Monday’s polls

We have three GB polls today, from Populus, Ashcroft and YouGov.

  • The twice-weekly Populus poll has topline figures of CON 32%, LAB 34%, LDEM 8%, UKIP 14%, GRN 5% (tabs).
  • The weekly Lord Ashcroft poll has topline figures of CON 34%, LAB 31%, LDEM 7%, UKIP 14%, GRN 7% (tabs). Labour have dropped five points since Ashcroft’s previous poll, but this will be largely a reversion to the mean after they jumped up five points a week ago.
  • Finally the daily YouGov for the Sun has topline figures of CON 35%, LAB 32%, LDEM 7%, UKIP 15%, GRN 6%. The three point Tory lead is the largest that YouGov have shown since back in January 2012.

Two three point Tory leads on the same day. All the usual caveats apply – it is only two polls and Populus showed a two point Labour lead. It wouldn’t be the first time that two polls have popped out on the same day showing something unusual, only for it to turn out to be pure co-incidence when polls in the following days showing everything back to normal. Keep an eye on it though…

319 Responses to “Monday’s polls”

1 4 5 6 7
  1. @Couper

    He tends to clearly separate his posts and tweets on polling from those on gambling. I find him very fair and impartial on polling, perhaps with slight unconscious anti-Tory bias

  2. Having read UKPR for about a year decided to join the fun in the lead up to the most interesting election in a generation.
    On the hard left of the Tory party, I will endeavour to keep my bias to a minimum

  3. Howard Alert

    There is another one purporting to be me.

  4. @Couper

    “But SNP would support Labour in the confidence vote forcing Labour to bring down their own government.”

    No, events would pan out more than this:
    1. SNP fails to support a Labour budget (which in the old days would have precipitated a general election)
    2. Labour has another go, signalling that 3 below will be the consequence if it fails. Same result.
    3. At which point Miliband walks out of Number 10, posts the keys through the letterbox and announces that Labour is withdrawing from government given it’s inability to pass legislation without SNP support.
    4. To a chorus of “no you can’t do that, the FTPA doesn’t let you”, Miliband replies “I just have. I now challenge any other coalition of parties to form a government and demonstrate that they can carry a vote of confidence, and if they can’t then we need a fresh GE.”
    5. We go two weeks without anyone claiming to lead a government and without a new one being formed and carrying a vote of confidence, at which case the time limit in the FTPA is exceeded.

    I can’t see any outcome other than a fresh general election at that point.

  5. Guymonde

    You are mistaken. I am not a tin hat.

  6. @ OLD NAT

    I went to Scotland Votes and plugged in 27%-28% for Labour, 44-46% for SNP and 18-20% for Conservative. 2-4% for LD and 2-4% for UKIP and 2-4% for SGP

    I got:

    SNP 46
    Labour 10
    Conservative 2
    LD 1

    The 2nd Conservative seat mleted away at 18% and
    became an SNP seat. Both Conservative seats are on the border.

  7. Phil Haines
    “I can’t see any outcome other than a fresh general election at that point.”


    On another matter, would it be easier if we all changed our names to Howard?

  8. Andy Shadrack

    Yes but remember that, as far as we know, Scotland Votes operates under Scottish UNS parameters.

    Presumably it has different methodologies for the Holyrood section in order to allocate list seats, but we don’t know,

    Best treated as a bit of fun until we know more.

  9. @Phil Haines

    Of course no one can force Ed M to carry on as PM if he doesn’t want to but it does put the ball firmly in the Tories court.

    If the Tories think they can win a GE then they go for a GE & Miliband loses, if the Tories don’t think they can win they make the SNP and others offers they can’t refuse.

    Far more likely that Miliband deals with the SNP and others in order to avoid this kind of shambles.


    I know the SNP are a burden to the Labour party just as much as the Labour party are a burden to the Tories and so on. It’s called politics

    Learn to live with it.


  11. And they’re all a burden to the electorate…

  12. TIM N
    “Good evening all . Thanks for all the welcomes.
    Will be interesting to see what the YG poll brings.
    My prediction Con 36 Lab 33”

    Think I might take a punt on that VI as well but the Sun are tweeting some movement for two parties so it could be an increase on the Tory and Labour VI but my guess it’s going to see a dip for UKIP.

