Two new polls today – the daily YouGov poll for the Sun and the monthly Survation poll for the Daily Mirror.

Survation in the Mirror have topline figures of CON 28%(-3), LAB 34%(+4), LDEM 10%(+3), UKIP 19%(-4), GRN 4%(+1). Lots of sharp changes there since their previous poll, but usual caveats apply – the Tory lead in Survation’s previous poll was rather unusual in itself, today’s large Labour lead also unusual, hence the large changes from one to the other. Note also the drop in UKIP support – Survation consistently show the highest UKIP support, so while 19 is large compared to other pollsters’ figures, its a notable drop from Survation.

Meanwhile the daily YouGov poll for the Sun has topline figures of CON 35%, LAB 33%, LDEM 6%, UKIP 14%, GRN 7%. A two point Conservative lead from YouGov, the first time they’ve shown that for just over a month.

The bigger picture remains the same. The Conservatives probably haven’t moved ahead, or we’d be seeing that across most of the polls, when actually they are averaging out at a tiny Labour lead. Neither is there is big swing to Labour, or we’d be seeing that across most of the polls, when actually they all just seem to be showing normal variation around the margin of error. In terms of the Labour vs Conservative race, 2015 so far has been largely static. The only trend that may be meaningful is the drop in UKIP support.


Now that Survation have published their monthly poll we can compare UKIP’s January and February scores across all the pollsters (I’ve taken an average for those companies who publish more than once a month). There does seem to be a pretty consistent fall in UKIP support, perhaps slightly obscured by the fact that the most frequent pollster, YouGov, shows one of the more modest drops and the second most frequent pollster, Populus, changed their methodology at the start of February in a way that increased UKIP support.

359 Responses to “Latest YouGov and Survation polls”

1 3 4 5 6 7 8
  1. There are actually 5 National Health Services (everyone always forgets about poor Isle of Man).

    Interestingly Jersey and Guernsey do not have NHSs although they do cooperate with the other health services on things like training and payscales with some common elements.

  2. Apparently there was a major conflict in the garden this afternoon between two blackbirds.

    I can only assume the one who had been singing from 3am for months seeking a female mate was a bit miffed that another fellow had nipped in and reaped the reward of all of his hard work.

    That’s how socialism started!

  3. @ Martyn

    Hmm, be careful. If your model is linear all you are doing is testing to see if it changes linearly, which is not quite what you want.

    The first point I would make is that my comment was posted as a quick riposte to @AW’s suggestion that Ukip VIs have fallen reliably since Jan 1. I wasn’t claiming to offer a definitive analysis: I was just pointing out that not all the evidence supports the claim in question.

    The second it that your statement above is wrong, and my guess is that you wouldn’t have written it if you hadn’t been as exhausted as you describe. Simple linear regression does not test whether a function changes in a linear fashion, It imposes a best-fitting straight line on whatever sample of data points are under examination and then – based on certain distributional assumptions – it can tell you whether the intercept and slope differ reliably from zero. What I stated yesterday was if you fit a straight line to the Ukip VI figures from Jan 1 onwards the slope of this line does not differ significantly from zero. In other words, on this evidence Ukip support has not been dropping over this period.

    The next sentence in your comment suggests to me that you are well aware that the method doesn’t test whether something changes linearly. As you say, you can perfectly well carry out a linear regression on the graph points of an explicitly non-linear function like a sine curve, and by doing so it is perfectly possible to find whether this non-linear function has intercept or slope that differ from zero (a sine curve oscillating around a line at 45 degs would have a non-zero slope).

    Apologies to others for all of this stats-talk, but the main point is that we don’t yet know whether Ukip VI is genuinely falling or whether the changes we can all see are simply manifestations of MoE.

  4. Jasper22

    Well according to Wiki:

    Since the Laws in Wales Acts 1535–1542, which formally incorporated all of Wales within the Kingdom of England, there has been no geographical or constitutional basis for describing any of the territory of Wales as a principality

    so I think someone’s been watching too much Wolf Hall. (Henry VIII had a great uncle called Jasper)

  5. Talk of fishing for p-values always reminds me of Neuroskeptic’s classic:

    The Nine Circles of Scientific Hell

  6. Northumbrianscot

    I don’t think you are allowed to count health systems that are outside the UK.

  7. As there is a Prince of Wales, but no King (or Queen) of Wales, then surely it is a principality by definition?

  8. Pete B

    As there is a Prince of Wales, but no King (or Queen) of Wales, then surely it is a principality by definition?

