There have been two new polls out today – both show a sharp reduction in the Tory lead.

The weekly ICM poll for the Guardian has topline figures of CON 47%(-1), LAB 33%(+5), LDEM 9%(-1), UKIP 4%(-2), GRN 2%(-1). The Conservative party’s support drops only a little, but Labour jump up five to 33% (their highest since June 2016 according to Martin Boon). The fourteen point lead is larger than most other polls – this is for methodological reasons (ICM’s demographic based turnout model gives a large boost to the Conservative party, otherwise it too would likely have been producing a single-digit lead). Note that ICM have also tweaked their method slightly to hide the option of UKIP for respondents in seats where UKIP aren’t standing, though this will likely have only a small effect. Full tabs are here.

Meanwhile a Survation poll for Good Morning Britain has topline figures of CON 43%(-5), LAB 34%(+5), LDEM 8%(nc), UKIP 4%(nc). Changes are from the previous Survation telephone poll a week ago, rather than their online poll at the weekend. Full tabs are here.

Fieldwork for both polls was over the weekend, after the Conservative manifesto launch. While there was already a general movement towards Labour before the manifestos, the post-manifesto polls suggest a further and sharper movement since then. I wrote a while back about how manifestos rarely have much impact in general elections – while it’s impossible to prove a causal link, the timing certainly suggests this is an exception! Perhaps it’s because when elections are five years in the making most policies have already been announced and focus-grouped into things that won’t scare the horses. Or perhaps just because manifesto launches rarely go as wrong as the Conservative party’s appears to have.

Whatever the reason, the question now is whether this is a temporary narrowing that will reverse when (or if) the focus of attention moves onto other subjects, or whether we are heading for a somewhat tighter race than many people expected.

238 Responses to “Latest ICM and Survation polls”

1 3 4 5
  1. @Anthony Wells

    “We don’t have any evidence to measure it, and our own opinions will be irredeemably skewed by our own political opinions. The last thing we did is a predictable parade of Tory supporters saying they think she came across well and Labour supporters saying they think she struggled.”


    People are not all irredeemably partisan. This is why scientists can make counter-intuitive discoveries that confounded their expectations. Also, he’s talking about the undecided, who may be less invested than the Labour and Conservative voters you compare them to.

  2. Anthony,

    “The last thing we did is a predictable parade of Tory supporters saying they think she came across well and Labour supporters saying they think she struggled.”

    Well as neither, I think she struggled to come across badly!


  3. Jim Stott
    “If the Tories are mostly taking UKIP votes and increasing their lead in leafy shire seats they already hold, that will not do them much good on 8th June”

    UKIP came second in over 100 seats in 2015. Admittedly it was often a distant second, but many of these were Labour seats in the Midlands and North. Therefore gaining UKIP votes there might well see a few Tory gains.

  4. What do people think the chances of a hung parliament are if Labour increases to close the gap and the Conservatives lose points, do you think a hung parliament is a real possibility and if there is one will it be with Lib Dems or the SNP?


    Well as neither, I think she struggled to come across badly!

    Ha. Tactfully put.

  6. Carfrew

    I think both The Guardian and Independent will back Labour this time.

  7. RAF

    Without the turnout model or reallocation of DKs it would have been CON 44%, LAB 36%, LDEM 8%, UKIP 3%

    With the turnout model, but without reallocation of DKs, it would have have been CON 48%, LAB 33%, LDEM 8%, UKIP 3%

    With the turnout model and the reallocation of DKs (ICM’s final topline figures) it was CON 47%, LAB 33%, LDEM 9%, UKIP 4%

    We can’t tell for sure what the result would have been without the turnout model and with the DKs re-allocated, but presumably it would have been something like CON 43%, LAB 36%, LDEM 9%, UKIP 4%.

    And of course, that’s no turnout model *at all* which would be silly. Again, not possible to work out exactly what the figures would have been using ICM’s old turnout model based on people’s answers, but it looks like it would have been something along the lines of CON 45%, LAB 36%, LDEM 8%, UKIP 3%

  8. It is also possible people have different expectations. Some might think that while overall it wasn’t great, played a bad hand well.

    Also, polling usually releases its results quickly, which provides an incentive to game the system.

    Thus, there may be people who are honest in their analysis, they think someone did badly, but they are going to say otherwise because the results will be broadcast to others.

    Which is partisan, but a different kind of partisan. They aren’t fooling themselves, only others.

    The way polling is conducted tends to restrict its diagnostic utility while amplifying its influence. Then they have sum at extra to complain about how fickle we all are!!

  9. @Mike Pearce

    It’s possible but boy have they given Corbyn a bashing. Main reason it’s possible is if they think Corbyn might go easier on the Brexit solution which is overshadowing normal politics as you no doubt know.

    This won’t alter their general liberal outlook, it’s just that they may put liberal concerns over preserving free movement/single market over other things.

  10. @Anthony

    Many thanks. There is no way I could have worked that out for myself!

