I’ve only had a couple of hours sleep so this is a very short comment on lessons from the polls at the election. The two best performing traditional polls seem to be those from Survation and Surveymonkey. Survation had a one point Con lead in their final GB poll, Surveymonkey had a four point lead in their final UK poll. The actual lead is 2 or 3 points depending on if you look at UK or GB figures. Congratulations to both of them. While it wasn’t a traditional poll, YouGov’s MRP model also came very close – it’s final GB figures were a four point lead (and some of the individual seat estimates that looked frankly outlandish, like Canterbury leaning Labour and Kensington being a tossup, actually turned out to be correct).

Looking across the board the other companies all overstated the Tory lead to one degree or another. The actual share of the Tory vote was broadly accurate, rather it was that almost everyone understated Labour support. I have a lot of sympathy with Peter Kellner’s article in the Standard earlier – that to some degree it was a case of pollsters “trying too hard”. Companies have all been trying to correct the problems of 2015, and in many cases those changes seem to have gone too far.

A big gulf between pollsters that many commented on during the campaign was the attitude to turnout. The pollsters who were furthest out on the lead, ComRes, ICM and BMG, all used methods that pumped up the Tory lead through demographic based turnout models, rather than basing turnout on how likely respondents said they are to vote. This was in many ways a way of addressing an issue in 2015 polling samples that contained too many of the sort of young people who vote, weighting down young turnout (and turnout among working class respondents, renters, or less well educated – different pollsters used different criteria). This wasn’t necessarily the wrong solution, but it was a risky one – it depends on modelling turnout correctly. What if turnout among young people actually did rise, then pollsters who were replicating 2015 patterns of turnout might miss it. That may be what happened.

That said, one shouldn’t jump to conclusions too quickly. It may be a case of how demographic turnout models were applied (by weighting the whole sample to match 2015 recalled vote and then separately weighting different demographic groups up or down based on likelihood to vote there’s a risk of “double-counting”). Most importantly, the YouGov MRP model and the Surveymonkey survey both based their turnout models on demographics too, and they both got the election right, so clearly it’s an approach that has the potential to work if done correctly.

Personally I’m pleased the YouGov model worked, disappointed the more traditional YouGov poll had too big a lead… but that at least gives us something to learn from (and for most of the campaign the two showed a similar lead, so rolling back some decisions and learning from the model seems a good starting point).

And with that, I’m going to get some sleep.

2,448 Responses to “Post-election thoughts”

1 45 46 47 48 49
  1. “The ruthlessly electorally efficient Tory party has only been able to command a majority in Parliament for just 2 years and 1 month in the past 20 years. They might want to reflect on that”

    Or 20 years in the last 38.

    Depending which arbitrary start point you pick.

  2. @Cloudspotter

    I’m not sure that a direct comparison works though, as its likely that there are a fair number of ‘leftwing’ older people who decided to vote for the Conservatives to guarantee Brexit.

    Once the issue of Brexit is resolved then many of those may be more inclined to support Labour again and rebalance the scales slightly.

    Also, anyone who thinks that they can rely on ‘time’ alone to lop off the advantage of political rivals as voters slowly die off is probably being optimistic (and a bit distasteful).

    I believe there was a survey only a year or so ago that suggested the generation born since 2000 are the most right-wing generation ever in terms of the political opinions they express. I remember reading it …. will try and see if I can find it.

  3. @ Candy:

    Well the Cardington airship great shed that used to be the scene of fire tests was sold off, once the Building Research Establishment was privatised.

    Doubtless the facilities for tests have been minimised to save taxpayers` money.

  4. Concerning the article cited above by voice of reason.

    The last bit, after the evidence of a move to right on social policy by younger people, may be shown to have been correct.

    “The “deterioration” of the NHS and state education services could mean that “things could suddenly kick the other way”, he said.

  5. Just catching up with that tower fire in London poor old UK having a bit of a rough time of it lately .

  6. [snip…]To get back to polling: the Greens in the end voted Lab, partly I suppose because there were too few credible Lib-Dem alternative candidates. & why would they vote for H. Brexit.

  7. MARKW

    Sorry TOH ;-)

    No problem, everybody is entitled to post their views (in a non partisan way).

    My view remains we will leave probably without a deal. Painfull but worthwhile in the end.


    Back to the garden and the allotments. Investments doing brilliantly at the moment. Taking profits now as the bull run must come to an end soon.

  8. @neil wilson

    True but past “glories” don’t shape the present. Such a poor record in the past two decades is pause for reflection.

  9. Test

  10. @Donald
    How did you know your mum was a ‘borderline nazi’? Did your family go on holiday to Poland every September?

  11. @CANDY

    But taking the landslide of ’97 is has a wee bit of a distortion. Whilst not as neat as comparing 97 and 17, if you compare ’79 and 97.

    25-34 1979

    Con 43
    Lab 38
    Lib 15

    1997 45-54
    Con 31
    Lab 41
    LD 20

    In this period shift to the left as you get older.

  12. @David Welch

    The Building Research Establishment still exists, and still does tests on fire – they’ve just gone global:


    It’s completely wrong to claim that no tests are done in the UK on building materials. Not sure why people are making stuff up about this.

