This morning’s Mail on Sunday had a new poll of South Thanet which they built up into a UKIP “covering up” an unfavourable poll showing them headed for defeat. ComRes have subsequently released the tables for the poll here, revealing it was actually commissioned by ChartwellPolitical, an agency founded by two former UKIP staffers.

First let’s cut away the Mail’s hyperbole – it’s really not a “LOSER POLL!” and doesn’t show Farage heading for a humiliating defeat. UKIP are one point behind the Conservatives, with Labour one point behind them – CON 31%, UKIP 30%, LAB 29%, LDEM 5%. Given the margin of error, one couldn’t confidently say which of the three parties are ahead. What it actually shows is an extremely tight race. However it’s significantly less positive for UKIP than the previous polling in Thanet South – a Survation poll back in February that showed UKIP ten points ahead, and it resulted in UKIP attempts to rubbish the poll and its methodology last night.

In terms of methodology, the poll is mostly done using the same methods ComRes use in their national telephone polls – same turnout weighting and filtering, same squeeze question, same treatment of don’t knows. There are two important differences between the way ComRes do their national and constituency polls though. First respondents were prompted with the individual candidate names, secondly the poll was NOT politically weighted (if it had been, it would probably have been better for the Tories to some degree, depending on how much false recall ComRes allowed for in setting targets).

Most of the criticism of the poll last night (including some from UKIP themselves) was frankly complete nonsense. I can only assume a lot of it was sourced from “what some bloke on Twitter reckoned”. To sum up, the difference isn’t because candidates weren’t named – they were. It isn’t because 2010 political or turnout weights were used – they weren’t. It wasn’t because people who are unlikely to vote were included – they weren’t. It wasn’t because ComRes reallocated people by 2010 vote – they don’t. The idea that the question wording mentioning “your local MP” favoured Laura Sandys seems somewhat stretched, given the question included candidate names and Laura Sandys wasn’t one of those candidates.

To look at the more substantive things people have asked though, the initial voting intention question in the poll found Nigel Farage ahead. What put him behind in the final figures was weighting by turnout and squeezing the don’t knows. Neither of these are strange and unusual, they are ComRes’s normal method and are perfectly justifiable.

Looking at turnout first, ComRes found that Labour and Conservative voters said they were more likely to vote than UKIP voters. In Survation’s poll in February they actually found just the same thing, and their approach to weighting by likelihood to vote is very similar to ComRes’s (there is a difference in how they treat people who are very unlikely to vote – Survation weight them down very heavily, ComRes exclude them. In practice this difference has minimal effect). The difference between the Survation poll showing a ten point UKIP lead and the ComRes one showing a one point Tory lead is NOT turnout weighting.

The other difference is don’t knows. In the Survation poll people who said they didn’t know how they would vote were ignored. In ComRes, they were asked a follow up “squeeze” question – how would they vote if they legally HAD to. For people who still didn’t give an answer, ComRes asked if they identified with any party, and took that as their most likely vote. In practice these squeeze questions helped Labour and the Conservatives, but didn’t squeeze out much in the way of extra UKIP support.

There is nothing at all methodologically “wrong” with this poll… but then, there wasn’t anything “wrong” with the Survation poll in February either. There are different methodological approaches, and there are good arguments to be made for and against them, but we don’t have the evidence to say which is right. More importantly, a lot of the difference here isn’t because of methodology… it’s just because ComRes found fewer Ukippers and more Labour and Conservative voters than Survation did. Perhaps that’s because UKIP have lost support since February. Perhaps that’s just normal sample variation. We can’t tell, we can only say that South Thanet may still be a tight race after all.


748 Responses to “ComRes poll of Thanet South”

1 11 12 13 14 15
  1. So if there are two blocks Con-LD-DUP against Lab-Progressive alliance,one of them should get a majority,unless there are parties who won’t support neither block? But it seems only UKIP would not support anyone

  2. @OldNat

    I’m not sure your question was redundant in fact.