  13. Pete B


  14. PETE B

    Can’t argue with that. :-)

  15. I am being to think that an unintended consequence of the fixed term act is to make calling an election much more difficult in that it makes it hard for the PM not to look like they are throwing into the towel effectively making it very hard to win the following election.

    A more likely scenario is to engineer removing your leader and then try to suggest that the election is to give the “New” leader a fresh mandate.


  16. @Couper

    I think you are overplaying the SNPs hand. It really isn’t that strong. It’s smoking in Scotland. But only in Scotland. The SNP will poll at best 5% in GB. That’s 95% of the electorate voting for other parties.

    Clearly that is too crude a metric, as the SNP are likely to be kingmakers of the UK government (whether through coalition or c&s). But the same was true of the LDs, and look what has happened to them.

    The SNP will believe their position is different to that of the LDs as they already run Holyrood with an overall majority and will likely poll 40%+ in Scotland at the GE. But is it that different on a UK level? And will either Lab or the Tories give the SNP DevoMax if they do not have to? The Tories have only argued for EVEL as it appeared at the time that the Tories would always have a shot at a majority on English only affairs. However, polling is showing that as Labour recedes in Scotland it is getting closer to the Tories in England.

  17. More on potential model biases.

    Yesterday, I posted a comment suggesting that certain seat-projection models may be systematically biased against the Tories. Since then, I have done more checks to fill in some of the gaps and this has firmed up my views about the biases.

    Specifically, I have looked at each of the models that publish seat-by-seat projections asking how they call each of the 133 seats in which Lord Ashcroft has done constituency polling since May 2014. All classifications are based on downloads taken earlier on today. Two of the models (May2015 and ElectionForecast (EF)) explicitly make use of Ashcroft constituency polling data. The third – Electoral Calculus (EC) – does not. As you know, May2015 has a feature allowing you to run calculations that disregard the use of the Ashcroft data – falling back on the Strong Transition calculations borrowed from the EC team.

    Naturally, one would expect models using Constituency Polling data (dubbed CP-models, to abbreviate) to mirror the polling data, and my tests using the Euclidean Distance metric duly confirm that they do. (Accuracy falls with passing time due to the inclusion of drift adjustments.) Non-CP models are unlikely to mirror the Ashcroft seat projections as well as models that build those numbers into their very calculations. So, I am not even going to compare the two for accuracy.

    Instead, what I am looking at (in more detail, for those who read yesterday’s comment) is systematic biases between the two types of model. It might have been the case that they came up with broadly similar profiles for the 133 seats, with differences seen only in the precise seats contributing to the party totals. However, the results show that this was not the case.

    The two CP-models produced similar profiles across the 133 seats:

    Election Forecast: Tories win 37.5 seats; Labour – 62 seats; LibDems – 16; SNP – 15; Greens – 1 seat and Ukip 1.5 seat. [EF has a number of constituencies in which the top two parties are assigned identical VIs, and in these cases I have assigned half a seat to each party].

    May2015: Tories – 41; Labour – 58; LDs – 17; SNP – 15; Greens – 1 and Ukip 1.

    It goes without saying that neither model is suggesting that Labour are way in the lead across the country. It is just that Ashcroft happens to have concentrated his polling in seats where Labour are currently ahead, and so the two models just reflect back the choices he had made.

    What about the non-CP models?

    Electoral Calculus: Tories – 35; Labour – 84; LDs – 5; SNP – 8; Greens – 1 and Ukip = 0.

    May2015 (with Ashcroft switched off): Tories – 33; Labour – 79; LDs – 5; SNP – 14; Greens – 1 and Ukip -1.

    Again very similar to one another (not surprising, given that they use the same methods). But the profiles are strikingly different from those for the CP-models. In the absence of constituency evidence, the models predict the LibDems will lose the vast majority of their current seats polled by Ashcroft. EC doesn’t seem to have kept up with the full extent of the SNP post-referendum surge. Most critically, however, the non-CP models are far more Labour-leaning that the CP models are. The labour margins are 49 (84 minus 35) for EC, and 46 for May2015 with Ashcroft switched off. The corresponding margins for the standard version of May2015 is just 17, with EF projecting a margin of 24.5.

