We don’t have any more information on how the British Polling Council’s review of the election polls will progress beyond it being chaired by Pat Sturgis, but several pollsters have given some thoughts today beyond the initial “We got it wrong and we’ll look at it” statements most pollsters put out on Friday. Obviously no one comes to any conclusions yet – there’s a lot of data to go through and we need thoughtful analysis and solutions rather than jumping at the first possibility that raises its head – but they are all interesting reads:

Peter Kellner of YouGov has written an overview here (a longer version of his article in the Sunday Times at the weekend), covering some of the potential causes of error like poor sampling, late swing and “shy Tories”.

Martin Boon of ICM has written a detailed deconstruction of ICM’s final poll which would be have been an interesting piece anyway in terms of giving a great overview of how the different parts of ICM’s methodology come together to turn the raw figures into the final headline VI. Martin concludes that all of ICM’s techniques seemed to make the poll more accurate, but the sample itself seemed to be at fault (and he raises the pessimistic possibility that sampling techniques may no longer be up to delivering decent samples)

Andrew Cooper of Populus has written an article in the Guardian here – despite the headline most of the article isn’t about what Cameron should do, but about how the polls did.

Finally ComRes have an overview on their site, discussing possibilities like differential response and the need to identify likely voters more accurately.

674 Responses to “The polling post-mortem – some pollsters’ thoughts”

1 12 13 14
  1. @ RAF

    It is a good point, but you forget that all but Conservatives will also fight with each other.

    My point was purely about Labour. Can any of the candidates (potential ones too) reframe the economic narrative. I doubt it. It would be like inventing a new language that instantly accessible to the audience.

    But yes, the Conservatives could have a much harder job to sell their narrative.

  2. @Millie

    @AW answered that Q a while ago. The LDs did receive an incumbency bonus but as they tanked it only reduced the scale of the defeat in seats they lost anyway.

  3. Millie,

    I think the main thing that gave a big LD incumbency bonus credibility was Lord A’s constituency polling, and in particular the SVI/CVI differences which were probably interpreted as supporting the case for incumbency/tactical voting in favour of the LDs. Here is a quote from Lord A’s own commentary on his 1st April batch of polls.

    “As ever, the Lib Dem vote was significantly higher in all seats when respondents were asked to think about their own constituency and the candidates likely to stand there, compared to the standard voting intention question simply asking which party they would vote for in an election tomorrow. Indeed the party was well behind in all eight seats – rather than ahead in four – on the basis of the standard question. However, I have not gone so far as to name individual candidates, as the Lib Dems do in their own private polling.”

    As we now know, the actual GE results did not follow the constituency polling pattern for the LD seats, and we would have done much better by just assuming UNS.

  4. @Mr Jones

    I didn’t need to lie, ‘cos I could see what you were up to!! Honestly: bait and switch?? It’s not exactly sophisticated!!

    If there’s no objective centre, that would reinforce my point that it’s difficult to define.

    Some might have a discussion over whether there’s an objective centre or not but being as it’s your problem rather than mine, I gave an answer that worked both ways. Because in truth, either way, it’s tricky. Now, if you are done quibbling needlessly… We all have lives to lead you know!!

  5. I don’t really think we can say, this far out, what the factors will be at play in the 2020 general election.

    After all, in 2010 half the people on this board were predicting that the coalition would last more than a couple of years, and that Labour would reap a huge harvest of former LD LoC voters. Not to mention that the economy would still be in dire straits as austerity would prevent a recovery.

    The two big known unknowns are the conduct and outcome of the UK referendum, and the 2016 Holyrood election (and subsequent behaviour of the SNP). Added to that there are of course any number of unknown unknowns that mean that the election may well hinge on something we haven’t even conceived of yet.

    Let’s work out why noone except TOH and Robin Hood was able to predict the result of the last election before we start predicting the result of the next one.

  6. @Carfew / MrJones,

    It’s getting a bit dull now. Can’t you just agree that there’s an objective centre but that it’s a different objective centre for each of you?


  7. bah, earlier post should have been “coalition wouldn’t last”.

  8. @Hawthorn

    If it was not for the SNP, Labour would be on 272 seats.

    Labour would still have lost, but don’t try to blame Ed Miliband for the rise of the SNP.

    We would have been talking about 30 gains from the Tories next time which is a very different kettle of fish.’

    Absolutely right. The collapse in Scotland makes Labour’s performance worse than it actually was – in England Labour did somewhat better than in 1992. If Labour can get back to 260 in 2020 , it would be able to contemplate a minority Government with SNP/Plaid/LibDem support. I suspect that the LibDems will be in no mood to help the Tories out again and will be looking for some revenge. I suspect that we will hear much less about the largest party having the first right to form a government – indeed I don’t recall Ashdown advancing that argument prior to the 1992 election. Had Kinnock and Ashdown been able to outvote Major after that election I am pretty sure that a Lab/LibDem deal would have been reached – regardless of the Tories still having most seats.

  9. RAF

    Thanks both.

    Your replies confirmed my impression from simply looking at a large number of results. As EiY says, we all would have done much better predicting on the basis of UNS. The LD slump was remarkably uniform.

    I noticed that Lynne Featherstone, who was held up as an example of a popular incumbent ( and I think she probably was ) basically did no better than anyone else.

