ICM post-debate poll

ICM did a rapid “who won” poll after the Paxman interviews with Cameron and Miliband tonight. Cameron won over Miliband by 54% to 46% – we won’t know until the tables appear, but I expect that going to show Labour voters thinking Miliband won, Conservative voters thinking Cameron won, the rest splitting evenly and not many minds changed. Still, we’ll see if there is any longer term effect in the next week’s polls. I wouldn’t have imagined there will be – while the debate was new in 2010, events like tonight’s Paxman interviews or the leaders’ Question Time style event next month aren’t new – there were things just like that in2005 and 2001 and they made bog all difference to anything.

Meanwhile tonight’s YouGov poll for the Sun has topline figures of CON 36%, LAB 34%, LDEM 7%, UKIP 13%, GRN 5%. A Tory lead today after some ties and Labour leads so far this week, suggesting the underlying average in YouGov’s daily polls is still neck-and-neck.

UPDATE: ICM tables are now up here (thanks to that fine man Tom Clark!). As expected, perceptions of who won fell pretty much along the lines of pre-existing party support – 84% of Conservatives thought Cameron won, 74% of Labour supporters thought Miliband won, Lib Dems were split, Ukippers thought Cameron won, Greens thought Miliband won. Note that while the sample was demographically and politically weighted to be nationally representative, it was a very heavily Labour sample in terms of current voting intention: the pre-debate voting intentions of the sample had an 10% Labour lead (thus are the difficulties of doing things like this – people who watch programmes like this are different from your average voter!)

365 Responses to “ICM post-debate poll”

1 4 5 6 7 8
  1. Considering how much everyone “enjoyed” the beebs political spectrum centre test a week or so ago I felt it might be good to share another.


    It may have already been pointed out by others especially in the run up to the last election but they have a new site for this election and I thought it might amuse some of you.

    FWIW I got Green 50% Labour 40% Lib Dem 10%

  2. Has anyone else noted that what pushes up the LD vote up in the March 25th -26th poll is the LD value of 15% on the Scottish crossbreak?

    Right next to it is an England and Wales LD value of 6%. So a 15% Scottish LD value pushes LD up to 7%.

    So YouGov and Pannelbase, for example, are not far apart if you remove the YouGov Scottish crossbreak anomaly, which if I understand the Scottish local government by-election results is an aberration of LD support anyway.

    @ OLD NAT

    Where are you getting your stats from? Can you post the url please?

  3. The story from the uninteresting constituency in the Liverpool corner of the NW continues.

    Visits so far: Labour and Green (neither will visit again, so they said).

    Leaflets: Green 2, Labour 1 – it is because a new Green leaflet (actually two in one) arrived today. Quality is better than the previous, and much more misleading. It keeps on moving from 2010 elections, local elections and EU elections depending on the suitability of the data, also referring to the fact that they are the second largest party in the council – true, Labour 78, Green 4, LibDem 3, Liberal 3, independent 2 (they even used the LibDem style column charts – the LibDems should have copyrighted those)

    The Labour leaflet is purely local, about the achievements of the local councillor.

    I feel being left out of the General Elections of 2015.

  4. Over 400 complaints now made re-pro Cameron bias in last night’s programme – particularly from Burley who kept asking Milliband supplementary questions without having treated Cameron likewise. She even threw a question at him that an audience member had earlier directed to Cameron.

  5. Andy Shadrack

    Which stats?

  6. Rivers10

    I got 83% SNP : 17% SGP..

    Mind you, it wasn’t hard to ignore the party policies which concentrated on English issues, in a supposedly Scottish data set!

  7. I cannot say I had the stomach to watch the interviews end to end – but despite what is said to be Media bias I thought Miliband was relatively fluid – particularly in the SKY bit I saw….

    And I honestly can say I do not think it was either a debate in any meaningful sense or will change any minds on either side. I think these sorts of political interviews and Town Hall style meetings with the public are such an established formula they are pretty well factored into the public’s mind and tend to attract an audience accordingly….

    In the end next week make have more impact – simply because of the novelty of both debate forum and the number of leaders – and the fact it has Farage max factor – which will garner Media attention.

    I’m pretty much convinced the public dislikes interviewees being harried and interrupted by the headlining star interviewer – and for all it makes for a Colosseum-like entertainment – it hardly persuades by argument.

