United Kingdom General Elections: December 12th, 2019

[jaichind]

The lastest yougov poll does it both ways (exclude BXP for voters where BXP have stood down AND just display BXP for everyone)

The result where one removes BXP where BXP is not standing gives

CON    42
LAB     28
LDEM  15
BXP      4
Green   4

The one that have BXP everywhere

CON   39
LAB    26
LDM   16
BXP     9

https://d25d2506sfb94s.cloudfront.net/cumulus_uploads/document/w3ohbvr6zt/Sky_TheTimes_VI_191112_w.pdf

By looking at 2016 Leave vs Remain votes it seems most BXP voters when not given BXP as a choice mostly went CON or Other (BXP rebels?).  For 2016 Remain voters by not having BXP as an option actually shifted some LDEM voters over to LAB (tactical voting now that they see BXP is not there to split the CON vote?)

Anyway this poll seems to show that in Southern seats (where CON mostly won in 2017) the BXP voter are mostly CON voters.  I suspect in the North the BXP voter are much more likely to be LAB Leave voters so BXP running there does take LAB votes that would otherwise go back to LAB if BXP is not running.

[jaichind]

Not this **** again.

a) the 'social grade' system was faulty even in the 1970s and is a complete disaster now; it in no way reflects the realities of a service sector economy with a very large elderly population (it is true, by the way, pensioners are routinely rolled into category DE by some polling firms).* I can go into far too much detail about this if anyone is interested: I have done before. But for now: how many people in Great Britain consider, for example, nursing to be a middle class occupation? I suspect not many. I wonder how many people (in the countryside, where this is relevant) consider farmers to be working class? No one.

b) even were this not so, YouGov's peculiar polling methods happen to make their internal numbers completely worthless anyway. It's some real voodoo nonsense. Rubbish.

c) even were this not so, different polling firms internals show very different figures and patterns, not just from YouGov but from each other. In fact the main thing that shows up when one monitors numbers for these categories over even a short period of time (say, a month) is how absurdly volatile they are. Which is not surprising as they are almost random categories by this point.

d) this actually takes us back to a) because there are serious issues with sampling for some of the categories, particularly C2 which is a notorious disaster.

e) even were all of this not so, poll internals are not polls or surveys themselves, but a way of making sure that the poll was conducted with an appropriately balanced sample. The way they are thrown around by people who know this but have commercial reasons to ignore that fact amounts to the pollution of public discourse.

f) you wouldn't give your bank details to a 'Nigerian Prince', you don't believe anything hawked by Matthew Badwin.

_{*And there are reasons to be a mildly dubious as to how thorough the ones that say they do not do this actually are.}

This is most likely too America-centric but why cannot the polls instead group the population by education (grad school, university, vocational, and high school etc etc)?

[jaichind]

https://twitter.com/britainelects/status/1195357004337840128

Continued proof that Brexit was just the holding pen for future tories.

I wonder what their methodology is?  Namely did they just ask for party support or excluded BXP as a choice for someone from a seat where BXP is not running.  If former then this result ia significant. If the latter it seems to be a statistical shift as long as we accept that most Southern BXP supports are mostly former CON voters.

[jaichind]

One also apparently has a massive (and unexplained) rise in Johnson's personal rating.

Was this Deltapoll? If so, when combined with the obviously odd business of LibDem support dropping by a third in a week and transferring all to the Tories, we can say 'bad sample' and draw a line through it.

Edit: O.K. I've seen their uploaded tables and... this is a bad poll, I'll just leave it at that.

What makes it bad ? Is it because it shows almost a third of 2017 LDEM voters going to CON ?  I agree that looks fishy.

[jaichind]

I am surprised Theresa May is running again.  Last few PMs seems to have retired from politics after they left the PM spot.

[jaichind]

It seems Yougov will do a snap poll after the debate.  Of course it is not clear who will watch this debate.

[jaichind]

YouGov / Sky News

Post debate scores

Trustworthy
Corbyn 45
Johnson 40

Prime ministerial
Corbyn 29
Johnson 54

Likeable
Corbyn 37
Johnson 54

In touch with ordinary people
Corbyn 59
Johnson 25

[jaichind]

Great Grimsby

CON: 44% (+2)
LAB: 31% (-18)
BREX: 17% (+17)
LDEM: 4% (+1)
GRN: 3% (+3)

(lot of undecideds mostly Lab 2017 voters ans a small sample size, but not great news for Labour)

Sort of fits the narrative that BXP eats into LAB leave voters that would not vote CON while in the South BXP tend to eat into CON base.  Of couse BXP is not running in most of the South.