    The analysis is based exclusively on crossbreaks from the five weekly YouGov polls. The formula calls for a Standard Deviation estimate to be plugged in. As you may recall, I had worked out an estimate for this in the past (at 4%, the equivlent of 8% MoE). If I were to try mixing in full polls as well, it would not be at all clear what SD figure to use.

    My personal view is that this particular analysis doesn’t add a great deal to what we could already see from your rolling averages. All it does is pick up a couple of real kinks in the VI trajectory over the last few weeks. Short-lived changes are liable to disappear in rolling averages.

    But the indications are that the short-lived changes are not accumulating in any way. So your rolling averages provide as good a picture as any.

  3. Are all the normal polls going to be a day late because of the bank holiday?

    And I’m wondering when the usual torrent of polls is going to start. Soon I hope, this is getting a bit boring and frustrating when you are looking for trends and there’s only one poll to look at. It’s pointless looking for any trend from one poll, of course.

  4. In the most recent ten polls the Labour VI has ranged from 32-35

    The Conservative VI has ranged from 31-37

    But the combined UKIP/Con VI has only ranged from 46-49

    This suggests to me that the higher volatility of Con VI shares may be due to voters going back and forth between Con and UKIP.

    Therefore the target for the Cons for the last month looks pretty clear to me. Get 3 or 4% of the UKIP VI and the Cons will have the most votes and possibly the most seats.

  5. Unicorn

    Ta.

    My weekly summaries also have the advantage of being like that old jacket that my wife often tries to throw out – reassuringly boring. :-)

  6. @FarehamGrecian

    “Populus poll shows 5.3% swing to Lab in E&W. That’s Lab majority territory.”

    Murphy has just got to sort those pesky Jocks out and it all starts to look rather tasty!

    :-)

    Now that YouGov appear to have reverted to the mean after some rather extraordinary results last week, the polls are starting to support my hunch that UKIP will hold steady, the LibDems will recover a little and the combined scores for the two main parties are very unlikely to exceed the 2010 aggregate of 66%.

    I can only see the smaller parties gaining as this campaign continues. Put crudely, in this political era of disillusionment and alienation from the mainstream, the more voters see of the non-mainstream politicians, the more they’ll like them whereas I fear the opposite may be true for Labour and the Tories.

    The one thing that may change my hypothesis is a complete shift in the tone and approach from one or other of the two main parties, but having read insider accounts of their respective campaign strategies, I don’t think any of us should hold our breath.

    Bacon sarnies and orchestrated press monsterings all the way to the finishing line on May 7th seems to be what we should expect. Dear oh dear.

    Nige and Nicola will be happy though.

  7. @thesheep –thanks, I’ll look into that.
    @unicorn — thanks for that information.
    best, andyo

  8. Omni
    Good graph, thanks. What emerges to me that is swingback is always thought of as Labour to Conservative but that is not what this graph shews. What it shews is, from Oct 14, swingback from UKIP and Green to *all* three other parties, albeit glacial in the case of Labour and LD. Given these are net changes (see Spearmint churn analysis), it is visually(stress visually) a straight forward switch from UKIP to Con and (from Jan 15) from Green to LD.

    The overall impression is that if the trends continued, a few more LD seats would be saved and it would remain status quo Lab to Con. However the graphs presumably do not factor in SNP, assuming these are GB poll graphs. Labour’s chances in E and W are thus masked by the graphs.

  9. CB11
    See my early post this morning please.

  10. OMNISHAMBLES

    ‘It is fairly clear that swingback has been happening for around 2 months now. You need to look at the trends:
    http://www.electionforecast.co.uk/graphics/2015_pool_top_5.svg
    As you can see the coalition parties have been steadily gaining VI for a while. You can see it from the simple rolling average of polls and the EWMA charts above. No one can dispute swingback is happening.
    The debate on here IIRC was about crossover. And the real question is not if swingback is real, it’s will there be enough swingback to keep the coalition parties in power or not.’

