strange.
My sense is a Tory landslide is fading into the distance at the moment. Most likely result is a majority of between 20 and 60 but they could still fall short of 326 if the last week goes badly.
I'm thinking exactly the opposite. All the anecdotes point to a larger Tory victory. The last person to enthuse the WWC and the lower middle class was Blair. It could be 1983, 1997 or even 1931, although I think the different demoses in Scotland and NI preclude the latter.
I think that the plural of demos is demoi
I did O Level Greek. It is demoi. Similar to logos.
Wow looks like I missed an entire scandal's life cycle in the space of a lunch break.
Fwiw I just listened to the clip and it's definitely talent, you can hear the t at the end. In any case it's too unsubtle for Boris to say something like that, indeed when I first saw someone quote the phrase I presumed it was said by Pat Mountain or equivalent.
There are multiple clips though. In this one it certainly does sound like talent. In the other one that was posted, it's definitely colour, with no hint of a 't' at the start or end. It's all very strange.
Can't you see that's because the audio quality was much worse on that version? The audio wasn't coming directly from the microphone, and there were lots of sounds from the audience like the sound of cameras.
What a fucking stupid thing to tweet, It's so bad and silly I must check my prejudices: is it really real? Did he really tweet that?
It's nuts. Merkel seems to be giving him a bit of an incredulous glance too. You've got to be trying really hard to be in a picture with Trump and come out looking like the bad guy!
Note that Boris Johnson is by far the lowest ranked of all the higher of the pairs. There are huge numbers of voters hostile to both. How they resolve their distaste for both is going to decide this election.
Unfortunately, despite 'other alternatives are available', its likely they'd just hold their nose and vote for the other one.
And therefore, Johnson will get his majority.
You're assuming that Boris Johnson is the preferred choice of those who dislike both.
Note that Boris Johnson is by far the lowest ranked of all the higher of the pairs. There are huge numbers of voters hostile to both. How they resolve their distaste for both is going to decide this election.
Unfortunately, despite 'other alternatives are available', its likely they'd just hold their nose and vote for the other one.
And therefore, Johnson will get his majority.
I was thinking a few weeks ago about the start of 2017. I think a few posters on here remarked how campaigns very rarely changed polling number and broadly where you started was where you ended up. Of course, 2017 then became the exception that proved the rule.
And I do wonder if thats were we are here. Polling gap started about 10% four/five weeks ago, and despite a squeeze for the LD and BXP, that gap has remained steady. Short of a miracle of tactical voting, I can't see anything other than a Con majority next Thursday.
This is why the Tories framing the election as Boris vs. Corbyn was actually a smart strategic move. When your leader is unpopular, but the main opponent is even more so, you want to cut off the oxygen from all the other alternatives and consolidate your vote with a forced choice. Labour has played right along with that strategy...
I have updated my England only seats model based on 2017 party vote splits for Leave and Remain to allow for more extensive Remain tactical voting than hitherto, based on the evidence from constituency polling that Conservative voters in Remain leaning seats seem willing to switch to the LDs in contests where Labour is clearly out of contention. It also allows for more limited TV between Lab and LD Remainers. It now generates exactly the same number of Con seats (336) as the YouGov MRP for the same polling numbers (11.1% Con lead).
Key TV assumptions (calibrated to reconcile to the GB averages used in the YouGov MRP) in seats where the LDs are standing: 2017 Con Remainer vote: Seats where LDs not Labour are challenging split 47 Con, 41 LD, in other seats 70 Con 18 Lab. (GB average 67 Con, 21 LD). 2017 Lab Remainer vote: Seats where LDs not Labour are challenging or where Con and Lab are more than 10k apart without TV split Lab 70 LD 21, in other seats Lab 86 LD 4. (GB average Lab 77 LD 13). 2017 LD Remainer vote: Seats where LDs were 1st/2nd or close to Lab without TV split LD 83 Lab 7, other seats LD 63 Lab 27. (GB average LD 73 Lab 17).
