ICM’s monthly poll for the Guardian has voting intention figures, with changes from their last poll, of CON 39%(-4), LAB 34%(+2), LDEM 19%(+1). It was conducted between the 18th and 20th of April.

It is obviously a sharp reduction in the Tory lead when the polls had appeared to be stabilising with the Conservatives in a double-point lead over Labour. This seems to be counter-intuitive when the Labout government have been having a rather torrid time of it over the abolition of the 10p tax rate, but that doesn’t necessarily mean it wrong.

The poll would also have been influenced by media coverage of Gordon Brown’s visit to the USA. Coverage of Brown looking statesmanlike with other world leaders could have boosted him (compare with the polls showing a sharp narrowing of the lead in January 2008 after the Davos conference when the media also had lots of coverage of Gordon Brown looking statesmanlike). Arguing against that, the poll certainly doesn’t show any increase in Gordon Brown’s own ratings: David Cameron now has a lead of 8 points over Brown as best Prime Minister (37% to 29%, with Clegg on 8%).

Rather the most likely reason seems to be an increase in economic confidence – 55% of people in this poll said they were confident about their own financial situation, compared to 48% in February when ICM last asked it. On Friday we should have to chance to see if the boost in economic confidence or the recovery in Labour support are echoed by YouGov’s monthly poll.

UPDATE: In a comment on a post below Mark Senior asked whether having the fieldwork for this poll over a weekend may have made a difference. Well, actually the majority of ICM’s polls for the Guardian are carried out over the weekend so this isn’t unusual at all. Does it make a difference? Well, in theory it could if it produced samples that were different in a way that weighting didn’t correct. As we’ve commented here in the past, for the last 6 months or so ICM’s polls for the Guardian have had a tendency to show lower Tory leads than their ones for other clients, something that seems to be pure co-incidence given they are conducted in exactly the same way (confirmed by Nick Sparrow, ICM’s boss). I initially thought it might have connected with ICM’s Guardian polls normally being conducted over the weekend and their Sunday Telegraph polls being done midweek – but it also held true when Guardian polls were done mid-week. I suppose it’s still possible there could be that there is a midweek vs. weekend fieldwork difference, with pure co-incidence explaining the midweek Guardian polls.

40 Responses to “ICM show sharp drop in Tory lead”

  1. Might be an oversight but you don’t say what the poll scores are!

  2. Whoops :)

    I write them in notepad and cut and paste them in – must have failed to select all the text :)

  3. Certainly a surprise to those of us who have been feeling a little beleagured recently, hard to explain this away against what seems to be relentless bad news/criticism of the government in the media. If it is not a complete rogue, and if Boris fails against Ken on May 1st then people will surely ask why Cameron has failed to capitalise on events.

  4. waiting for someone to say it must be a rogue poll – shall I? “It must be a rogue poll”

    especially as Yougov now do about 20 polls per month meaning that theirs are not rogue

  5. thanks for the clarification Anthony , for some reason I was under the misapprehension that the Guardian polls were done early in the week and usually released on a Thursday . Although the changes look dramatic the 3 latest Guardian polls are within the M of E of an average of them so no rogues but perhaps some outlier(s) .

  6. Anthony , note my comment on another thread re the change in Populus weighting for past vote in their latest poll .

  7. If Conservative support really is collapsing for whatever reason you’d think the Lib Dems would be doing better. 19% is okay but down on recent elections. There seems to me no good reason why Labour would be picking up more than them.

  8. Here we go again!

    What will YouGov’s next Poll say?

  9. Nowadays we seem to believe polling is a very exact science, but the enormous fluctuations on an almost poll-by-poll basis (16 point Tory lead verses 5 point lead) must either mean that polling techniques are less accurate than we think or 10% or so of the population change their minds on an almost daily basis.

