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.

ukipdrop

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 4 5 6 7 8
  1. Allan Christie

    ‘Allo ‘Allo predicted it yeRS AGO.

    “The fallen Madonna with the big boobies”.

  2. @ Tony Cornwall

    What do you statisticians expect to achieve from the higher level of maths; more accurate interpretation of trends, or better forecasting, or something else? And are your results likely to be significantly different from, say simple graphical analysis?

    Assuming that I am amongst those you address, my own suggestion in the context of this site (and more generally, in fact) is that statisticians try to negotiate, and hopefully agree, rules for interpreting evidence.

    People can argue endlessly about whether Ukip support is going down, or more generally about whether the cost of living is going up, crime is going up or down and any number of other such matters. What statisticians try to do is suggest procedures or measures that are eventually adopted and accepted by a wider community. Within this definition I would say that all (serious) pollsters are statisticians. As an example CPI (and all the procedures that go into its calculation) is used to index the pensions and benefits of literally millions of people, as well as playing a role in trade union negotiations etc.

    Returning to more parochial matters, poll-watchers can either spend their time squabbling and disagreeing about what they see in the polling reports or they can accept certain ground rules and concentrate on different areas of uncertainty. Contributors here accept some ground rules but in some cases are reluctant to accept others. Everyone is scathing about ‘dodgy polls’ based, for example, on the non-random sampling of the readership of a single paper. In contrast, no one seems to have reached any agreement about what counts as ‘crossover’, or ‘swingback’ for that matter.

    Finally, why is a formal statistical approach more trustworthy than what you refer to as a ‘simple graphical analysis’? My own response to this is that simple plots risk remaining open to differing interpretation. For example, if you were to plot Ukip VIs since Jan 1 – either as a rolling average (as @Statgeek or @NC do) or in (say) weekly batches – then you would probably see what looks like a decline. But the question is whether it is a real decline or just yet another manifestation of MoE. Those who want to see Ukip fading will be inclined to interpret it as real. Others will see it as random variation. At the end of the day a graph alone cannot resolve this matter and you have to ask yourself how likely it is that the observed pattern would occur if the reality was thwt there was no decline at all. This is where formal stats trumps any kind of graph. Given certain assumptions (themselves frequently debated) this approach can assign probabilities to certain kinds of occurrence.

    I would argue that graphs can never act as a kind of referee in disagreements of this kind. But stats can and routinely do serve this function. Almost all scientific research is premised apon prior agreement of the statistical rules of the game, and tens of thousands of publications a year attest to its efficacy in this role.

    Is that enough of an apologia?

  3. Scotland 33-2 against the mighty afghan .

  4. @ Unicorn,

    In defence of simple graphical analysis (well, I would say that, wouldn’t I), I think it can give us a sense of what lies within the range of normal noise and what looks like movement, possibly before it crosses the threshold of statistical significance. To whatever extent we prefer being early to being right- and we must at least a little, or we wouldn’t be here; we will after all get a 100% accurate survey of voting patterns the night of the election- there is some value in spotting potential trends before we can be certain they are real.

    There’s a strong argument for simultaneously developing more rigorous statistical standards though.

  5. 07052015

    “Scotland 33-2 against the mighty afghan ”

    Doing better than US & UK forces then?

  6. Oldnat

    Or if Milliband could do a “Milliband?”

  7. @OldNat

    I now have a vision of you furtively lifting your glasses – “It is I, Leclerc”

  8. @PeteB

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

    Thanks for the vote of confidence but this is not, and never has been my thesis. I remain entirely agnostic about Ukip VI movements (other than to say that they are reliably underperforming their 2014 trends).

    Your comment suggests that I have claimed that Ukip support is holding up. All I have said is that none of us can tell on current evidence. It may be holding up (since jan 1, that is). It may be dropping. It could even be rising gently. All of these scenarios are compatible with the evidence.

    All I have done is challenge non-agnostic statements, such as the hypothesis that Ukip support is falling. I would be equally sceptical if anyone came along and claimed that their support was rising ..or even ‘not yet dropping sway’ -as you suggest I have claimed.