    No, not anymore than there being a Duke of Cambridge makes Cambridge a Duchy.

  9. YouGov Tory swingback, Its happening…

  10. No, it’s just an archaic term that has continued in the vernacular long after its original meaning was lost. A bit like Fife being a “kingdom”.

  11. Called last night by Populus and tonight at my mother in laws houses.

    Both asked a specific “Thinking of the named candidates standing in your seat”, so I suspect it is for Lord Ashcroft, which is good news for this seat;

    Ross Skye and Lochaber.


  12. @ Roger Mexico. Spot on Roger. Prince Charles has no constitutional hold over Wales.

  13. @OldNat

    I believe a lot of the more complex IoM cases end up in Manchester so there is a relevance to the discussion!

    We have an interesting piece of cross NHS infrastructure in Scotland – The State Hospital for Scotland and Northern Ireland in Carstairs. All the really dangerous psychotics from NI get sent to Scotland.

    Some might says that’s fair enough as Scotland has been doing the reverse for the past 400 years…

  14. @Peter Cairns (SNP)

    Surely, it’s not just a matter of CK’s personal popularity.
    If you are lucky enough to live in the constituency of Ross, Skye and Lochaber what could you possibly be unhappy about?

  15. @ Unicorn

    My post wasn’t against you, I certainly appreciate what you are doing and advocating (evidence-based claims). I do have a problem with using the model as a predictive model (it keeps on reminding me student recruitment models for universities, although they are not sophisticated at all). So there is nothing personal here, just the opposite. And I certainly appreciate that you try to use all the available data. Reading your posts, and your results sometimes I feel encountering a statistical picture of a meta analysis (which is barely ever published)

    We have random samples here – can we aggregate it? If so what is the epistemological basis.

    Do we have the same question as the ones that created the VEs? If not, what constraints do we have (if I’m refereeing any paper where the design doesn’t correspond to the methodological design – the original problem of the research was different from the one that the paper is about – I recommend automatic reject. Go back and make a different design).

    Do we actually know what we aggregating across different polling companies (and also within one)? can they be aggregated? The means are innocent (although I would like the standard deviation with it), but when we go to trends?

    We have the following problem: we have quite solid data on E&W and on S VI, and it is treated NOT as indicative of the HoC results, but as determinant (to various levels), which no narrative can support. The VI of GB, with trends can be interesting, but what story do they express in cool numbers, and are the constructs flawed?

    This is why I think extra caution is required with trending VI – I genuinely think that there are no trends in isolation (I can’t think the mechanism that create it) in a STATISTICAL sense. think of 1994-97. There was a trend, but which narrative did accurately describe the manifestation of the the trend in the numbers?

    This is why I suggested the Bayesian trial here: what data would say that Con/Lab would win? Horrible, but create the data. Match it what we have (VI) and analyse the two sets.

    Now for constituency, it’s even more arbitrary: creating a kind of truth table, and then applying a Boolean reduction. So, arbitrarily, what is the level of exposure to UKIP counsellors that would turn you away. What is the perceived threat of a party’s win that make you vote for the second disliked parties, etc.

    It would give you a few dominant and recessive variables with probably about 8 pathways in the UK that could predict voting.

  16. Richard

    The weightings on the last page of the YouGov London poll are presumably just a formatting mess-up rather than supernatural powers in predicting exactly who will reply to the survey. Especially since it says (c)2010 at the bottom, so four years in the future would be really impressive.

    I suspect it may have something to do with the change in poll commissioner from the Evening Standard to the Times, and maybe other alterations being required. Presumably the ES got upset at polls not providing exactly the answers that Boris wanted.

  17. Prediction:

    Lab 34%
    Con 33%

    Total @ 94% with SNP on 4% (45%)

  18. Northumbrianscot

    Remember that the most villainous families from both sets of Border Marches were exiled to Ulster, in order to bring peace.

    I assume you are descended from the more civilised people brought into Northumberland to replace them. :-)

  19. My guess:

    Con 38
    Lab 24
    LD 13
    UKIP 7
    Grn 12
    SNP/PC 6

  20. Wales hasn’t been independent since Owain in the 1400s – and even then his writ did not run large.

  21. @Roger Mexico

    So the only London regional newspaper (and a freesheet no less) has ditched London polling in favour of expensive tome whose online arm sits behind a paywall. Shame.