In addition to the above, in seats where the Greens are expected to poll 2,500+ without TV and which are not Con-Lab marginals (in many of which the LDs are standing down): 2017 Lab Remainer vote: add 10% to Green from Lab, otherwise add 1.5% to Lab from Green. 2017 Green Remainer vote: add 10% to Green from Lab, otherwise add 4.2% to Lab from Green.
Results: Using YouGov MRP poll share (42.9/31.8/13.6/3.1BXP/3.2G) Con 336 (+39), Lab 184(-43), LD 12(+4), G 1.
Scaling proportionately to 2/3 Dec YouGov share (42/33/12/4/4) Con 314, Lab 211, LD 7, G 1. (Note that the "other seats" TV adjustment has to be varied slightly to maintain the same national vote shares.)
The critical figure needed to give the Conservatives an overall majority of 326 is in the region of 308, depending on assumptions made for Scotland/Wales.
Note that Boris Johnson is by far the lowest ranked of all the higher of the pairs. There are huge numbers of voters hostile to both. How they resolve their distaste for both is going to decide this election.
Unfortunately, despite 'other alternatives are available', its likely they'd just hold their nose and vote for the other one.
And therefore, Johnson will get his majority.
I was thinking a few weeks ago about the start of 2017. I think a few posters on here remarked how campaigns very rarely changed polling number and broadly where you started was where you ended up. Of course, 2017 then became the exception that proved the rule.
And I do wonder if thats were we are here. Polling gap started about 10% four/five weeks ago, and despite a squeeze for the LD and BXP, that gap has remained steady. Short of a miracle of tactical voting, I can't see anything other than a Con majority next Thursday.
This is why the Tories framing the election as Boris vs. Corbyn was actually a smart strategic move. When your leader is unpopular, but the main opponent is even more so, you want to cut off the oxygen from all the other alternatives and consolidate your vote with a forced choice. Labour has played right along with that strategy...
Probably true. Labour simply hoped that the story of 2017 would repeat itself.
"leader ratings are a better guide to election outcomes than voting intention numbers."
I think as a general assertion that needs to be questioned.
The graph above shows Howard with a lead of about 15 points over Blair in 2005. Were the voting intention opinion polls proecting a Tory victory then?
And in 2010 the graph shows Cameron with nearly a 30-point lead over Brown. The result was a hung parliament, and I think the voting intention polls were a better guide then too.
I have updated my England only seats model based on 2017 party vote splits for Leave and Remain to allow for more extensive Remain tactical voting than hitherto, based on the evidence from constituency polling that Conservative voters in Remain leaning seats seem willing to switch to the LDs in contests where Labour is clearly out of contention. It also allows for more limited TV between Lab and LD Remainers. It now generates exactly the same number of Con seats (336) as the YouGov MRP for the same polling numbers (11.1% Con lead).
Key TV assumptions (calibrated to reconcile to the GB averages used in the YouGov MRP) in seats where the LDs are standing: 2017 Con Remainer vote: Seats where LDs not Labour are challenging split 47 Con, 41 LD, in other seats 70 Con 18 Lab. (GB average 67 Con, 21 LD). 2017 Lab Remainer vote: Seats where LDs not Labour are challenging or where Con and Lab are more than 10k apart without TV split Lab 70 LD 21, in other seats Lab 86 LD 4. (GB average Lab 77 LD 13). 2017 LD Remainer vote: Seats where LDs were 1st/2nd or close to Lab without TV split LD 83 Lab 7, other seats LD 63 Lab 27. (GB average LD 73 Lab 17).
In addition to the above, in seats where the Greens are expected to poll 2,500+ without TV and which are not Con-Lab marginals (in many of which the LDs are standing down): 2017 Lab Remainer vote: add 10% to Green from Lab, otherwise add 1.5% to Lab from Green. 2017 Green Remainer vote: add 10% to Green from Lab, otherwise add 4.2% to Lab from Green.