    This is a difference in the high 20s over a relatively modest number of constituencies. With discrepancies of this order, clearly both types of model can’t be right. It is therefore sensible to ask which type of model is currently making better projections. Your answer will depend on your own view of the status of Ashcroft constituency polling. If you believe this can’t be trusted then you should probably go with the projections of the non-CP models, and so take the view that Labour is sailing comfortably towards being the largest party. If not, then the CP-models are the ones for you to watch.

    All of this concentrates on the seats that have actually been polled over the last 9-10 months. Of course, the biases – if that is what they are – will probably extend to quite a few other marginal seats that have not yet been polled, and so the discrepancies could be much larger than this overall.

  18. It seems that the reality of coalition government is still simply opaque to many here.

    If parties make an Agreement for the length of a parliament, that’s that. Only an extreme event could bring that under pressure and yes, it sometimes does, but that is an unforeseen event by definition.

    Otherwise there is no sense in signing up to the Agreement in the first place.

    If polling stays roughly where it is, there will be almost certainly a Governing Agreement between Lab, LD and SNP, stress *if*.

    Very boring for those who want drama, but that’s probably the way it will be..

  19. What holds Ukip together? Having been to two Ukip meetings this week it is clear that some members views fit with the Thatcherite view of the Cons and some members views fit with the factory floor wing of Old Lab. What other party has ever had such diverse membership.

    The glue seems to be of dislike of EU by the right combined with a fear of large scale immigration by the left.

    This raises various questions. Here are two;
    Immediately how to predict churn in and out of Ukip before this GE;
    Longer term can the party survive beyond Farage?

  20. Unicorn

    Fascinating and instructive – as always.


    @”What holds Ukip together? ”

    Nigel Farage ?

  22. @ StanVax

    On the hard left of the Tory party…

    Really? The Tory Party has a hard left… who’d hae thought it.

    And hello & welcome, BTW. :-)

  23. BristolianHoward

    Coalition seems a very unlikely prospect.

    However, C&S under these new circumstances would probably involve some form of publicly disclosed agreement – otherwise it is minority government.

  24. Whenever we are told a poll is interesting it turns out to be dead boring.As this
    will probably be.

  25. @RAF

    The SNP don’t want to be king makers they want Scottish independence. So both Labour and Tory parties, who I assume are sticking with a unionist position, will have to be careful how they deal with the SNP. Similarly the SNP will have to be careful they don’t set back their ultimate goal.

    If Scotland returns many SNP MPs but gets nothing then perhaps that will actually aid the SNP’s long term goal.

    Interesting times


    Another thought provoking analysis, thank you.


    Your analysis raises the question why Ashcroft polls are so different to the others.

  28. Boring! Sun says the Lib Dems have hit a 25 year polling low

  29. CON 36%, LAB 34%, LDEM 5%, UKIP 14%, GRN 6%

  30. Tom Newton Dunn [email protected] · 1 min1 minute ago
    EXCL: YouGov/Sun poll tonight – Libs drop to 5%, a 25 year polling low. CON 36%, LAB 34%, LDEM 5%, UKIP 14%, GRN 6%

  31. Latest YouGov poll (02 – 03 Mar):
    CON – 36% (+1)
    LAB – 34% (+2)
    UKIP – 14% (-1)
    GRN – 6% (-)
    LDEM – 5% (-2)

  32. @ UNICORN

    My apologies I was not so much trying to suggest you had taken a certain position, though I do not read all posts, but rather to lay out a position as to why I thought such a position was erroneous.

    Neither UKIP or Green are shrinking in relation to the 2010 GE, but LD is and so I do not disagree that in comparison to 2010 GE there is a swing away from LD to other parties.

    But I object to that same statistical observation being made about UKIP and Green, where to my mind it is yet to be proven, statistically or otherwise, that the vote they recieve in the 2015 GE will actually be less than the votes they received at the 2014 European election.