    I think it was Pete B, or it might have been Chris Lane, who predicted a return for the LDs to their six seat situation of 40 odd years ago. This was scorned at the time, but I remember thinking that if there was an incumbency bias, then it should have applied similarly in Jo Grimond’s day.

    People vote very much upon party lines: local factors and loyalties to individuals count for comparatively little.

    Did our pollsters just over-analyse and over-complicate things?

  10. @Graham

    You’re right re Scotland. But it means that a “best case” scenario for Labour is that it will be contesting 2020 in a situation where the Conservatives can rerun the “coalition of chaos” arguments that caused Labour so much difficulty in 2015.

    I do wonder now whether the cause of the left in England and Wales might be better served if Scotland seceded.

  11. Actually, to a degree it seems that the design for the 2015 elections was to give all seats to SNP in Scotland and take away all seats from the LibDems. As all great plans it had some shortcomings and failed.

    (Joke of course).

    It comes from Farron’s claim that he was taking soundings from his fellow MPs (seven phone calls in all) before making the bid. It might have been misreported and he is canvassing the 10,000 new recruits (I do believe that many joined). Which then is a long time.

    Labour is not doing better (library vs street vs TV studio), but at least it could be done on a two-hundred and odd petals daisie (he loves me, she loves me not – OK, I’m in. I will appeal to aspiration, and I’m inspired) as only the PLP can recommend candidates.

    Greens are much better. We achieved the same as five years ago, but we did better (?), and it’s a solid basis for the future. What about learning from the LibDem mistake?

    I won’t comment on UKIP, because I would be accused with bias (of which I have none, of course), but Father Goriot is a good start (Balzac is good for every analogies).

  12. @Neil A

    There are indeed many unknowns. It’s not even a given that there will be a 2020 general election.

    Imagine if defecting to UKIP had been a viable option for some of John Major’s “b*****ds” after 1992. It is not beyond the bounds of possibility that a significant number of Conservative MPs will yet defect to UKIP if the EU referendum is divisive enough and Cameron is judged to have betrayed the Eurosceptic right.

    That is not a prediction, just an added layer of uncertainty.

    Mind I was one of those who for the first year of coalition thought it might not go the full course, so comprehensive had been the LD change in direction from their 2010 campaign. I underestimated the supine nature of the LD grassroot diehards and the potential for speed cameras on the M11 to derail the ambitions of the LD most likely to challenge Clegg.

  13. @Phil Haines

    Doubtless the Tories will try the same scare tactic – though it does not usually prove as effective second time round.Depends also on how much chaos we face over the next few years – a rerun of what happened to Major as his majority withered away would not put the Tories in a strong position to make that case credibility wise. There is a chance we will be back to a Hung Parliament before we get to 2020. After 10 Tory years people will also be more receptive to a ‘Time for a change’ message.

  14. Phil Haines
    You didn’t think the Tories would finish in front of Labour in 2010, you didn’t think the coalition would last, you didn’t think the Tories would win, let alone gain a majority this time.
    Do you not think that political preference is seriously clouding your judgement? Certainly you are not alone, many who were ” up the pictures” in 2010, were similarly wrong, (only more so) this time.
    I can only assume that many Labour supporters on this site, live and work with other Labour supporters. Therefore they totally under estimate the Tory Party.

  15. Graham
    Not unless there is a marked change in the accomplishments of your leader.

  16. For the sake of democracy, I really do recommend more forward thinking and less belly button gazing.

  17. I wonder if the stretcher vote will return this parliament. Not as many elderly MP’s nowadays especially on the Tory benches but votes could be close enough to require it. Especially over the Human Rights Act where there is at least two Tories who will never vote for its repeal( Grieve and Ken Clarke).

  18. Dianne Abbot London Mayor? Puhlease !

    Sadiq Khan? maybe

    Journalist Christian Wolmar Also declared candidature.

    Wonder when gorgeous George will announce his intentions.

  19. @NeilA

    “Let’s work out why noone except TOH and Robin Hood was able to predict the result of the last election before we start predicting the result of the next one.”

    Agreed on the second clause (provided the “we” applies to pollsters and not individuals).

    As to the first clause, every lottery has a winner. I’d suggest that these two individuals got lucky by backing their own political preferences rather than the groupthink. I doubt there’s any scientific or statistical basis to their success beyond wishful thinking that came good.

  20. New thread

  21. NEIL A
    @Carfew / MrJones,
    It’s getting a bit dull now. Can’t you just agree that there’s an objective centre but that it’s a different objective centre for each of you?


    Lol, typical police, turn up late after the incident is already resolved!!…


    “I doubt there’s any scientific or statistical basis to their success beyond wishful thinking that came good.”

    Certainly not wishful thinking on my part.

    I would suggest however careful and generally correct analysis of what the detailed questions in the polls were telling us about how the electorate felt about the economy, leadership, NHS, immigration, unemployment etc.

  23. @Raf

    At no point did NS say she wouldn’t deal with Tories over more powers, in fact it is clear that they are dealing. NS’s statements were all in regard to putting the Tories into government. And obviously she hasn’t done that the voters of England have.

    And yes something has changed: Scotland is being ruled yet again by a Tory government it didn’t vote for. It is clear that the UK does not and probably cannot work for Scotland. If Scotland had known in September there would be a Tory government this May, I am pretty sure it would have been a narrow Yes vote.

1 12 13 14