    Yes my mistake I was thinking about cross over. Your graph shows what we are seeing, conservatives have risen a little in the last 3 months or so, (Labour has as well but not so much) but in the last 3 weeks or so no real change. Certainly no evidence of cross over and swing back looks to have stalled for now.
    It was a half serious question of mine people were asking how do we know when crossover occurs. Always thought that it depended on what people understood by the term crossover. For some it meant a permaneant state of cross over which would take us up to the election. For others it could be a temporary blip.
    But as valid would be to ask when can we identify when the polls have turned definitively in favour of Labour. Like cross over, if you take the view it needs to be permaneant, I suspect we would only really know when we get to the election.

  11. @andyo

    You can start here for Bayes, and then continue if you still want to :-)

    http://betterexplained.com/articles/an-intuitive-and-short-explanation-of-bayes-theorem/

  12. What happened to the big 2 squeeze in the rest?

  13. @bristolianhoward

    I’ve never heard of swingback being thought of as Labour to Conservative. I guess it would be true in 1950s, but for me swingback is simply the governing party(s) gaining VI from anywhere.

    The situation with Scotland and the SNP actually make it easier to read those graphs. If Labour still had Scotland, an identical Labour/Tory voteshare would mean more seats for Labour. But because of the SNP and a number of other factors (like higher population increase in Labour constituencies), an identical Labour/Tory voteshare means a roughly identical number of seats. Roughly. This is unusual, but it makes it a little easier to see what those VI graphs would mean in terms of seats.

    Here is a graph that shows you what I mean
    http://www.ncpolitics.uk/wp-content/uploads/2014/12/skew3.jpg

    And the article
    http://www.ncpolitics.uk/2014/12/the-skew-aka-electoral-bias-which-party.html/

  14. @Mike N

    Working out when crossover happens.

    On Sunday night @Peter Ould posted this fascinating EWMA chart which provides a partial answer to this question.

    If the Lower Boundary (in red) lies above zero, this means that the Tories are reliably ahead of Labour. On the other hand, if the Upper Bound (in green) lies below zero, this puts Labour ahead.

    If you look at the graph you will see that Labour were reliably ahead up to Jan 26th and again for a bit after March 23. Between these two dates there was a period where Labour was first a whisker short of being ahead followed by a few days in which the Tories just fell a little short of being ahead.

    So, these very sensitive methods show that there has been no reliable crossover to date, but that it was very close indeed in mid March.

    My understanding is that @Peter Ould’s analysis is based on YouGov polls alone. You might get a different outcome if you threw in data from all the different pollsters. I suspect not, though, as this would add considerably to variance.

    So, the answer to your question is: “Not yet..but perhaps quite soon…”

  15. @Laszlo @Andyo

    For a more political take on Bayes theorem you can always try this: https://duncan3ross.wordpress.com/2014/11/26/why-david-cameron-is-wrong-the-maths/ in which I apply Bayes to detecting terrorists…

  16. Unicorn
    Thanks.
    “these very sensitive methods show that there has been no reliable crossover to date” which sounds/feels about right.
    And this is just YG.

    NeilJ
    “…I was thinking about cross over.”
    Me too.

  17. @ The Sheep

    Thanks. Yes, that awkward false negative that Fisher completely missed – and ignored even when he was told (probably because it would had interfered with his race based euthanasia proposal).

  18. ALEC
    Thanks for the two links.
    In the first, there is only really one statement which is relevant to my query:
    There are also differences in terms of the work people do, with those who do routine and manual work tending to have fewer years in receipt of State Pension than managerial or professional workers.