  10. Looking back at ICM/Guardian polls since last October they have consistently given a Tory lead ranging from 2 points to 6. Only last month did they a larger lead therefore they have gone back to their ‘norm’ and it could be argued that last months poll was a rogue, for them, and not this one

  11. Anthony,

    Given people concerns about poll volatility, for want of a better term, have there been any trends across the polls in terms of “Don’t Knows” (DK)?

    If the % of DK’s was rising showing that a growing number weren’t sure who to vote for then that same feeling could create greater churn with voters more likely to swing from one part to another based on the current news.

    After a decade of triangulation a combination of “there all just the same” or ” I can’t decide which is best” may mean that polls will jump about until closer to when an election is due or when one is actually called.


  12. WMA 41:31:18 CLead 9 (rounding). Poll makes no sense in the light of what is actually going on in politics, but recent ICM polls have tended to be fairly accurate compared with Retrospectives so we’ll have to see.

  13. ANTHONY said :- “Coverage of Brown looking statesmanlike with other world leaders could have boosted him” –

    This was surely a glib comment – he looked like a bumbling fool (reminiscent of Michael Foot on his excursions). Watching him with George Bush actually made Bush look more staesmanlike !! Even the media agreed with my view of Gordon Brown on his mistimed visit along with the Pope – another gaff !

    Oh yeah – the POLL from ICM is definately a ROGUE POLL / I can’t understand why the media still bother with some of these poor POLLSTERS .

    The one good point about rogue POLLS showing Labour creeping back up so close to an election means that the Labour voters will feel more relaxed and stay at home .

  14. From the text:

    Today’s poll suggests that Alistair Darling is a liability for Labour as chancellor: only 26% of voters think that he is the best chancellor of the three figures from each party. But his support is stronger than for the Conservative shadow chancellor, George Osborne, picked by only 20% of voters. Backing for Vincent Cable, from the Liberal Democrats, stands at 13%.

    That doesn’t make sense, does it? Unless they mean darling is a liabilty because his approval rating is less than his party’s. Even then though it’s a bit strange.

  15. Lukw.

    I agree entirely-and I think the former of your two suggestions is the more likely.

  16. Paul,

    Any poll that shows major shifts with not obvious reason why should be treated with caution. More so when you have the likes of Michael Crick on Newsnight suggesting that it was part of a gap narrowing trend since the budget.


    Factor out the extremes and the polls show a similar pattern of a Tory lead around 8%.

  17. Mike – (a) tone done partisan comments please, (b) “rogue polls” do not reflect badly on individual pollsters. A rogue poll is just a term for the 5% of polls that fall outside the 95% confidence margin, by definition 5% of every pollster’s polls will be “rogue polls”.

    Even if a pollster suspects that a figure is the result of normal sample error, they have little choice but to publish them or otherwise people would accuse them of hushing up results they didn’t like. Imagine how you would react if a newspaper didn’t publish a poll showing a 20 point Tory lead because they privately decided it was probably a rogue poll :)

  18. Matthew. I don’t think their conclusion can possibility be supported by those figures. Approval ratings for Alistair Darling suggest he probably is a liability, but you really can’t reach that conclusion from looking at a question showing people rate him as the best out of the three main parties treasury spokesmen!

  19. Adrian – there may be some differences between ICM’s polls for different clients if the days of the week the fieldwork was one has an effect, but it would be a minimal one. Last month ICM did three published political polls that all showed a sharply increased Conservative lead. The chances of all three being strange outliers produced by random chance are very low indeed.

    Only in hindsight will we be able to make a good guess about whether this poll is a outlier or reflects a genuine narrowing of the lead, but I think we can be pretty confident that last month’s were showing a genuine shift in party support following the budget. The only question is whether it is now being sustained or if it is fading away.

  20. I don’t get why people polled over the weekend are more likely to say they are voting Labour. What is the theory behind that explanation for this poll?

  21. “Today’s poll suggests that Alistair Darling is a liability for Labour as chancellor: only 26% of voters think that he is the best chancellor of the three figures from each party.”