    All that I have stated is that on the evidence I have seen it is not possible to make any of these assertions without being open to legitimate challenge.

  9. Northumbrianscot

    According to Allan Christie, I might be doing that in O&S!

  10. Unicorn
    I think you’re splitting hairs, to put it mildly. If I remember correctly, another poster claimed that UKIP support seemed to be fading and you replied with something similar to what you wrote above – to wit, that the data doesn’t support that hypothesis. I am agreeing with that.

    I’m quite happy to also agree that the data doesn’t support any other variation in UKIP VI.

  11. Old nat

    I will have you know my grandfather served in the seaforth highlanders but not in any of our magnificent campaigns in afghan.

    Scotland roaring along at 59-3 after 15 .

  12. 07052015

    I enjoyed “Scotland roaring along” in the women’s cricket on ice championships.

  13. I don’t know the details of the proposals for health in Greater Manchester but I do think that the general path being pioneered there is the most significant thing happening in current UK politics. It is no co-incidence that the Financial times kicked off their series on “Dis-united Kingdom” with a close look at the Manchester model.
    Colin asks if there is the possibility of Tory votes but a quick glance at the make up of Manchester Council would discourage as 95 of the 96 councillors are Labour.. and the other is Independent Labour. In Scotland, J Murphy is powerfully aware of the challenge the coherent city region poses to, for example, Glasgow and Edinburgh.
    In terms of health, all over the world the challenge is known to be the inter-linking of health and social care. As far as I can tell the most successful model has been Sweden where acute health has been transferred to local government. The few weeks leading up to a General Election may not be the right timescale for decision-making but it is the debate that must be had eventually.
    The answer to the post-code lottery issue must take the integration argument into account. For example if the NHS is provided across (in my case) Scotland but the care funding is very uneven by local authority how can there be equity? I am on a health board but I am a councillor in the lowest funded local authority in Scotland. How can there be equity of outcomes if the other two local authorities have much higher funding for care?

  14. @ Spearmint

    In defence of simple graphical analysis (well, I would say that, wouldn’t I), I think it can give us a sense of what lies within the range of normal noise and what looks like movement, possibly before it crosses the threshold of statistical significance

    Like everyone else, I pore over graphs and find them enlightening. I woudn’t for a moment question their communicative value. To the extent that they differentiate between noise and genuine change, they often do this by adding error bars or confidence limits. In such cases I don’t think they can be treated as an alternative to statistical analysis. Instead, they are drawing upon statistical calculations and presenting them visually. At the end of the day it is the statistical component that is the adjudicator.

    That said, I totally agree that graphs will always provide the earliest advance warning of a real change. You just have to live with the false positives that eventually come to nothing.

  15. @ PeteB

    Hair-splitting is my forte ;-)

  16. Barney
    There will never be equity of outcomes whatever system is used, because the world just doesn’t work that way. Just as one example – for reasons of geography and traffic patterns, it will not take the same average time for an ambulance to reach each hospital.

    So long as some figures re outcomes are published, people will be at liberty to choose where to live.

  17. @ BC

    To put it simple: what is the biggest constraint that obstructs the most patients being treated effectively in the NHS?

    If anyone really think that devolving health to Manchester is the solution ….

    There’s waste in the NHS – this is the nature of the beast. If you cut it out (ok, I grant that there are some that can be done simply), it will come back even stronger in a different form. The question is the responsiveness in some parts of the system and very routine, following the routine in others. It’s the most difficult thing to co-ordinate in any organisation be it private or public.

  18. @ OLD NAT

    This may be a leading question that you do not wish to answer, but have the SGP and SNP made a deal that SGP would not run in certain seats so that SNP would have a better chance of winning like Orkney and Shetland where I have been on a website that says SGP will not run a candidate there.

    Thus far I have identified 31 SGP candidates out of 59 seats and note that the SGP Convention is the weekend after next.

  19. @ BC

    The FT article was quite good, and follows the aggregates. But what is the distribution of aggregates (the study on social exclusion last year chose one of Manchester’s suburbs for case study)? How many people, companies left out? There’s no doubt that it’s much more pleasant to walk in Liverpool’s or Manchester’s downtown today than 15 years ago.