    The problem for the Media is fewer people are switched on and into the contrived studio fora that thrived in the first twenty five years of the TV interview. I suspect a lot of folk seek out information on the Internet in ways that persuade by association. The Obama campaign in 2012 was good at reaching a target audience through Facebook – friends of friends- tennis fan to tennis fan- baseball lover to baseball lover – opera goer to opera goer – etc etc.

  8. JAMES
    No computer tables, but the presentation slides are here

    Many thanks, and some interesting data. P37 of the PDF seems to explain the SNP surge pretty well and why Unionists weren’t expecting it.

    But overall it’s a pity that there isn’t any weighting info or at least a table comparing the sample to the actual total votes each way.

    Probably worth some searching for “Ailsa Henderson University of Edinburgh” who is shown as the author of the Powerpoint presentation from which the PDF was created.

  9. @rivers10

    Thanks that was fun. I got:

    Conservatives: 40%
    Liberal Democrats: 30%
    Labour: 30%

  10. @ OLD NAT

    Yesterday’s West Lothian by-election results. Their municipal website has not posted the results yet.

    PS: It looks like I will be in the UK to work on an election campaign by around April 3rd, so I’ll be able to see everything first hand then. I have been asked to come over, as I kept my UK citizenship.

  11. PS to my 16:01 WET post. Ms Henderson’s blog post refers to the study. See here.

  12. Andy Shadrack

    The Panelbase poll did weight for 2010 recalled voting All the results were weighted by age, sex, region and SEG and voters were also weighted to match the 2010 UK election result.

    They do have 2010 and 2005 votes on file for their panel, though you don’t need to fill them in on joining.

    This is the first GB-wide poll[1] involving voting intention they have done[2] so it’s a field they’re not terribly experienced in. All their previous polling has been regional for Scotland or the North East (they’re based in Hexham and the only major pollster not mainly located in London).

    So it will be a while before we get any idea of what house effects etc they have and the low Lib Dem VI may simply be an accident. We don’t know if this will even be a regular thing as there doesn’t seem to be a commissioner for it. Non-regular political pollsters often do a bit of it in the run up to general elections to advertise their availability and raise their profile (We even had one from India in the run up to 2010 – oddly enough they got the closest).

    [1] Strictly speaking it’s also the first UK-wide (Martyn must have fainted with pleasure), though I can’t really see the point without regional cross-breaks.

    [2] They did do some Scotland v rUK attitude research for Wings earlier in the year:


    which had did have crossbreaks and weighting based on 2011 Holyrood and 2010 Westminster respectively.

  13. The Guardian website has a chart of no less than 12 different forecasts for the election result from different groups and organisations.

    The range is:
    Con 233 to 296
    Lab 261 to 312
    LD 10 to 48
    SNP average 41

    Not a lot of consensus in the forecasting trade

    Providing a link got me on the naughty step!

  14. @Statgeek

    Sorry bout that. I’m blaming my new computer. Everything is different and that includes Excel, Word et al. I’m missing shortcuts and icons etc. I made the mistake of starting with the new data, then having the old…I had a feeling that would muck things up.

    You haven’t been ‘upgraded’ to Office 2013 have you?

    I got that for the first time, with my new job last September.

    I initially hated it, but our relationship has got worse sinve that zenith.

  15. @johnb160

    This is something similar that comes from the LSE forecasting conference happening right now


  16. Graham

    The BBC continue to show ever increasing bias in all there reporting against DC and Cons.

    eg from a Radio 4 report ” The Lab party have announced there will be no NI increase if they win the GE. This follows the announcement there will be no VAT increase, now also stated by the Cons.”

    The BBC constitution states they should be impartial in there election coverage. Ignored by most presenters.

  17. “Not a lot of consensus in the forecasting trade”

    Rallings and Thrasher I like very much. not only because they are roughly where i am in terms of seats but because they are the only ones to use local elections which, even though they are skewed and give only a fragmentary picture, are at least real elections.

    Nate Silver the god of all this stuff always used local elections or their equivalent.

  18. Andy Shadrack

    I hope you’ll visit Scotland when you are over.

    Council websites don’t usually put up the official reports till the next week.

    @britainelects gathers all kind of election results, though they are really poor at reporting on Scottish local elections, as they try to force the (effectively AV) by elections in an STV system into a FPTP model.

    For by elections here, I dig out the latest previous election results from the council websites, then compare them to the actual results tweeted from the count.

  19. Hurrah, a BBC bias thread. Anthony is going to love this.

    @ Catty and Statty,

    You haven’t been ‘upgraded’ to Office 2013 have you?