    I certainly do not believe that anyone is claiming that SNP, UKIP and/or Green will obtain less than the votes they received at the 2010 GE.

    So swing or squeeze, as it relates to 2010 weighting, is not a good measure for those three parties. For Labour, Conservative and LD maybe that is a correct measure.

  33. That is a wacky poll.

    Not the Tories- if swingback really has kicked in we should start to see some 36%s for them- but does anyone other than Chris Lane think the Lib Dems are declining at this point?

  34. Tony Cornwall
    I’ve commented before that it will be difficult for UKIP to formulate policies that satisfy both wings, which is presumably why very few policies have been revealed yet, even after their latest conference, where they were supposed to be unveiled.

    The underlying unifying factor (IMHO) is old-fashioned patriotism, which is widely sneered at and/or denigrated as ‘racist’ by some.
    This might be why UKIP has more older supporters than other parties, as they can remember a time when patriotism was not a dirty word.

  35. @Tony Cornwall

    It raises lots of questions and I’m afraid I am grappling around for answers. But I don’t think we are looking at a difference betwen the Ashcroft polls on the one hand and “The Rest” on the other.

    What is happening is that polling technology – applied at the constituency level – is being pitted against a set of mathematical algorithms uses to convert national (or regional) VI changes into shifts in individual constituencies. The real question I suspect is whether it is safer to trust tried and tested polling methods (albeit used across a small geographical area) or whether it is best to place your faith in Strong Transition, or UNS calculations or various other algorithms of the kind.

  36. TIM N
    “Good evening all . Thanks for all the welcomes.
    Will be interesting to see what the YG poll brings.
    My prediction Con 36 Lab 33

    Not far off the mark for a new poster.

  37. Hmmm. So Con lead over Lab down to 2% and the poor Lib Dems being squeezed further.
    Any thoughts on why Con have risen to 36 % during a quiet few days. That’s a sharp rise from where they were on YouGov a couple of weeks ago.

  38. When was the last 36 for the Tories on YG?

  39. @ Unicorn,

    Thanks for that analysis. Fascinating as always!

  40. All this talk of SNP deals may be premature if the Cons are moving ahead.

    What has happened over the last few days to give Cons this boost? They only thing is the tuition fee announcement and that should have helped Labour.

  41. The Lib/Dems really are a finished article. Nick Clegg should just pull the plug on his party. They will be bankrupted after all they lost deposits are counted.

  42. Pete B

    “This might be why UKIP has more older supporters than other parties, as they can remember a time when patriotism was not a dirty word.”

    That may help to explain why one party in Scotland, whose support is concentrated among older people, has adopted “Patriotic” as a self-descriptor.

  43. Yougov is really odd right now. I cannot think what might be driving these apparent changes.

  44. Another Yougov Con lead on a Tuesday.

    So that’s 8 YG Con leads this year, all but one on Mondays or Tuesdays. [The eighth was on a Wednesday, so it’s lucky for Labour the election is on a Thursday.]

  45. Its only one poll so the drop in Lib Dems should perhaps excite us too much. Perhaps more important is that It confirms that Tories ahead of Labour slightly for another night though difference v small.

  46. @ Neil A,

    2014-10-03. Middle of conference season.

  47. Couper

    Other than massive events, views among floating voters are seldom influenced by a policy announcement or an event.

    Views shift slowly and are affected by all kinds of “mood music”.

  48. @ Alan Christie

    I wasn’t far out. You were spot on with your dip for UKIP. I think 36% is the highest Con score with YG for at least a year.

  49. @ Couper,

    What has happened over the last few days to give Cons this boost? They only thing is the tuition fee announcement and that should have helped Labour.

    Rich pensioners worried about having their pensions taxed? But I’d have thought anyone likely to begrudge it was voting Tory already.

    I’m a little puzzled as well, but I’m guessing it’s a combination of random variation and the beginnings of swingback for the Tories. Ukip are down a bit and they’re enjoying a corresponding boost. They actually crossed over on the YouGov average last week, so it’s not entirely surprising we’re seeing a few leads for them.

  50. Cross-over during March anyone on YG if not the UKPR Ave but at least level pegging looks likely at some point.

1 4 5 6 7