    I note they also also filtered for whether people were on the electoral roll; and excluded respondents intending to vote Ukip where no Ukip candidate is standing. These seem like very sensible measures. Do you know if YouGov will follow suit?

  11. RAF

    They didn’t exclude those people, just didn’t offer the choice of UKIP (I know it from a not-so-friend friend).

  12. Mike I do not think so.They will both give a long opinion piece about tactical voting .

  13. The Welsh poll looks very interesting.

    I commented the other night that the Labour odds in Delyn and Alyn & Deeside looked very generous. Wish now I’d put my money down!

    Looks like the Tories will do well to hold on to Vale of Clwyd and Gower if this carries on.

    But a week is a long time…..

  14. Carfrew

    Yes agreed although the attacks in The Guardian have lessened in recent weeks and with The Lib Dems way behind in the polls they are now much more positive to Lab.

  15. DEZ. Yes that is a possibility but both papers are certainly attacking Con and any tactical voting opinion piece would be reflective of this.

  16. Mike Independent won’t endore anyone

  17. @Sea Change

    “In a word no. He has an absolute mountain to climb from 23%.
    Clearly, Labour would rather start at 50% odd and now be 44%-47%”


    In several words, you are missing my point.

    Labour might prefer to start from a higher base but this is hard to achieve when if you announce policy early the opposition have time to match it or outflank.

    The government have the advantage in that they can execute policy. This is why they usually get toppled through events, making big errors and so on rather than opposition gambits.

    In 1945 obviously Churchill could match that whirlwind of a manifesto… But after a war, people would be more receptive to policy of rebuilding a better future etc.

  18. I think TM has blown her chance (albeit a small one) to have a Blair sized majority, she will probably still win but with a majority of maybe 20-30. I expect to see polls in the next few days in the 42-38 range which would indicate a hung parliament but I doubt it will hold up and LAB will drop a couple of points to give Cons a small Majority. TM really has to nail Brexit or her Legacy will be screwing up this golden chance for a landslide.

  19. @Laszlo


  20. @Connor123 at the moment a hung parliament is still unlikely. Polling averages still show the Tories comfortably ahead (12-14 points) with outliers at 9 points. On the right hand side of this website are polling histories of previous elections. Note 1987 and 1983 where the incumbent party got large majorities there were 4 and 9 point outliers. Note also how that differs with 1992 and 2015 where the result was much closer. Realistically until the outliers hit 2/3 points we are still in majority terrirory. At the moment it looks like somehwere between 60-120 (rather than 100-150) but still firm. Still 2 weeks to go though so lots to play for and lots can change.

  21. Aargh

    “In 1945 obviously Churchill COULDN’T match that whirlwind of a manifesto”, I meant…

  22. @Mike Pearce

    “Yes agreed although the attacks in The Guardian have lessened in recent weeks and with The Lib Dems way behind in the polls they are now much more positive to Lab.”


    Just taking a look at the Indy they seem to be being a bit more Corbyn-friendly too. A few weeks ago I posted several,headlines from just one issue that where all rather unfriendly.

    Corbyn has picked policies that seem hard to really attack strongly… even on here…

  23. Raf,
    You may have missed my suggestion (so many posts) that the Turnout filter and DK reallocation in effect makes the ICM numbers a forecast based on current responses.
    In this sense their 14 is in line with the other leads around 10.
    Also, in line with some peoples view that come June 08th the gap will be wider than it is now.

    Of course the around 10 could drop some more in which case ICM may well drop by a similar amount but I expect they will have larger les than the other pollsters; although of course MOE could mean it is much larger nearer 8 more or the same.

    FWIW – IMO the TM interview will have negligible impact either way – I like Peter’s ‘ She struggled to come across badly’ with its inherent ambiguities best.

  24. Matt Wilson

    I believe the Indy endorsed the Tories albeit not stridently in 2015 so it may not stay neutral.

  25. Just to confirm the Indy came out in favour of another Con/Lib coalition so it did not remain neutral

  26. I notice that yougov got the present situation wrong when they asked the question on the Conservative care proposals. They failed to say that people in care homes have the whole of their estate, including their home if they have one or a share in one, included in the calculations of their assets now. They will be better off under the new proposals. Has AW any observations?

  27. I have made a revised election forecast based on the YouGov UK poll of yesterday, and the YouGov Wales poll of today (which makes no difference to the result). Conservative majority is 28 seats, with all of their net increase in seats of 7 coming from Scotland. Labour are up two seats on 2015.
    That is the Conservatives, for a 7% increase in vote share gain 7 seats, and Labour does much better under Jeremy Corbyn than under Ed Miliband in 2015 or Gordon Brown in 2010.

  28. @ManicBeancounter


    Fail to win a few seats in Scotland and losexa few in the South to Lib Dems, and the majority is gone.

  29. Watching the 10pm news tonight, I think the script is all going wrong. We are supposed to be moving onto Brexit and the IRA, instead we are still on manifestos.