    PP 19-20 of your 2nd link has a good statement of the facts, with:
    At the moment (table 3-4), men born in Scotland have a slightly lower life expectancy than the rest of the UK (75.3 years compared to 77.7) but a substantially lower healthy life expectancy (60.1 years compared to 63). At the point of retirement,
    overall life expectancy is similar (16.4 and 17.6 years) but Scottish men can only expect 8.8 years of healthy life. A practical consequence of this is that expenditure on disability benefits is 22% higher than in the UK as a whole but at the moment
    pensions spend is only 4% higher[51].
    In combination with a closer study of the demographic profile of Scotland (figure 3-1), this starts to identify the reasons why an independent Scotland needs to expand its population if it is to meet its pensions liability. Unless the current situation is
    addressed, the proportion of retired people will become substantially higher than it is in the UK and the effect will be compounded by ongoing health problems in the working age population.

    Overall, I’d agree fully with the 2nd part of that, which is one of the more important reasons why the Scottish Government has been trying to get Westminster approval to allow more immigration into Scotland since 2007. I’m not so sure about the
    first part, though. As a 65+ pensioner, fortunately still in the “healthy life” stage, I’d regard current plans to cut that overall expectancy by 10% as pretty dire.

    Although I’ve been mainly an expat, I’ve had a number of UK employments and as a result had a few different pension pots. The first one was an occupational scheme I took early at 55, when I became semi-retired, and as with most schemes made no
    distinctions in the type of work performed when in service. Equally, the one pot I haven’t yet taken, partly because of the new rules which have just become law, allows me to seek an annuity, but none of the proposals I have received require any
    information on health issues beyond any pre-existing conditions.

    Yet actuarial tables somewhere must show that, for example, former miners or building workers who reached adulthood in the 70s or even the early 80s, will have been subject to working conditions no longer tolerated and quite possibly life-limiting but
    are the ones just approaching the newly increased pension ages whilst fearing that they’ll disappear into the distance. Whatever else we disagree on, I suspect that neither of us approves Westminster’s funding of everything out of current income, but
    that’s what has put us where we are.

  19. In the past it was normal for the election date to be announced a calendar month before polling day – in 1983 the election was declared on May 9th with polling on June 9th with the same pattern in 1987. Campaigning,however, did not really get under way for a further ten days – ie just three weeks from election day. Yet here we are a week into the campaign with more than four weeks left of this nonsense!

  20. Omni
    I just meant swingback ‘in terms of this election’ of course. I see no ‘swingback’ in terms of Green to LD in terms of being a governing party, whatsoever, IMO. UKIP to Conis another story of course..

  21. UNICORN
    Apologies for my failure to switch off bold after the word ‘dropped’.

    I’d be interested in a poll on UKPR posters who wished for a preview button. 100%?

  22. @littleredrock
    i have been plugging that line for a while now,based only on how soft UKIP’S vote is.It has a core yes,but not in the teens percentage wise and that is from a UKIP candidate.
    He reckons 2-3 seats at best,but they want to give the tories a bloody nose in the hope the change to a more right wing leader.His words.
    I think there will be a lot of floating UKIPpers who go into the ballot box wavering about what their vote could do to sway the marginals in favour of Labour.If as you say the tories can get 3-4% back,which is very dooable they will win a small majority.I suspect we will hear much more vote UKIP get Lab in the next few weeks.

  23. Could someone point me in the direction of England only polls please?

    Thank you.

  24. CROSSBAT11
    Murphy has just got to sort those pesky Jocks out and it all starts to look rather tasty!

    Talking of which, does anyone know if anyone has commissioned polling for tonight’s Scottish “debate”?

    NB to anyone who hasn’t looked at the STV page linked above, the “debate” will only be between Con, LD, LiS & SNP. The BBC Scotland one will have SGP & UKIP, too.

  25. @GRAHAM

    you’re right; it is exhausting; it is also the problem with all fixed term elections – look at the USA. The trouble is handing the initiative of fixing the date of an election to a PM effectively makes it a matter of fixing the election to the narrows of party advantage. Perhaps we should have a draw with the Speaker pulling the winning ticket for the date from a tricorn hat – just think how happy Mr Bercow would be to discomfit the treasury bench….