Results: Using YouGov MRP poll share (42.9/31.8/13.6/3.1BXP/3.2G) Con 336 (+39), Lab 184(-43), LD 12(+4), G 1.
Scaling proportionately to 2/3 Dec YouGov share (42/33/12/4/4) Con 314, Lab 211, LD 7, G 1. (Note that the "other seats" TV adjustment has to be varied slightly to maintain the same national vote shares.)
The critical figure needed to give the Conservatives an overall majority of 326 is in the region of 308, depending on assumptions made for Scotland/Wales.
Edit: "2017 Con Remainer vote: Seats where LDs not Labour are challenging split 47 Con, 41 LD, in other seats 70 Con 18 Lab....." should read "70 Con 18 LD".
Based on data from 1979 the Mori net leadership satisfaction data point to a Tory lead of 8.8pp rather than the 12pp of their headline poll. In the ten previous elections in the sample the actual result was in line with the prediction from the leadership ratings (within about 1pp) 7 times. Of the other two elections, the result was in the same direction as the headline poll (ie better for the government in this case) twice (1979 and 2005) and in the opposite direction once (1992). A similar error to 1992 would put us at around 6pp, hung parliament territory.
I was hoping someone had done this. Thanks. So that's a 1-in-10 chance of the Tory vote share lead being held down to 6pp, which might be a hung Parliament if the SNP/Lib Dems can win enough seats from the Tories.
That would make the 4/11 from Ladbrokes on a Conservative majority the value bet, wouldn't it?
DYOR but yes I would say so. Putting together everything we know I would say the chance of a Tory majority is probably about 85%. One thing I wonder is whether the AS stuff may have artificially depressed Corbyn's net favorability - people have heard all this bad stuff about Corbyn and know they are meant to disapprove but perhaps aren't as fussed by it as they make out. But perhaps the same holds for Johnson and some of the bad vibes attached to him. The main forecasting error for the model was 2005, when Blair won despite being 15pp behind Howard. Maybe people knew they were meant to disapprove of Blair because of Iraq, and do marked him down in the survey, but ultimately didn't see it as that big an issue. Just a thought. Overall it has excellent predictive power.
Interesting. I think this will actually resonate more with older Tory remainers than most younger voters, as he's largely unknown to them except as someone carrying the heft of the ex-prime ministership.
On the other hand, there is a recent trend for the old to appeal to the young, and Heseltine, Major and Clarke have all become iconic for the most committed young remainers.
Yes, they are shoring up their power base for the inevitable leadership election. Piddock is a bit overpriced in my view. I think she will retain her seat OK.
"leader ratings are a better guide to election outcomes than voting intention numbers."
I think as a general assertion that needs to be questioned.
The graph above shows Howard with a lead of about 15 points over Blair in 2005. Were the voting intention opinion polls proecting a Tory victory then?
And in 2010 the graph shows Cameron with nearly a 30-point lead over Brown. The result was a hung parliament, and I think the voting intention polls were a better guide then too.
The leadership-result link doesn't work too well with seats, but with % of the vote it's a pretty tidy relationship. Cameron 2010 ended up with about a 7% lead, and Blair2005 somehow scraped through with sub3%, neither of which are too far from the trendline.
Wow looks like I missed an entire scandal's life cycle in the space of a lunch break.
Fwiw I just listened to the clip and it's definitely talent, you can hear the t at the end. In any case it's too unsubtle for Boris to say something like that, indeed when I first saw someone quote the phrase I presumed it was said by Pat Mountain or equivalent.
There are multiple clips though. In this one it certainly does sound like talent. In the other one that was posted, it's definitely colour, with no hint of a 't' at the start or end. It's all very strange.
Can't you see that's because the audio quality was much worse on that version? The audio wasn't coming directly from the microphone, and there were lots of sounds from the audience like the sound of cameras.
The quality of the audio in the other clip isn't that bad. I'm not sure what to believe. After the BBC apparently covered up Boris's wreath-laying faux pas, it doesn't seem such a stretch of the imagination to suspect them of a little bit of audio manipulation.