    In terms of momentum, it points to a further closing of the poll lead.

    But I am nervous that the experts are questioning the extent of poll movements we are seeing, saying they are unprecedented.

    The mantra is people know who they are going to vote for and policies and campaigns don’t change much at all.

    We shall see soon enough…

  30. @JimJam

    Apologies for not seeing your earlier post.

    I did reply to the one above but it got eaten by the site’s firewall. It must have been hungry. Basically, I agree with you on pretty much all points.

  31. Richard

    I think it might be a steady drip feed of Brexit, economy, Corbyn etc rather than an all out assault which might come over at a bit desperate. But the Tory u turn on social care will be tough for the Tories to shake off. But the drip feed strategy might just shift then back up to 47% – 49% which they are not that far away from.

  32. Been catching up posts, so may have missed discussion on this. However, my impression is that May is denying having made a U-turn, and her argument seems to be that there was always going to be a discussion process about details, and a cap might get included at that stage. In other words, the policy remains as written, with no cap.

    I got the impression she was becoming exasperated at keep getting the same questions when she had already told them there was no change to the policy, because there has been no change to the policy.

    Which rather brings us back to where we began this, except May has been damaged by the back and forth as well as by the policy.

    The news suggests that there is considerable division within the tory party, and indeed that May sneaked the manifesto past people. It seems she has been seeking a manifesto with maximum leeway to raise more money and few promises to spend it on anything. This would seem to be no more nor less than a continuation of the row over Hammond’s NI proposals, which got defeated largely by reference to the manifesto. This is a manifesto to prepare the ground for more of the same. It is consistent with Osborne’s position about Brexit.

    Rather than defusing the situation, surely all this is simply placing in starker relief that the tory manifesto is preparation for more and deeper austerity, whereas labour is promising at least some spending relaxation, albeit funded through tax rises. The two are wholly different approaches.

    Wonder what voters think, but this isnt a one off issue over care funding, but an example of contrast between the two parties which will continue to play through to the election.

  33. New thread.

  34. RMJ1,
    my understanding is indeed as yours that people in care homes currently do not enjoy nearly as much protection as under the new proposals. However, most people spend relatively little time in care homes but much more receiving help in their own home, where the value of the home IS protected. People tend to go to care homes only as a last resort, perhaps not least because of the financial implications. However, people going into care homes frequently take a significant turn for the worse and die sooner than those remaining in their own homes, limiting the cost.

    The figures I have seen demonstrate that there would be a net increase of money going to government, and in some cases quite marked amounts. Perhaps also significant is fear that the risk of losing most of any inheritance is now greater.

    Also, any proposal to raise the proportion taken by government raises the question of the next step, because this isnt a solution to the current shortage of care.

    And while this has become the focus of attention, it is consistent with the tone of other measures in the manifesto.

  35. @ JONESINBANGOR 10.04pm
    It is very unlikely that the Tories will lose seats in the South. Here they won most of the seats in 2015 and are now polling over 50%. It is in the North West, Yorks & Humber, West Midlands where most of the Conservative marginals are. Here they have large gains – over 10% on 2015 in these Labour Heartlands. But the Labour seats tend to have large majorities, so a smaller increase in the Labour vote keeps the seats Labour. It is here where the local issues/campaigns cause the seats to change Party, rather than the National or Regional swing.
    Scotland is different. In 2015 the SNP nearly cleaned up with 50% of the popular vote. In the local elections on May 3rd there was an anti-SNP, anti-INDYREF2 vote, benefiting the Conservatives, and giving Labour a larger share of the vote than the Westminster Opinion polls (despite it being normal to be below, due to Independents at the local level). In forecasting the council seats, I was fairly close on the Con seat gain (+150 v +164 actual), but did not forecast the hardening of the Labour vote against the SNP.

  36. A friend just posted this comment on Facebook…

    “I placed a £40 bet on Labour winning most seats yesterday on betfair, an hour ago they offered me £55 to withdraw it, now they are offering £60.”

    Don’t know anything about Betfair or any other bookies. Is this common practice? Anything to be read into it?

    For what it’s worth, I think the tories will win, but not by anything like they were expecting, maybe a 20-30 seat majority. The tories campaign tactic of keeping May out of range of scrutiny, and letting Labour lose the election, rather than going out to win it, is proving to be very short sighted. “No plan survives first contact” etc.

    Of course if the tories win a majority equal or close to the number of seats they win in Scotland, Kezia Dugdale may have some answering to do considering her recent comments on voting tory to keep SNP out.

    As it is, I think the tories will be lucky to get more than 3 in Scotland. The English media has, imo, massively overplayed the “tory surge” in Scotland, and outside traditionally tory rural areas, it is non-existent.

  37. AW

    Do you have a handy summary of the different polling companies, if they use phone or online, and what methods they use to model turnout? Would be rather useful to see how these factors come into play when comparing the polling vs results, and seeing which changes most effectively overcome the past overestimation of the LAB vote.

1 3 4 5