  26. Barbazenzero

    I’d be interested in a poll on UKPR posters who wished for a preview button. 100%?
    ——————-
    Or an edit button after posting

  27. @Barbazenzero

    Highly unlikely.

    IIRC there was only one company (ICM) providing a snap poll of the Salmond / Darling debates, and that was probably due to the level of interest throughout the UK.

  28. @ MIBRI

    I don’t think that an additional 3-4% would necessarily give Cons an overall majority.

    And some UKIP waverers will vote Labour.

    So I think largest party is a realistic goal based on this alone. A OM would require something else as well.

    (2-3 UKIP seats seems likely to me whether it polls 10% or 16%)

  29. JAMES
    IIRC there was only one company (ICM) providing a snap poll of the Salmond / Darling debates, and that was probably due to the level of interest throughout the UK.

    I fear you’re correct, unless any of the London media think that Davidson, Murphy & Rennie stand a chance of blunting Sturgeon’s popularity. The figures for the numbers watching will be interesting, though.

  30. @Laszlo, thesheep
    Thanks for those links — I think I can start to get to grips with this now — having avoided it for so long!
    best, andyo

  31. PAUL
    Or an edit button after posting

    That has much to recommend it, but it does allow posters to change their view rather than just correct typos, etc.

  32. apparently my postcard about my holidays is in moderation: it says little, is totally non-partisan and attempts humour: I really should know better!

  33. The debate tonight should have at least one significant UK wide issue:

    Unless I am reading the runes completely wrong, my guess is the question of whether EM & Labour will join with the SNP to vote against a Tory government in a confidence motion if they have the numbers.

    I think Sturgeon will be determined to get that answer from Murphy, But, will Murphy manage to evade answering or will he end up admitting that Labour will vote with the SNP and so lead to a complete meltdown in the Tory press.

  34. Omni

    Many thanks.

    Really enjoying the comments of all you professional pollsters night after night.

    My lay-person gut instinct tells me that if Lab get to 280 (or thereabouts) there will be more than enough anti-Tory helpers on hand in the next House of Commons.

  35. Does anyone know how many SAFE seats have cons and lab right now?

  36. A few months ago I suggested that swingback, although statistically a smooth process – and it has been that at times – it might also happen in lumps. So a bout of anti-swingback (cringe) is not a big shock.

    I still have the two main parties around 280 seats, Conservatives likely to be slightly ahead on seats (say 2-5). If some other parties decide to gang up on them, they can’t actually stay in government even with most votes and most seats. And then the fun begins of trying to sort out an alternative.

    As for UKIP, I don’t think those voters who have come from somewhere else, lent them their votes and are now considering returning will necessarily go back to where they started.

  37. COUPER2802
    Unless I am reading the runes completely wrong, my guess is the question of whether EM & Labour will join with the SNP to vote against a Tory government in a confidence motion if they have the numbers.

    STV say:
    The debate will be in three main sections. In the opening section, the four leaders will face questioning by Bernard Ponsonby.

    In the second segment, the leaders will take turns to make statements and then face questions from the studio audience. The order was chosen by lots with Ruth Davidson going first, followed by Jim Murphy, then Nicola Sturgeon and finally Wille Rennie.

    In the final segment, there will be an open debate, with the party leaders all facing questions from the audience. They will then have the opportunity to make a closing statement. Again, the order was chosen by lots with Jim Murphy going first, Nicola Sturgeon second, Ruth Davidson third and the final statement going to Willie Rennie.

    I think there’s a fair chance Ponsonby will ask Murphy that question, but if he doesn’t then it doesn’t look as though Sturgeon will have the opportunity to ask it, except rhetorically. I’d be surprised if someone in the audience doesn’t manage to squeeze it in somehow, though.