Yes, they are shoring up their power base for the inevitable leadership election. Piddock is a bit overpriced in my view. I think she will retain her seat OK.
My understanding is Boris has been using this phrase "people of talent" regularly on the campaign trail.
If only there were more people of talent on offer for us to vote for?
If only there were some / any people of talent on offer for us to vote for?
FTFY - but we really have a choice of idiots...
A bit unfair. There is the very occasional person of talent(!) in all parties. It is just unfortunate that they tend to be consigned to the outer fringes of the parties. It is the leadership of both main parties which is as weak as I recall. Whilst we may have disagreed with Major, Blair and Ashdown on many matter, they all look like Churchill next to the current clowns.
A few things to take from this People of Colour/Talent episode.
1) Twitter has destroyed journalism in this country and it is very close to destroying the very essence of truth. Journalists simply do not operate within editorial boundaries on Twitter. Channel 4 need severely reprimanding for their role in this affair as an example to all.
2) A great many people got exposed as being part of the hive-mind of Twitter. In one respect this episode is a salutary lesson for them but will it be heeded? Doubtful.
3) By accident, this episode reveals the dangers in the next few years of the deep fake videos. If you have the hive-mind of Twitter searching for a gotcha moment and you have someone else ready to supply it, then its clear we are about to enter a new era where truth is destroyed. I'm sure our enemies are rubbing their hands with glee.
Very fair, but in this case the thinking (and hearing) was conditioned by Boris' past record of far worse casual racism. I suspect there would have been much more investigation had he not had such a record.
Liking this idea....I now have the MRP data into my python setup....
Would be good to compare the accuracy of the initial results against the accuracy of 2017. If the 2019 results are shwoing similar defelctions as 2017 (i.e. undercounting Labour by the same amountin the same constituency) then you could dynamically recompute the MRP results to take account of the error
I already calculated a RMS error for 2017 for the big 4 parties.
Here’s a parallel universe scenario for you. At the weekend the Labour and LibDem (/Remain Alliance) parties jointly announce that the risks of giving Bozo a majority for the next five years are so great that they have agreed to which party in each seat voters should give their support. And posted the recommendations on all of their websites. How much difference if any do we think this would make?
For Labour I’d have thought there isn’t really any downside (yet I am sure they would oppose it the most vigorously). For the LibDems there is the risk that being tainted with Corbynism would lose them some Tory remainers, whilst maximising the chance of pulling Labour voters across in key seats like Esher against Raab.
Unnecessary for the LDs in seats like Esher and Wimbledon in the SE. The Labour vote in those seats has melted and it is only a case of whether it is enough to secure a LD win. Wimbledon possibly, Guildford probably, Esher unlikely. The remaining Labour voters are purely tribal and will not shift.
I have updated my England only seats model based on 2017 party vote splits for Leave and Remain to allow for more extensive Remain tactical voting than hitherto, based on the evidence from constituency polling that Conservative voters in Remain leaning seats seem willing to switch to the LDs in contests where Labour is clearly out of contention. It also allows for more limited TV between Lab and LD Remainers. It now generates exactly the same number of Con seats (336) as the YouGov MRP for the same polling numbers (11.1% Con lead).
Key TV assumptions (calibrated to reconcile to the GB averages used in the YouGov MRP) in seats where the LDs are standing: 2017 Con Remainer vote: Seats where LDs not Labour are challenging split 47 Con, 41 LD, in other seats 70 Con 18 Lab. (GB average 67 Con, 21 LD). 2017 Lab Remainer vote: Seats where LDs not Labour are challenging or where Con and Lab are more than 10k apart without TV split Lab 70 LD 21, in other seats Lab 86 LD 4. (GB average Lab 77 LD 13). 2017 LD Remainer vote: Seats where LDs were 1st/2nd or close to Lab without TV split LD 83 Lab 7, other seats LD 63 Lab 27. (GB average LD 73 Lab 17).