    This is why I’ve started using OpenOffice. Is it slightly unstable and does it mysteriously delete saved files sometimes when it crashes? Hell yes it does, as the Great Leader would say. And yet it’s still a better piece of software than Office 2013.

  20. @ Pete B

    Yes – I have been outed.

    This will be an intersting election for me, as I have one relative voting Conservative, another considering UKIP and a third reputedly thinking of LD or Green.

    We all started out as Conservatives.

  21. Last night’s television.

    Very few viewers
    No slips
    No gaffes
    No knockout blows

    and so, no change.

    Another opportunity for the Conservative party to kickstart swingback gone. And less than six weeks to go. At what point do they start getting bolder?

  22. @Spearmint

    Domestically, I’m on LIbreOffice (okish). It seems stable on a Linux platform.

    The stats add-on in Excel is good, but I discovered the open source ‘Sofastats’, that is reasonable, even if the layout a touch unfriendly.

    I’d love Minitab, but at £1236, no way.

  23. If UKIP is leaking voters to the Conservatives, as some seem to think, does that mean that Farage will go for Cameron at the 7 leaders’ debate? Or will he be frightened that that will aid Miliband too much?

    It would be very interesting to know what Farage’s ideal election outcome is.


    Everyone knows the BBC is biased, they are just not sure in which direction.

    I did their little quiz and am 100% Tory. It is rather easy to identify each party’s policy so I could have chosen to be anything I wanted. Batching policies together is not at all useful as there are some things I could have chosen from each offering.

  25. Little Red Rock

    Good point. For now Con are UKIPs biggest rival as it is in the main Con voters who are switching to UKIP.
    So I think Farrage is more likely to attack Cameron than Miliband. I also think a Lab Govt might suit UKIP more as they can build up their base in readiness for 2020.

    Its all becoming a bit Machiavellian.

  26. Spearmint,

    To be Fair it was a response to the Burley was biased view which I won’t add to as it would be unhelpful and Anthony would be disappointed.

  27. FWIW,

    I think Farage wants an outcome that enables the UKIP to demand a referendum on EU membership in the Autumn as a price for C&S.

  28. OLDNAT
    just a question, but if ‘yes’ won the vote across the board in 16 to 49 year olds and lost it in 50+ year olds by fairly similar margins, how,with very high turnouts across the board, did ‘yes’ lose the overall vote 45:55 when the 16 to 49 year olds make up more of the electorate than the 50+’s do?

    Office 2013 …. I initially hated it, but our relationship has got worse sinve that zenith.


    I started my love/hate relationship with Office 95 pro running on Windows NT 3.51 when it first became available. As one of the earliest desktop SQL capable applications it survives today with many specialist software builders still using Access 95 databases because of their virtual indestructibility when working with the DAO.

    “Improvements” up to and including Office 2003 were generally worthwhile but I don’t recommend anyone moving to Office 2007 or newer without exceptionally good reasons to do so.


    I think Farage will big up Miliband because he wants Labour’s anti EU vote and doesn’t want to alienate them. His core Tory constituency is composed entirely of Cameron haters so he won’t lose any of them by siding with Miliband. It is all getting a little too transparent and will eventually cost him because he will lose Tory floaters.

  31. I was struck by the ‘under promise, over deliver’ mantra from Miliband. That’s surely one of the biggest cop-outs any politician could say before an election. Cameron looked as nervous as I can ever remember him.

    Anyway, I found the whole spectacle rather distasteful, crass, and hugely contrived. Paxman was Paxman, and Kay Burley was…well, Kay Burley.

  32. Spearmint –

    “Hurrah, a BBC bias thread. Anthony is going to love this.”

    Indeed – thanks Spearmint. To those who don’t know, BBC bias is one of those issues we do NOT discuss here. Past experience suggests it’s not something that its possible to discuss in a non-partisan way. Let us just assume that everyone thinks it is either neutral, or biased against the party they support, and move on.

  33. David Colby

    I suggest you ask the University of Edinburgh if you are confused about their figures.

  34. @JimJam

    If that is right he is aiming for a pretty small target i.e. Tories plus Non UKIP friends = approx. 318 – 322 or he is hoping to get a lot of MPs.

    If his aim is to take a big chunk of the Conservative Party (from MPs to local activists) then he probably needs DC to stay as leader since it is the antipathy towards DC that would be a big part of any defectors’ motivation.

    Does anyone think that DC plans to spend five years as leader of the opposition? It seems very unlikely to me.