  38. Albert

    None in Scotland or Northern Ireland – can’t help with E&W. :-)

  39. @BarBaZenZero

    That seems a rubbish format.
    Well never mind there is another debate tomorrow which includes the Greens and UKIP, which I was looking forward to for comic relief from the UKIP guy. However, Nicola may have to wait until tomorrow to put Murphy on the spot.

    I also hope someone can pick up Murphy on the way Scotland is funded, that is as a % of spending not taxation. It doesn’t matter how much tax we raise from mansions or bankers if spending is cut, Scottish block grant is cut.

  40. COUPER2802
    That seems a rubbish format.

    Not ideal, certainly. I thought the 2011 debates included an opportunity for the leaders to question each other. Perhaps I’m mistaken or perhaps it was only the BBC one.

  41. Nicola can just interject when Murphy plays the ‘vote SNP and get Tory’ card. Salmond showed in September, if you interrupt on behalf of the audience, you get away with it.

  42. @BarbaZenzero

    Willie Rennie’s a lucky chap. What are the odds on going last twice in a row?

    Also Jim Murphy’s luck continues to fray. Nicola Sturgeon succeeds him twice.

  43. @COUPER2802

    I don’t think that is really the issue. It’s unlikely that Labour will ever vote for the Tories in a Queen’s speech – regardless of how the SNP voted.

    The issue is whether the SNP would support a Labour government by voting for their Queen’s speech?

  44. @MIBRI If as you say the tories can get 3-4% back,which is very dooable they will win a small majority.

    That might be the ‘dooable’ if most UKIP voters were ex-Tories, but there’s a link below to a sample of 7000 people, which suggests that less than 1 in 3 were.

    For the Tories a 3-4% boost from UKIP would only happen if the majority of ex-Tories abandoned UKIP.

    UKIP voters were also asked whether they’d prefer a LAB or CON government and split 28 to 35 – which suggests that there is no huge pool of Tory Kippers ready to turn the election

    http://www2.politicalbetting.com/index.php/archives/2015/04/07/camerons-view-of-ukip-voters-effectively-being-just-tories-on-holiday-not-supported-by-polling/

  45. @WB: “my postcard about my holidays is in moderation: it says little, is totally non-partisan and attempts humour”

    That’s what comes of lying on the beach.

  46. @Rory Hughes

    It is pretty clear SNP will support a Labour’s Queens Speech [1], then try amendments later. Even if Labour gives them nothing they have the politcal tactic of getting left wing Labourites on side supporting their amendments and leaving EM to rely on Tory votes or abstentions.

    For example the minimum wage Labour want £8 an hour, SNP £8.70. SNP can put down that amendment.

    Then for the Holyrood elections SNP have a list of issues where ‘Labour voted with the Tories’ to beat LiS with.

    I don’t think UK Labourites have anything at all to fear from SNP help.

    I think though that SNP would forgo political games for a C & S deal, just my feeling, looking at what they are saying I think Sturgeon & Hosie favour C & S, where they can get workable devolved powers for Holyrood.

    [1] Until Labour puts Trident renewal in that Speech, then its a problem. I think that is not until about 2018 or 19 so maybe that can be kicked into the long grass for now.

  47. RAF

    “What are the odds on going last twice in a row?”

    For the LDs in Scotland the chances are extremely high. :-)

  48. Blair is backing Miliband. Amazed that this place of ‘up to the minute’ political current affairs didn’t speculate on how that might affect polling.

  49. Survation/Jewish Chronicle poll of Jewish voters:

    CON 69% LAB 22% LD 2% UKIP 1%

    Looks like Mike Freer will get a big majority in Finchley & Golders Green; Hendon could be very close despite the small margin Matthew Offord won by in 2010.

    Israel/ Middle East a big issue for Jewish voters and LAB/ Miliband seen as weak on it.

1 11 12 13 14 15