The worst thing you can do in analysing polls of voting intention is to get excited at polls that show something exciting and different and ignore those that show the same old pattern. Occassionally the unusual poll will herald a genuine movement in public opinion – after all, whenever there is a change, one poll has to pick it up first. More often than not, the unusual poll will turn out to be a freak result, the product of unusual sampling or methods. If there is genuinely a change in public opinion, other polls will pick it up sooner or later, so it’s always wise to withhold your judgement.

Today we have one of those unusual polls, and we have the overexcitement you’d expect. ICM’s monthly poll in the Guardian has topline figures of CON 39%(+3), LAB 33%(-2), LDEM 8%(nc), UKIP 7%(-2), GRN 7%(+3) (tabs). This is pretty odd all round – a storming six point lead for the Tories, up on thirty-nine percent; the Greens and UKIP equal on seven percent.

In the Guardian’s write up they are rightly dubious, and include a welcome caveat from ICM’s Martin Boon about the inevitability of random variation and the sample perhaps being a touch too Tory. I’ll just leave it with the usual caveats – it’s one poll, and an odd looking one at that. Sure, it could be the start of some Tory surge, but if it is we will see it echoed in other polls today…and luckily enough we have at least three of them.

Populus this morning had topline figures of CON 33%, LAB 33%, LDEM 8%, UKIP 15%, GRN 5% (tabs). The Conservatives are up two points (possibly helped by an update in weighting targets), but no big tory lead.

Still to come are the weekly Ashcroft poll and the daily YouGov poll. Come the end of the day, the way to judge where we are is too look at all them as a whole – not fixate on the unusual one.

UPDATE: Lord Ashcroft’s weekly poll has topline figures of CON 33%(-3), LAB 33%(-1), LDEM 9%(+3), UKIP 13%(+3), GRN 6%(-1). Changes are from a fortnight ago – Ashcroft took a week off to avoid bank holiday fieldwork. As with today’s Populus poll, there is nothing here to support the big Tory lead in the ICM poll. Full details are on Lord Ashcroft’s website here.

UPDATE2: Finally the daily YouGov poll for the Sun has topline figures of CON 33%, LAB 34%, LD 8%, UKIP 13%, GRN 6% – a one point Labour lead. Putting all four polls together that ICM poll looks very much like an outlier. Such things are an unavoidable part of polling – and well done to Guardian for reporting it in a heavily caveated way within the context of other polls showing no movement, rather than getting all excited about it.

459 Responses to “Latest ICM, Ashcroft and Populus polls”

1 2 3 10
  1. Firstish?

  2. Oh my. *Trying very hard not to fixate on the unusual poll*

  3. how can UKIP go from 18 to 7 in a matter of days, something wrong with the polls there

  4. ~8% is actually what I expect UKIP to get on election day, but I doubt they’re already that low.

  5. @John Smith

    Well, it’s two different polling companies you are comparing for a start.

    Guardian says that (in recent polls at least) UKIP have scored 4.5% less in phone polls than online polls, on average.

  6. High correlation between UKIP VI and Con lead – despite analysis showing that many UKIP are not former Con voters.

    “Vote UKIP get Ed” might just be the way forward for the blues.

  7. @John Smith

    ICM always very low on UKIP. No real evidence to tell us whether phone polls understate them or online overstate. European election polls had the online overstating them a tad but not enough phone polls to judge them.

  8. Perhaps ICM are simply proving that their Wisdom Index is, indeed, more accurate than ICM’s conventional polling?

  9. Scotland cross break from ICM
    Labour 1
    SNP 47
    LibDem 2
    Con 9

    Big Tory revival the tactical voting must be paying off.

  10. Conservatives leading Labour in Scotland? Have I misread that?

  11. Populous shows Ukip on 15% and ICM show Ukip on 7%. One of them is going to be very wrong in 3 weeks time.

  12. i can’t get a clear screen of the data,is their any figures for England and Wales alone on the ICM poll ?

  13. And that 8% has to go / come from somewhere, and it might not be 50:50 Labour, Conservative, so the risk is not just that UKIP % is wrongly forecast, but the whole game.

  14. what I find very odd about the ICM poll is that any major party has polled 39%. Coming on the back of the recent trend showing a shift towards Labour this surely has to be seen as a very strange outlier

  15. Read the commentary from Mr Wells, people !!

  16. Obvious outlier from the Tories’ best pollster.

    Still, nice bit of balance to the Labour six pointer last week – should even up the sidebar again.

    Am I right in thinking ICM’s figures might even get the Tories a tiny majority? (Chortle).

  17. Question for Anthony:

    In the article, ICM mention they are aware of an unusually large number of 2010 Tories in the sample.

    If this is the case, why do polling firms not adapt their methodology of reaching the headline figure to correct this – or is that dodgy mathematics?


  18. Looks like the Lib Dems have sown up the prize (if one is on offer) for worst pun of the campaign.

    From the Guardian live blog, re: Umuna and Balls appearing to contradict Jim Murphy about whether Labour will continue to make some cuts after 2015/15:

    “Now the Scottish Liberal Democrats have responded to “[b]Jim Murphy being Chuka’d under the bus by his colleagues[/b]” (after Ummuna said that “ the leader of the Scottish Labour party will not be in charge of the UK budget” – see 1.30pm), with party President Sir Malcolm Bruce saying: “The left hand of Labour doesn’t know what the other left hand is doing. In their frenzied attempts to try and restore some economic credibility they have only demonstrated their economic incompetence.””