    So Farage probably wants Cameron to limp on, aggravating the right wing of the Conservative party, possibly prevented from calling a referendum by the Lib Dems but needs UKIP to do sufficiently well that his position as leader is secure. And he needs to win Thanet South.

  35. Survation have a poll on English devolution for Labour List – available via this Labour List page –


    “62% of respondents agree that too much of England is run from London (compared to just 13% who disagree). Nearly three-quarters (73%) of respondents in the North and over two-thirds (69%) of those in the Midlands believe that too much of England is run from London, compared to just 53% in the South and 47% in London.”

    This is reminiscent of the thinking within LiS during the Thatcher years which led to legislative devolution “to protect Scotland from the Tories”.

  36. @ Roger Mexico

    Thanks for all the background information, but you do need to consider that the Panelbase LD 2010 voter value gave a 2015 respondent value of 23 and YouGov’s unweighted value was 24.

    Then there is the fact that the March 25th to 26th YouGov poll has an English crossbreak of 6 for LD, while Lord A has them on 8.

    A party cannot simultaneously go in two directions at once. One of them has to be wrong. And depending on which one is right, it may make or break which of Labour or Conservative has the largest number of seats in the HoC,

    Finally at 5.1% at the end of the third week in March that is a 30.1% decline since the first week of February. I am not saying that cannot happen to a Green Party, but I have not seen it happen to a Green Party where the membership is rising from 30,000 around the December Solstice to 57,800 by around the middle of March.

    The two trends are going in polar opposite directions, that’s what perplexes me – and I see more ground strength in a campaign that is now running 550 candidates UK wide as compared to 335 in 2010.

    If the pollsters are overstating Labour, as some on this list keep saying, then at who’s expense? Green, LD or Conservative?

    As ever I appreciate your patient comments back to me.

  37. @ Little Red Rock,

    No knockout blows

    Au contraire– we got a killer new pledge! Katie Hopkins vowed to leave the country if Miliband becomes Prime Minister.

    I reckon that takes Labour well into “workable majority” territory.

  38. @ Spearmint

    600+ seats I should think.

    I call it the Paul Daniels factor.

  39. Spearmint.

    Paul Daniels still here!!

  40. JimJam

    The greatest broken promise tragedy of them all.

  41. TV debate saw no discernible change to GE betting odds:

    Most Seats:
    Tories: 2/5
    Labour: 15/8

    Next PM:
    Cameron: 1/2
    Miliband: 11/8

  42. @Old Nat – yours of 2.53

    Re: Armadale & surrounding hamlets

    Many thanks for the %s info. Not much sign of enthusiasm for Labour in an area which, for many years, was strong mining country.

  43. @Richard

    It helps in the sense that the issue is the chart format dimensions. It doesn’t in the sense that the specific solution applies to Excel 2003 and I’m on Excel 2007, and that solution doesn’t seem to work for me.

    In addition the vba needs to apply to charts which take up the whole chartsheet, rather than ones which are inserted into worksheets. It’s an activesheet / activeobject thing I think. I haven’t figured out the syntax yet.

    Many thanks for the heads up though. If push comes to shove I can just change the dimensions on each chart.


    Nah, it’s Excel 2007. I’m moved from Vista to Win7 and I had forgotten to take the qat files with me, so my old shortcuts and so on are missing. As a result, I go the long way round on many regular tasks, and am just needing a few days of practice.

    Notice how I change to Win7 in 2015…I’m way behind the curve, and I like it that way. Let others solve the problems of IT compatibility. I pay for my product. I don’t pay to test it for MS. :))

  44. OLDNAT

    For those people who believe in English Devolution, the Referendum in the North East in the 2000s is a dreadful set-back, because it seemed to be such an emphatic rejection of the idea.

  45. @ProfHoward

    Except that was on a specific proposal that was not really embraced by anyone. In the light of what has happened in Scotland and London, and if there were a good proposal, then it might not go the same way again.

    Worth noting that of course that was Northumbrian Devolution, not English Devolution…

  46. If the leak to BBC is anything to go by, and it remains on the screen, the Conservatives have lost the election today (in my opinion).

  47. I guess (although I cannot be sure, of course) that being in a coalition makes leaks even easier.

  48. Laszlo

    Well, somebody has been a naughty boy or girl!

  49. It’s the second today. The first one wouldn’t have much effect (trying to make it difficult for civil servants to pay their subs), and it seemed to come from the LibDems.

    The second one is more serious if one recalls Cameron’s concern for the frail and disabled shown yesterday. It is an ideal billboard.

1 4 5 6 7 8