In addition to the above, in seats where the Greens are expected to poll 2,500+ without TV and which are not Con-Lab marginals (in many of which the LDs are standing down): 2017 Lab Remainer vote: add 10% to Green from Lab, otherwise add 1.5% to Lab from Green. 2017 Green Remainer vote: add 10% to Green from Lab, otherwise add 4.2% to Lab from Green.
Results: Using YouGov MRP poll share (42.9/31.8/13.6/3.1BXP/3.2G) Con 336 (+39), Lab 184(-43), LD 12(+4), G 1.
Scaling proportionately to 2/3 Dec YouGov share (42/33/12/4/4) Con 314, Lab 211, LD 7, G 1. (Note that the "other seats" TV adjustment has to be varied slightly to maintain the same national vote shares.)
The critical figure needed to give the Conservatives an overall majority of 326 is in the region of 308, depending on assumptions made for Scotland/Wales.
Edit: "2017 Con Remainer vote: Seats where LDs not Labour are challenging split 47 Con, 41 LD, in other seats 70 Con 18 Lab....." should read "70 Con 18 LD".
Isn't that 47/41 split potentially very bad for Con in the sense it would imply LD are at least within touching distance in a lot of Con / LD marginals - in turn implying they might well win quite a few.
The chart is showing net figures. If you just take the positive ones then the GE2005 anomaly doesn't exit. Blair had a higher satisfaction level than Howard ahead
We are though in outlier teritory, because not only is Corbyn there with the lowest of the lows but Johnson is the lowest of the highs. Also Howard had a 15+ point lead over Blair but lost heavily, which shows us that if there is any predictive power in this chart there is also a lot of variability.
Theory: more women tend to want to spend squillions and squillions of quid on public services, benefits and just about everything else. More men tend to worry about where all the money is meant to come from.
*****
The Ipsos headline VI snaps quite neatly into line with the trend of all the other recent surveys, doesn't it? The mean Tory lead in all nationwide polls with fieldwork completed November 25th or later now stands at 9.75%, there's still no sign of any of the parties moving significantly up or down, and there are now only five-and-a-half campaigning days left to go.
YouGov's new Scotland-only numbers aren't too bad for the Conservatives either.
So the real question, following this and the other week, is how much money will the Government make from privatising Channel Four, and who gets the related bung? Might find pensioner’s TV licences for a few years....
Happens in all parties. The Tories in Guildford have a man, Dennis Paul, as a leading figure and ex-councillor, who ran a libellous campaign against the blogger Tim Ireland, falsely accusing him of paedophilia.
I'm happy to defer to those who have listened to the better quality audio.
Listen to it here
It is unusual (but not gramatically wrong) to say "people of talent" rather than "talented people".
No it isn't. In fact it's a far more common phrase than 'people of colour', a recent import from the US and mostly used by the Guardian-reading classes.
Neither is exactly admired by any criterion, but the most striking contrast is the Good leader/Bad leader rating:
Boris 36% Good leader, 41% Bad leader Corbyn 16% Good leader, 57% Bad leader
Mmm, but Corbyn is ahead (in the sense of less bad) on most of the other criteria. Classically Tories lead on "competence/leadership" and Labour leaders win on "empathy/understands people like me" (CorbynJohnson -19/-30). I think Johnson does convey an impression of dynamism which Corbyn's judicious replies lack, but people mostly don't actually like what Johnson says.
As in other polls, the gap for gender and age is gigantic - we're used to age correlating with conservatism, but I've never seen an age split quite like it.
Theory: more women tend to want to spend squillions and squillions of quid on public services, benefits and just about everything else. More men tend to worry about where all the money is meant to come from.
This is view on society that died out in the 80's. I'm guessing you are in the second of your two categories.
Only because the previous Ipsos was conducted so long ago. The polls have been broadly consistent for nearly two weeks and show no clear evidence of movement in favour of any party over that period.