  19. Just eyeballing the recent months polling it seems that ICM have often thrown up odd Tory numbers – usually higher than the other polls conducted around the same time, but, on one occasion at least, also lower. The Labour number seems pretty much with the pack. Looking at last election’s polling ICM was with the pack for the Tories then also. Anyone explanation, clever people?

  20. Sorry folks – it’s my fault.

    In the last thread I mused that Labour seems to be having a decent run, adding ‘this means that tonight’s YG will show a Tory 6% lead’.

    Got the wrong company, but I was right!

  21. Mibri

    C 40 : L 35 L UK 9 : Grn 8 : LD 7 : PC 1

  22. The really barmy thing about ICM, is that it had the Conservatives on 28 and Labour on 33 in their December poll. That was also blatant nonsense. So we now have two screwy polls in the last 5 months,

    If you take the rolling average of the 7 ICM polls since the Indyref (back to October 2014), then Labour are a fraction ahead, which would make this consistent with the normal polling. Of course, the Tory leads are the three most recent polls.

  23. This is the first sensible poll I have seen this year. Especially in Scotland. If only it was an accurate poll. Back to the nail biting.

  24. @neil a
    “Am I right in thinking ICM’s figures might even get the Tories a tiny majority?”

    I think you need a 7%+ Tory lead for an overall majority. So, no. According to an analysis I read, once the Tories get to ~315 seats, it gets difficult to add more. I.e. you need more swing for fewer seat gains past that point.

  25. Very interesting piece on why Conservative campaign appears so odd by John Lanchester on London Review of Books blog.

  26. Couper

    LOL who ever was predicting 7 tory seats for Scotland a few days ago must be feeling quite smug.

    Wrong, but smug.

  27. @Roland

    “You are awful, but I like you”

  28. Ashcroft National Poll, 10-12 April: CON 33%, LAB 33%, LDEM 9%, UKIP 13%, GRN 6%.

  29. Ashcroft:

    CON 33%,
    LAB 33%,
    LDEM 9%,
    UKIP 13%,
    GRN 6%.

  30. Lib Dems and UKIP both up, mainly at the expense of the big two parties.

  31. Why no mention of today’s YouGov poll for The Sun which showed a 3-point Labour lead?

  32. @Neil A On those figures I would forecast about 305 for the Conservatives, but some of the forecasting methods would probably give them about 315-320+ seats.

  33. Let boredom resume.

  34. Well, if ICM balances out Panelbase, Ashcroft balances out Populus.

    Polldrums! ;)

  35. So – not so good for Cons, down 3, suggests that ICM really was a bit of a rogue.

  36. Ashcroft changes

    CON 33 (-3)
    LAB 33 (-1)
    LIB 9 (+3)
    UKIP 13 (+3)
    GRN 6 (-1)

  37. Ashcroft’s tiny (52) Scots sample

    SNP 45 : L 16 : Con 13 : Grn 9 : UKIP 7 : LD 6

    Not quite as daft as ICM, but still meaningless, other than SNP in big lead.

  38. Latest YouGov / The Sun results 12th April – Con 33%, Lab 36%, LD 7%, UKIP 13%, GRN 5%; APP -14


  39. Well here’s something that has changed

    Britain is now heading in the “right direction” as opposed to the wrong direction

    That’s good to know

  40. Toby

    The YG tables were out at 6am today – old news, when there is a flurry of polling.

  41. @ Toby

    We had that last night, got the T-Shirt etc etc!


    That sort of statement makes me think of the ending of Superman (with Christopher Reeve).

    Glad that the UK continues to rotate anti-clockwise on the North Pole.

  43. Toby – that’s a bit old now, came out last night!

  44. Party ratings in Ashcroft Poll among swing voters:

    Labour + 2.60
    Conservative – 2.47

    Funny that the leader ratings are the other way round. I wonder if today’s manifesto launch will have any effect on Miliband’s personal ratings?

  45. Leader ratings in Ashcroft Poll among swing voters:

    Miliband -5.75
    Cameron +0.87

    So, does it come down to people being influenced more by the Party versus the leader of the Party?

  46. Newspapers have hundreds of library photos of the leaders from which to choose. If you look at the front page (on the net) of Cameron and Miliband, what would you say was the editor’s intention? Don’t answer that please, I know you are all too intelligent to require interpretation from me and i don’t want to start partisan exchange This is not unusual I might add, for the Grauniad. I don’t know why they do it, but I have been noticing that they are doing it.

    Perhaps a general point about presentation is safer ground. Is it possible do we think, that press photo images can operate as dog whistles? I am thinking of, as examples, the endless Hague one of him in the baseball cap and Kinnock falling down on Brighton Beach. I am not suggesting partisan behaviour by editors but is it just a juvenile or brainless need to poke fun of political leaders?

  47. I was talking of the Guardian front page.

  48. ICMs a nice Poll.

    If only !

  49. Do we remember the phrase ‘Gold Standard’ for ICM?

    Had I been the ICM chief and / or Guardian’s, I would have ensured that it was either withheld or published with a much stronger disclaimer, just for my own reputation’s sake.

  50. @SWEBB

    For the last two days you have correctly given the YouGov result a few hours before the official release by quoting the vote share figure – before 8pm – from the top of their “Election Centre” page. Do you (or anyone else) know what time of day that figure is being updated?

1 2 3 10