    Being a popular chancellor, especially in difficult times, is a rare thing.

  22. How do we find out the single-digit scores of the minor parties? I’m interested to know how the Greens did. Perhaps it’s not statistically significant, but still.

  23. Anthony

    With reference to the Oracle’s comments (which are all basically the same) and your mild admonition others of us could comment that Cameron looks like a lightweight Blair clone or Clegg looks like an apprentice PR man—but we don’t generally. However I don’t think the Oracle should be moderated as all blogs need a bit of spicing up from time to time as long as it doesn’t stray into the realms of abuse.

  24. Anthony,

    Is there a MRS policy on polls that seem rogue?

  25. Paul – look at ICM’s polls since the Conservatives regained the lead in October. Their polls for the Guardian have tended to show the Conservatives 2 points lower than their polls for every other client.

    I’ve checked with Nick Sparrow and the polls are done in exactly the same way, there is no methodological difference between them that could explain the difference.

    The most plausible explanation I can come up with – other than pure co-incidence – is that ICM polls for the Guardian tend to be done over the weekend, while their polls for the Sunday papers tend to be done in the week.

  26. Anthony:

    Presumably pollsters don’t reveal the identity of the client to the people they are polling?

    But to the people doing the polling know?

  27. Osborne is a liablity.

  28. Don’t think he’s a liability, just not a particularly good man to have as a shadow chancellor, one feels he is somewhat overpromoted at present and his presence next to a very young looking leader is also not helpful. If only they could pursuede Clarkey to come back.

  29. Perhaps maybe the majority of people realise they are better off with the new tax system… never mind the poorest.

    Good news in the short term for Labour, but that thinking is more along the lines of the conservatives policy (or what many think a Tory policy should be anyway) so in the long run it might mean people prefer a less “lefty/care about the poor” approach – which seems about right. I think after 10 years of higher taxes people are starting to go back to the “charity begins at home” way of thinking, especially with the squeeze right now.

  30. even if ICM’s poll is right it still would not give brown majority cameron would be biggest with the LD’s holding the balance of power, i think this may just be a one off poll

  31. Quiet reminder… it’s still at least 18 months to an election. The economy may not be the big election winner that the Tories hope will save them… the opinion polls are still very volatile, even at what might possibly be the low-point for Labour (if Labour MPs keep their nerve and don’t do a repeat of the Tories under Major)… It could all look VERY different by, say, October 2009 or May 2010. So: to Labour supporters (the small band here): don’t despair – To Tory supporters: don’t count your chickens two years too soon – To LD supporters: just remember that you really don’t want Cameron – and to Peter: Och Ay, it’s St George’s day!!

  32. Three cheers for St George, England,s favourite Turk…..


  33. And St Andrew would be from eh, modern day israel I suppose?

  34. “Three cheers for St George, England,s favourite Turk…..”

    Put like that it’s as daft as St. Andrew-favourite Galileean of Scotland, Russia, Romania, Greece, Amalfi, and Luqa in Malta.

  35. St Andrew was chosen by a lot of nations because he was an apostle and missionary who spread Christianity, as opposed to killing a dragon??????


  36. Well, Peter, I could name quite a few dragons who need slaying… but I might get a reprimand from Anthony for being too partisan :)

  37. JohnH,

    Well when your finished with Dragons you can always go on to Windmills.


  38. The Dragon story was coined by the Crusaders.

    St. George was a soldier who rose through the ranks. He was tortured & executed by the Roman Emperor of the time for disobeying an order to persecute Christians.

    He would seem to be a pretty good role model for anyone or any country.

  39. I saw a Tory canvasser on St Georges Day – a sight in recent years as rare as dragons. In fact doubt one’s been seen on this street since Harold Wilson devalued the pound. Strikes and inflation – the good old days.

  40. Yes wolf, they were the OLD days. With inflation now at 2.5 (and the old RPI even went down last month) – maybe we need a better sense of history!