    Is it a cluster, is it outward or inward oriented, etc.

    Studying regional economic development in the UK is fascinating, because you ca discount the government (with the exception of Heseltain’s or Mandelson’s departments).

  20. @ Tony Cornwall

    “….What do you statisticians expect to achieve from the higher level of maths…”

    Fortune and glory, kid. Fortune and glory…

  21. @ Barney

    The best health model is Denmark’s. It is also very expensive. It’s easy if you have social support for it. Can you see a manifesto saying 1% more tax and stroke patients will have a humane treatment and rehabilitation?

    If yes, then you say that you dedicate this tax income for a particular purpose. What’s the next? … Ok, we aren’t that keen on stroke patients, we rely on professionals, and not the stupid voters how we allocate the tax income …

  22. Andy Shadrack

    There was talk of a “Yes Alliance” immediately after the referendum, but it came to nothing – probably (though I have absolutely no inside information) because it was apparent that the SNP didn’t need such.

    SGP will maximise their political strength next year on the Holyrood List, and their stated intention was to fight “more than half” of the Westminster seats (though in reality with an eye to 2016 not 2015).

    O&S is an expensive seat to fight given the cost of flights/ferries across two island groups and there are more votes on the mainland parts of the Highlands & Islands region.

    The Greens will also want to limit the number of lost deposits sent to the Queen’s personal purse, since they aren’t normally enthusiastic Royalists! :-)

  23. Andy Shadrack

    I should probably add that the SGP are a pretty slick professional bunch, with a politically astute and media savvy leadership.

    Having been in Parliament for the last 15 years, they have learned a lot that their E&W colleagues still have to learn, I fear.

  24. @ OLD NAT

    I fear you are right, but not true perhaps of Lucas, Lambert and Scott-Cato who have all been or are MEPs. The southwest Greens appear to be very well primed for this election.

    Never send a spokesperson out without ensuring they are briefed :)

  25. Unicorn

    Nice summary of statistics vs “eyeballing a graph”.

    It’s a little surprising to see statistics denigrated on a polling site, after all polling is using statistics to infer information about the general population from a sample of 1000.

    Wouldn’t it be better to ignore all the reweighting shenanigans and just eyeball whatever number you’d like to see instead?

    Exactly the same applies to the question, “is the UKIP VI declining?” To which the answer appears to be “more likely than not, but it’s still within reasonable bounds of possibility that it isn’t”.

    If people want to jump either side of the fence at this point that’s up to them, flying your colours doesn’t actually add any useful information though.

    Graphs are useful to spot potential trends, but without proper statistical treatment, any claims are pretty much open to being accepted or rejected depending on the persons mood or opinion.

  26. @ Alan

    Weighting if applied incorrectly actually misleads the public and the voter into thinking something is happening when it is not.

    In the United States they have invented the term “push-polling”, which means you try to drive voters into making certain decisions based on data your gleaning from the polls.

  27. Today’s YG Scottish crossbreak

    SNP 48% : Lab 22% : Con 19% : LD 3% : UKIP 6% : Grn 1%

    Mean of last 20 YG Scottish crossbreaks

    SNP 43% : Lab 25% : Con 18% : LD 5% : UKIP 5% : Grn 4%

  28. Can anyone account for the massive reweighting given to Conservative voting intention in both today’s and yesterday’s You gov polls for the sun. It seems disproportionate when compared to the 2010 V I in the unweighted data.

  29. @ Andy

    Others can correct me if I am mistaken, but I think is generally accepted that with very rare exceptions [1] all current British polling is done in good faith with at least the aim of being unbiased.

    If you think that some of the pollsters are inadvertently getting thins wrong then – apart from drawing attention to this (as you have already been doing) – I would suggest it would be helpful to try to infer some of the likely consequences of the bias. Which parties are likely to benefit from the miscalculation? Which might suffer? In that way the rest if the community can try to build in adjustments for any systematic error that might be built into some of the polls.

    [1] the exceptions include ‘dodgy’ newspaper polls that make no attempt to present w balance picture

  30. Government approval at -16. I think that’s a landmark in recent times.

    BARNEY

    @” I do think that the general path being pioneered there is the most significant thing happening in current UK politics.”