  22. @Mr N

    Greens too high.

  23. @Roger

    Thanks – I agree as the actual poll has different figures for weighted vs unweighted sample rows. I really just want to see the ethnicity weighting to see if Yougov are managing to get more BME’s onto their panel.

    I’m interested to see if the immigration leaflet has caused any drift from Labour, and this poll may be the only indication that something may have moved- but not really if it looks like last months with a massive upweighting of BME respondents.

  24. @Statgeek

    I’ll go with Con 35 Lab 34. That would qualify as “interesting”.

  25. YouGov/Sun poll tonight – Tories have a 6-point lead: CON 39%, LAB 33%, LD 6%, UKIP 12%, GRN 7%

  26. That’s my prediction before some of you end up at A&E

  27. Is that a constiueny poll for East Renfrewshire Allan? ;-)

  28. @AC

    So professional. I nearly believed you. Loool

  29. @ Allan Christie. You had me worried for a minute!

  30. YouGov/Sun poll tonight – Labour and Tories tied: CON 33%, LAB 33%, LD 8%, UKIP 15%, GRN 6%

  31. Well I got the Labour VI spot on!!

  32. @Allan Christie

    A bit sad really!

  33. @AC

    What’s interesting about that?

  34. And what is interesting about that poll? More Polldrums.

  35. I’m impressed by how that poll managed to have no interesting features whatsoever.

    Well, I guess Lib Dems on 8% might signal a slight recovery.

  36. JASPER22
    Twitter reporting a poll due later which has “surprises”

    Got a link to that tweet?

  37. It is interesting as Unicorn is being proved right – UKIP back up at 15 again….I’m starting to think this stats stuff is worthwhile….

  38. Shall we just cut out the next few months of nonsense and declare it a score draw now?

  39. Business as usual then. I thought we would see another Con lead tonight and the start of a trend.

  40. It seems to confirm Unicorn’s thesis that UKIP’s support is not yet dropping away.

    Is that a constituency poll for East Renfrewshire Allan? ;-)

    I wish Ashcroft would poll in our seat.

  42. Tory support slumps by 2% in one day as Labour draw level.

    The Straw effect, maybe or is 33% the new Tory ceiling?



    “The Straw effect, maybe or is 33% the new Tory ceiling?”

    You’re clutching at straws.


  44. @ Catmanjeff,

    What, and miss the excitement of hearing the phrases “Long-term economic plan!” and “You can’t trust the Tories with the NHS!” repeated four trillion more times?

  45. If DC doesn’t secure a majority (again) & Boris is an MP, but DC continues as PM, what % VI at GE do we think he’d need to survive 12 months, most votes? 2%, more ?

  46. It’s looking like only “an event” will change the polls.

    Now, if Cameron or Osborne could “do a Madonna” on live TV …..

  47. @ Martyn

    “don’t the NHS guidelines these days say that confidence intervals should be quoted instead, or have they changed since the new lot got in?”

    They do, but you you can include your significance testing in appendix, and guess what….

    NICE is smart in it, and they catch you in no time if you are trying to play the game with significance, but in management reorganisation of NHS it’s all p value (e.g. Hospital or in community care).

  48. @Laszlo

    Thank you for the posts about the proposed changes to the NHS in Manchester.

    I saw a debate on C4 news in which a Doctor was saying breezily that all will be well whilst being unable to give direct answers to the questions being asked.

    That so much power over the health of Mancunians would ultimately lie in the hands of a Mayor is worrying indeed. IIRC, didn’t Mancs reject electing a Mayor previously ?

    Additionally there’s things like currently all patients requiring treatment having the right to ask for it at any hospital of their choice. If this scheme goes ahead, it would not be in the interests (financially) of NHS Manchester to treat anybody from outside the GM boundary.

    That this is being announced as a ‘done-deal’ is atrocious.

  49. OLDNAT

    “It’s looking like only “an event” will change the polls.
    Now, if Cameron or Osborne could “do a Madonna” on live TV”

    LOL, I just seen the clip on the news.

  50. @Spearmint
    “I’m impressed by how that poll managed to have no interesting features whatsoever”.

    That made me larf (as R&D would say). It reminds me of the poem (Hugh Mearns apparently)

    Yesterday, upon the stair,
    I met a man who wasn’t there.
    He wasn’t there again today,
    I wish, I wish he’d go away…

1 3 4 5 6 7 8