It doesn't necessarily follow that Labour won't tighten the gap again over the remaining days of the campaign (or that the polls are right to begin with,) but there's plenty of evidence to suggest a reasonably stable Tory lead of around 10% at the moment.
A few things to take from this People of Colour/Talent episode.
1) Twitter has destroyed journalism in this country and it is very close to destroying the very essence of truth. Journalists simply do not operate within editorial boundaries on Twitter. Channel 4 need severely reprimanding for their role in this affair as an example to all.
2) A great many people got exposed as being part of the hive-mind of Twitter. In one respect this episode is a salutary lesson for them but will it be heeded? Doubtful.
3) By accident, this episode reveals the dangers in the next few years of the deep fake videos. If you have the hive-mind of Twitter searching for a gotcha moment and you have someone else ready to supply it, then its clear we are about to enter a new era where truth is destroyed. I'm sure our enemies are rubbing their hands with glee.
Very fair, but in this case the thinking (and hearing) was conditioned by Boris' past record of far worse casual racism. I suspect there would have been much more investigation had he not had such a record.
The chart is showing net figures. If you just take the positive ones then the GE2005 anomaly doesn't exit. Blair had a higher satisfaction level than Howard ahead
So is it net satisfaction, or positive satisfaction which predicts outcome? And if it is the second why did you publish net satisfaction?
"leader ratings are a better guide to election outcomes than voting intention numbers."
I think as a general assertion that needs to be questioned.
The graph above shows Howard with a lead of about 15 points over Blair in 2005. Were the voting intention opinion polls proecting a Tory victory then?
And in 2010 the graph shows Cameron with nearly a 30-point lead over Brown. The result was a hung parliament, and I think the voting intention polls were a better guide then too.
The leadership-result link doesn't work too well with seats, but with % of the vote it's a pretty tidy relationship. Cameron 2010 ended up with about a 7% lead, and Blair2005 somehow scraped through with sub3%, neither of which are too far from the trendline.
I think it's important to look at seat movements from election to election as well as vote share outcomes. In 2005, Howard's small-ish lead helped achieve an 80 seat swing against Blair, but because of the low base from 1997/2001, this didn't help much. Cameron had a bigger lead in 2010 (and a higher personal figure) and got a ~200 seat swing. May and Major had leads too small to matter much.
Unclear as yet whether Johnson's Cameron-sized lead matters, given how unpopular he is personally.
It doesn't matter whether Boris said talent or colour. Nor does it matter that C4 has apologised. People will believe what they want to believe irrespective of the truth.
The truth matters. It wont effect what people believe, but there will be a fact about what he said.
Neither is exactly admired by any criterion, but the most striking contrast is the Good leader/Bad leader rating:
Boris 36% Good leader, 41% Bad leader Corbyn 16% Good leader, 57% Bad leader
Mmm, but Corbyn is ahead (in the sense of less bad) on most of the other criteria. Classically Tories lead on "competence/leadership" and Labour leaders win on "empathy/understands people like me" (CorbynJohnson -19/-30). I think Johnson does convey an impression of dynamism which Corbyn's judicious replies lack, but people mostly don't actually like what Johnson says.
As in other polls, the gap for gender and age is gigantic - we're used to age correlating with conservatism, but I've never seen an age split quite like it.
We have never had a country where the retired own such a big share of the capital and the workers own so little. So its still just those with capital vote Tory and those without vote Labour seen along age rather than class.
Theory: more women tend to want to spend squillions and squillions of quid on public services, benefits and just about everything else. More men tend to worry about where all the money is meant to come from.
This is view on society that died out in the 80's. I'm guessing you are in the second of your two categories.
You may be right. I'm speculating. Could always be that women are breaking against Johnson so much more heavily because a lot of them (the younger ones especially) quite simply regard him as a sexist pig?
In which case, if the Tories had been able to find a leader with at least some presentational flair, yet without Johnson's baggage, Lord alone knows how far ahead they would be by now.