    I agree-and the economic devolution deal which preceded it.

  31. This weighting back to 2010 has no basis in current voting reality for Lib Dems, UKIP and Green, the latter two who were not even running a full slate in 2010.

    In Scotland, various local government elections and latterly the 2014 European election the number of elected Lib Dems has declined, dramatically in some instances.

    Thus what pollsters are doing weighting to 2010 is distorting current reality and misleading the public and the voters into thinking something different than what has actually been happening.

    ICM, for example, by giving someone who did not vote in 2010, but who is declaring they will definitively vote in 2015, a .5 value instead of 1, is in effect distorting the voting values of all 18 – 21 years olds, who were aged 14 – 17 in 2010.

    The use of 2010 voting data, particularly for use in giving “don’t knows” and “will not say” a numerical value invalidates the poll results, in my opinion.

    A more honest approach would be to ask a respondent how they voted when they last voted and then to ask them which election that was. When up to a quarter of youth are stating they intend vote Green, as signified by Ashcroft and YouGov, in some polls, but ICM and others are claiming youth support is at 6% for Green, then it is clear that distortion is happening.

    Another way to determine ‘”don’t knows” and those who indicate they may change their minds before May 7th is to ask them which party is their second choice.

    In the 2010 GE 88.1% of voters opted for Conservative, Labour or LD. Currently the polls are indicating somewhere between 72% to 74% are that way inclined, but due to weighting back to 2010 that figure could actually be much lower.

    At some point one has to ask are some pollsters purposely trying to supress voters from supporting a party other than the Conservatives, Labour or the LDs.

    It is not enough to call it a “house effect” when it is a universal practice that undervalues some respondents answers as to which party they are supporting while overvaluing others. That is called distortion.

  32. Meanwhile…Scottish Cricket is in Crisis with both fans calling for change!

    Peter.

  33. Andy Shadrack

    “At some point one has to ask are some pollsters purposely trying to supress voters from supporting a party other than the Conservatives, Labour or the LDs.”

    The polling companies (as opposed to some obsessive employees like Anthony who strive for the best in political polling accuracy) are commercial companies more interested in profit than anything else.

    For most (if not all) political polling is a sideline, which can raise their commercial profile.

    If investing expensive staff time in developing the most accurate political weightings doesn’t improve the bottom line, why would they do it?

    A company may decide that getting the most accurate snapshot in February will produce no extra revenue, while accurately “predicting” the May result would.

    Much easier then, to guess that, as in the past, minor parties will be squeezed in the election, so a spurious accuracy (they would hope) can be achieved by squeezing them now.

    Remember Deep Throat’s dictum “Follow the money”.

  34. @OLDNAT

    ”Today’s YG Scottish crossbreak

    SNP 48% : Lab 22% : Con 19% : LD 3% : UKIP 6% : Grn 1%”

    Headline – ”Poll shock – Tories close to just 3 points behind Labour in Scotland…an upset is likely…”it’s 1979 all over again” says one jubilant Tory candidate…while Murphy refuses to accept it is all over in East Renfrewshire..”

    OK, I’m getting completely carried away again amn’t I… ;-)

  35. @ROLL A HARD SIX

    No you are obviously completely right in your analysis. Don’t panic Mr Mainwaring, don’t panic!!

  36. Ok Peter, so who’s the other fan?

  37. Barney,

    Below I( have copied and pasted headlines from the King’s Fund, “Improving the public’s health”. This follows the Marmot review of 2010.

    Scottish research into health inequalities, as I am sure you know, offers its own solutions but are broadly similar to those headlines below. I hope I am not in breach of Anthony’s rules in posting this.

    “Nine key areas that can improve public health and reduce inequalities

    What is in the resource?

    You can access specific web resources for each section below. The report focuses on practical actions that local authority officers and teams can take in these nine key areas:

    the best start in life
    healthy schools and pupils
    helping people find good jobs and stay in work
    active and safe travel
    warmer and safer homes
    access to green and open spaces and the role of leisure services
    strong communities, wellbeing and resilience
    public protection and regulatory services (including takeaway/fast food, air pollution, and fire safety)
    health and spatial planning.
    You can read the full report here: Improving the public’s health: a resource for local authorities, including an overview written specifically for local authority leaders.”

    sam

  38. Roll A Hard Six

    :-)

    That’s not far from some of the responses there used to be to individual crossbreaks!