[jaichind]

Sort of fits the narrative that BXP eats into LAB leave voters that would not vote CON while in the South BXP tend to eat into CON base.  Of couse BXP is not running in most of the South.

The essential problem with this narrative is that there is no reason to believe that the Brexit Party vote does not = (some of the) people who voted UKIP in 2015 and, well, that really isn't what that pattern looked like.

Sure.  But the 2015 UKIP vote in the North are also made up of pre-2015 pro-Leave LAB voters.  They went back to LAB in 2017 when many like myself thought UKIP would be a gateway drug to CON but it seems that this bunch has a real resistance to voting CON.  So if BXP did not run here I suspect many of them would not vote or vote LAB but not CON.  So my guess is that BXP running in places like this helps CON.

[jaichind]

My main concern about this latest batch of polls are that the CON+BXP vote share seems to be dropping from something like 48% to something like 46%.  Any CON gains are mostly from gains from BXP due to real shifts in support or methodological changes to take into account that BXP will only be running in half the seats.  At this stage CON gains are maxed out vis-a-vis BXP and any more gains will have to be from LAB or LIB.  But the recent trends seems to be the other direction.

[jaichind]

I am surprised the Alex Salmond scandals are not hurting SNP that much.

[jaichind]

The CON-LAB gap does to be trending downward.

[jaichind]

Interestingly, the Scottish govt is underwater on every issue that was polled: -1 on economy, -4 on Justice, -8 on education, -12 on NHS. In every case, the SNP loses voters from it's 44% topline. Seems clear that Yes/No polarization is carrying the day here.

https://twitter.com/britainelects/status/1202863050434514946

Also, here are the changes with their last Scottish poll. I think it's rather important since such a dramatic Brexit drop points to it being the effect of candidates standing down, meaning that 6% is all in the highlands and borders which the Tories need to defend.

Since CON did not win THAT many seats on Scotland in 2017 BXP did not really stand down in that many seats.  The BXP polling collapse seems more about tactical shift to CON than adjustments for BXP candidates standing down.  Not that it would make that big of a difference given how wide the SNP-CON gap is.

[jaichind]

https://www.bloomberg.com/news/articles/2019-12-06/johnson-is-heading-for-a-majority-labour-and-tory-officials-say

"Boris Johnson Is Heading for a Majority, Labour and Tory Officials Say"

That looks likely to result in a Tory majority of between 20 and 35 seats in the House of Commons, officials from both parties said. 

[jaichind]

When the campaign started I had hopes that LAB can be driven to below 200 seats.  I had similar hopes in 2017 but it was clear during the campaign it was not to be.  It seems the same this time as well.  Oh well.

[jaichind]

https://mobile.twitter.com/ElectionMapsUK/status/1203004409304821760

The Clock is ticking down on Labour at an alarming rate. Corbyn needs to win the Debate tonight, and decisively so it would seem. If he makes it about Trustworthiness he can, I am convinced.

Or for this whole Andrew Neil interview, or there lack of, thing to fester

[jaichind]

Any idea about seat projections?

Electoral Calculus is currently predicting a Conservative majority of 20 with an overall 64% probability of a Conservative majority.

With SF no taking their seats is that not de facto majority of 27

[jaichind]

https://twitter.com/ElectionMapsUK/status/1203347824789856256

If this is replicated in the rest of the weekend polls then Labour may still have some fight left in them. Compared to their last weekend poll, Labour still gains, but it is less stark - that was similarly a 10% Tory lead but both Labour and the Conservatives had higher percentages.

But why do I advise caution? Last time Comres had a 5% Other vote - a extreme outlier when polling just GB, with no NI polled (PC only gets 1%, SNP is polled and gets 3 to 4% but the tweet has character limits). So we should wait and see if this is replicated in the multitude of other weekend polls.

I got a bad feeling about this.  Nightmares of 2017

[jaichind]

https://twitter.com/britainelects/status/1203828988579385350

Uhhhhh.... I guess Labour peaked? Could be over-herding.

Survation in their final 2015 and 2017 polls went against the convention (unusually large lead for CON in 2015 and unusually small lead for CON in 2017) and were correct.  Could lighting strike the third time ?  Of course this is not their final poll which I imagine would come out the day before the election.

[jaichind]

When does the latest Yougov MRP results come out?

[jaichind]

Yougov MRP regional vote shares

[jaichind]

https://www.bloomberg.com/news/articles/2019-12-11/johnson-s-poll-gamble-just-got-a-199-million-vote-of-confidence

"Boris Johnson’s Poll Gamble Just Got a $199 Million Vote of Confidence"