Note that Boris Johnson is by far the lowest ranked of all the higher of the pairs. There are huge numbers of voters hostile to both. How they resolve their distaste for both is going to decide this election.
Assuming the Conservatives win, I expect the next round of local elections will be torrid.
2019 will be the election the Conservatives wish they'd never won. There will be a recession in the next five years, and (rightly or wrongly) their handling of Brexit will get the blame.
We will also see house prices down meaningfully (15-20%) in nominal terms. The young won't thank the Conservatives, and the old will be livid.
Johnson will turn out to be just the man you don't want in charge for this. But having remodeled the Conservative Party in his own image, he'll prove difficult to shift.
If the Labour Party have chosen a moderate as their leader, this won't matter too much. But if they manage to pick Corbyn 2.0 (younger, fitter, less economically sane), then the country will lurch to the hard Left in 2024, and it will all be pretty horrible.
It's a good thing I don't own any overpriced London real estate... oh wait...
Comments
That would make the 4/11 from Ladbrokes on a Conservative majority the value bet, wouldn't it?
For International Women's Day someone highlighted Empress Irene. I politely reminded the person that she'd had her own son so brutally mutilated he died of his wounds. At this point the individual blocked me, deleted the original tweet so my reply would be very difficult to find then posted a new one claiming they'd made a 'grammatical error'.
Nor does it matter that C4 has apologised. People will believe what they want to believe irrespective of the truth.
And therefore, Johnson will get his majority.
I was thinking a few weeks ago about the start of 2017. I think a few posters on here remarked how campaigns very rarely changed polling number and broadly where you started was where you ended up. Of course, 2017 then became the exception that proved the rule.
And I do wonder if thats were we are here. Polling gap started about 10% four/five weeks ago, and despite a squeeze for the LD and BXP, that gap has remained steady.
Short of a miracle of tactical voting, I can't see anything other than a Con majority next Thursday.
Brown
Cameron
May
https://twitter.com/katyballs/status/1202944668482392064
Key TV assumptions (calibrated to reconcile to the GB averages used in the YouGov MRP) in seats where the LDs are standing:
2017 Con Remainer vote: Seats where LDs not Labour are challenging split 47 Con, 41 LD, in other seats 70 Con 18 Lab. (GB average 67 Con, 21 LD).
2017 Lab Remainer vote: Seats where LDs not Labour are challenging or where Con and Lab are more than 10k apart without TV split Lab 70 LD 21, in other seats Lab 86 LD 4. (GB average Lab 77 LD 13).
2017 LD Remainer vote: Seats where LDs were 1st/2nd or close to Lab without TV split LD 83 Lab 7, other seats LD 63 Lab 27. (GB average LD 73 Lab 17).
In addition to the above, in seats where the Greens are expected to poll 2,500+ without TV and which are not Con-Lab marginals (in many of which the LDs are standing down):
2017 Lab Remainer vote: add 10% to Green from Lab, otherwise add 1.5% to Lab from Green.
2017 Green Remainer vote: add 10% to Green from Lab, otherwise add 4.2% to Lab from Green.
Results:
Using YouGov MRP poll share (42.9/31.8/13.6/3.1BXP/3.2G) Con 336 (+39), Lab 184(-43), LD 12(+4), G 1.
Scaling proportionately to 2/3 Dec YouGov share (42/33/12/4/4) Con 314, Lab 211, LD 7, G 1. (Note that the "other seats" TV adjustment has to be varied slightly to maintain the same national vote shares.)
The critical figure needed to give the Conservatives an overall majority of 326 is in the region of 308, depending on assumptions made for Scotland/Wales.
https://twitter.com/DavidHerdson/status/1202934174782500865
I think as a general assertion that needs to be questioned.
The graph above shows Howard with a lead of about 15 points over Blair in 2005. Were the voting intention opinion polls proecting a Tory victory then?
And in 2010 the graph shows Cameron with nearly a 30-point lead over Brown. The result was a hung parliament, and I think the voting intention polls were a better guide then too.