  39. Morning all from the deep deep south

    My that’s a purdy little damn fine Scottish cross break hell boy it’s hotter than a billy goat’s ass in a pepper patch.

  40. OLDNAT

    “Andy Shadrack
    I should probably add that the SGP are a pretty slick professional bunch, with a politically astute and media savvy leadership.”

    Sadly, one Green candidate is repeating the claim that the O&G industry is given subsidies of billions of £s each year while the renewables industry receives hundreds of millions. It is not disclosed that this is a reference to different rates of VAT. Most UK companies pay VAT at 20% I think. For fuels, the VAT rate is 5%.
    The subsidy is unreal. It is the additional amount that would be paid (by both the renewable sector and the O&G sector) if VAT was levied at 20%. The reason that the O&G sector would pay billions and the renewables sector millions is because the O&G sector is so much larger.

    This kind of campaigning dismays me. I admired Patrick Harvie last summer.

    sam

  41. Sam

    This could become an exceptionally long and tedious thread, if we all quoted examples of “one candidate” of a party that selectively, and misleadingly, quoted statistics.

    650 seats @ an average of 5 candidates per seat = 3249 more!
    :-)

  42. Thanks for the very interesting discussion regarding p values etc.

    I think as people who analyse polls we have to be honest and recognise that we are in the business of trying to support statistical syllogisms rather than apply a properly scientific method. The source material, whilst useful, is simply not to up to the latter task.

  43. UNICORN
    Hair-splitting is my forte ;-)

    In reading psephological analysis, it seems to me, it is normal to recognise the interest of users in judging, for example, the importance of data about attitude to migration in relation to age and education, in voter responses to party policy. For that purpose I find the distinction which your hair splitting may provide between MOE and a true trend very useful.

  44. OLDNAT

    “slick”…..”professional” …”astute”? :-)

    sam

  45. Andy Shadrack

    Which is why production of raw data it an important step to lend confidence to the numbers produced.

    Push polling isn’t even polling. It’s typically a thinly veiled smear campaign. Such as “Should XYZ stop beating his wife?”. Done in large numbers without a care about producing a worthwhile result (apart from a shift in voter attitude about XYZ). It has no relevance to this site.

  46. The Times has an article today regarding the Russians asking for up front payment from the Ukraine for supplies of Gas.
    This could be Putin turning the economic screw.
    Today the Ukraine, tomorrow Europe.
    This could explain why Merkel and Hollande are going easy on the Russian Bear.
    We all need to tread carefully with “Uncle Ivan.”

  47. ”SNP 48% : Lab 22% : Con 19% : LD 3% : UKIP 6% : Grn 1%”

    How about:

    “Lib Dems claim ecological high-ground as their VI soars to 3x the Greens.”

    “SNP unable to reach 50% – Opponents say it’s a sign they’re starting to panic”

    “UKIP set to take Scotland by storm as VI jumps to 8x 2010 results”

    I’m wasted as a chart tinkerer.

  48. Statgeek outed as Pressman. :)

  49. Today’s YG Midlands/Wales crossbreak

    PC 1% : Lab 41% : Con 29% : LD 4% : UKIP 17% : Grn 7% : Oth 1% :

    Mean of last 20 YG Midlands/Wales crossbreaks

    PC 2% : Lab 36% : Con 33% : LD 5% : UKIP 16% : Grn 6% : Oth : 1% :

  50. Jasper

    Considering the Ukrainian currency is on a massive slide and is worth about a quarter of what it was last year after running out of foreign reserves (and has fallen 20% today). I’d say the Russians were being astute about their credit arrangements.

    If they wanted to turn the screw, they would be exercising a clause in their loan and demand an immediate repayment of the $3 billion due to Ukraine’s debt levels exceeding a threshold.

    I suspect most international suppliers would demand payment up front and being paid for in any currency except for the Hyrvina.

1 4 5 6 7 8