Crass stupidity would be as charitable as I could be.
https://twitter.com/adamcooperF1/status/1202945427865292801
Unless I'm reading it wrong, Blair was less popular in 2005?
One thing I wonder is whether the AS stuff may have artificially depressed Corbyn's net favorability - people have heard all this bad stuff about Corbyn and know they are meant to disapprove but perhaps aren't as fussed by it as they make out. But perhaps the same holds for Johnson and some of the bad vibes attached to him. The main forecasting error for the model was 2005, when Blair won despite being 15pp behind Howard. Maybe people knew they were meant to disapprove of Blair because of Iraq, and do marked him down in the survey, but ultimately didn't see it as that big an issue. Just a thought. Overall it has excellent predictive power.
On the other hand, there is a recent trend for the old to appeal to the young, and Heseltine, Major and Clarke have all become iconic for the most committed young remainers.
The leadership-result link doesn't work too well with seats, but with % of the vote it's a pretty tidy relationship. Cameron 2010 ended up with about a 7% lead, and Blair2005 somehow scraped through with sub3%, neither of which are too far from the trendline.
https://twitter.com/darrengrimes_/status/1202921805859708928
https://yougov.co.uk/topics/politics/articles-reports/2019/12/06/corbyn-and-johnson-failing-impress
Neither is exactly admired by any criterion, but the most striking contrast is the Good leader/Bad leader rating:
Boris 36% Good leader, 41% Bad leader
Corbyn 16% Good leader, 57% Bad leader
It's getting ugly out there.
I already calculated a RMS error for 2017 for the big 4 parties.
( https://docs.google.com/spreadsheets/d/18rrY4Yj07ne1QlhFYI0JuEl82dEXn5K4EdChKc8MRMo )
*****
The Ipsos headline VI snaps quite neatly into line with the trend of all the other recent surveys, doesn't it? The mean Tory lead in all nationwide polls with fieldwork completed November 25th or later now stands at 9.75%, there's still no sign of any of the parties moving significantly up or down, and there are now only five-and-a-half campaigning days left to go.
YouGov's new Scotland-only numbers aren't too bad for the Conservatives either.
This is turning against the Conservatives ...
@NCPoliticsUK
·
5m
Election day weather forecasts from the @metoffice
:
London 9°, showers
Plymouth 11°, wet/sunny intervals
Nuneaton 8°, showers/sunny intervals
Manchester 7°, showers
Edinburgh 6°, cloudy
Inverness 6°, cloudy
Cardiff 10°, wet
Belfast 7°, cloudy
As in other polls, the gap for gender and age is gigantic - we're used to age correlating with conservatism, but I've never seen an age split quite like it.
It doesn't necessarily follow that Labour won't tighten the gap again over the remaining days of the campaign (or that the polls are right to begin with,) but there's plenty of evidence to suggest a reasonably stable Tory lead of around 10% at the moment.
burning hot take, pal. Burning.
Debates between Biden and Trump will be so so so stupid...
https://www.theguardian.com/politics/2015/apr/09/labour-moves-ahead-of-tories-on-the-day-the-polls-turned
Unclear as yet whether Johnson's Cameron-sized lead matters, given how unpopular he is personally.
In which case, if the Tories had been able to find a leader with at least some presentational flair, yet without Johnson's baggage, Lord alone knows how far ahead they would be by now.
We will also see house prices down meaningfully (15-20%) in nominal terms. The young won't thank the Conservatives, and the old will be livid.
Johnson will turn out to be just the man you don't want in charge for this. But having remodeled the Conservative Party in his own image, he'll prove difficult to shift.
If the Labour Party have chosen a moderate as their leader, this won't matter too much. But if they manage to pick Corbyn 2.0 (younger, fitter, less economically sane), then the country will lurch to the hard Left in 2024, and it will all be pretty horrible.
It's a good thing I don't own any overpriced London real estate... oh wait...