  13. @cloudspotter

    (Except REDRICH, good post BTW)

    Cheers – my grandfather always told me never trust the Daily Mail (because of the general strike) and never underestimate the Tories ;-)

  14. YouGov enjoying the “I told you so”:


    Yes, yes, well done. However, can MRP capture the following:
    1/ Tactical voting (significant impact in Scotland and many English seats) – possibly the seat level info will show it as a factor but copying across demographics will miss it (at best the model can flag a difference between the small sample info it finds in specific seats and the large “copy-over” approach – that would be great info!)
    2/ Seat specific issues (e.g. first incumbent (not sure any proof on that?), popular/tarnished MP, etc.)
    3/ Regional factors (the sheer number of demographic divisions means MRP will have to prioritise some factors over others but the narrower you focus the sample the higher MOE and the broader the expansion of “copy-over” demographics (e.g. treat whole nation as one and reduce the number of demographic parameters) the higher the risk of missing a regional/specific small demographic factor such as old people in S.East/London scared by dementia tax but old people in poorer postcodes less bothered, Hendon Jewish vote, etc.)
    4/ Tightening up the 95% confidence intervals. This GE was unique in many ways and hence a lot of people expected a very wide range of outcomes 300-420 CON seats in my case subjectively and incorrectly biased to the upside (notably the herd pollsters and betting markets all centred on 7%ish CON lead and CON 355ish seats). Keen to see the next update of the model

    It is certainly great to see different approaches and I enjoyed attempting to reverse engineer YouGov’s approach (and hopefully many of us shared in the Canterbury and student town vote tips!) but after 25years trading I’m going to stay wary of putting all my eggs in one basket!

    P.S. Any chance of seeing the final prediction? Last update is still showing 7June although maybe its just the date update that didn’t get refreshed? Even better a post election update given the potential for another GE sooner rather than fixed term.

  15. @Redrich

    That’s assuming that you believe that voting for Blair was switching to the left or whether the fact that the centre of gravity had shifted in British politics and Labour had shifted to the right meant that it picked up naturally ‘right-wing’ voters.

  16. @Redrich

    25-34 1979

    Con 43
    Lab 38
    Lib 15

    1997 45-54
    Con 31
    Lab 41
    LD 20

    Those people are now 65-74, and according to Ashcroft the 65+ group broke as follows:

    Con 59%
    Lab 23%
    LD 10%

    So the cohort who were 25-35 in 1979 made a tiny shift left in 1997 (3 percentage points) possibly because they didn’t see Blair as that left, and then went very right wing in 2017.

    The cohort that was 25-35 in 1997 has shifted from 28% Con in 1997 to 40% Con in 2017, a much sharper move right.

  17. @ CANDY/REDRICH – very useful to look at the past, history doesn’t always repeat but it does echo!
    Could you repost with comparison to the overall vote % in each election (+/- %s to see how different each demographic was to the overall vote) and/or provide URL for your source info.

  18. Should have added, that the 1979 25-34s really shifted to the LibDems in 1997, not Labour which went up only 3% in 1997 for that group.

  19. 97 election data is also heavily skewed back tactical voting to make sure the Conservative Party didn’t get anywhere close to getting back into power. Tony Blair ended up with a far bigger mandate in parliament than his voter share actually suggests….

    you can also argue that by seducing voters in Scotland and Wales with devolved power it further biased the data in those markets ….

    even if those steps led to an overall reduction in the Labour vote over time and made it much harder for Labour to form a government …

    which is still a really fascinating question based on the polling. Can Labour actually win a majority if ~50% of Scotland seats go to the SNP for the foreseeable future?

  20. “Can Labour actually win a majority if ~50% of Scotland seats go to the SNP for the foreseeable future?”

    Yeh even with 50 Scottish seats they would have been short of majority in 2017. Did the tactical voting work in reverse in 2005?

  21. Candy the point I was making is that the nature of the 97 election distorts the data comparison – as it was a landslide and Tory vote was relatively low. Comparison between ’83 and ’01 which in essence were both landslides also points in different direction to your assertion.

    As others have noted there is evidence to suggest at all times in their lives boomers have shown an increased likelihood to vote Tory over Xers and millenials. There are other factors at play – appeal of party policies, shift in loyalties amongst WC and MC voters, where the centre of politics is etc.

    The key point in this debate is the relative extent to which some people in the cohort move to the right as they get older and the extent to which that offsets the current generation of older voters. The traditional explanation for the over 55’s being predominantly Tory was due to social / economic factors. Currently, I think its due in part to the pro-pensioner policies that have been in place since 2010.

    Personally I think the younger generations (under 45) in general have a different set of experiences and social views that give a tendency to favour Labour. Remember more of the over ’45 were able to get on the property ladder. This does not been Labour victory’s are predetermined in the future – its more that it will be the Tories that need to adjust there offering to win over these voters.

  22. @ Spaceman

    It’s pretty obvious if a party leader comes up with a left field proposal like free tuition fees in a manifesto it can completely skew an election, alternately the same can happen if a party leader messes up by not clarifying a social care policy properly. This could easily happen in the next election and we’ll all be left scratching out heads as all our lovely theories go straight out of the window.

  23. sorry ‘mean’ not ‘been’

  24. @Alec

    The phrase “You can’t do safety on the cheap” springs to mind.

  25. @alec

    Until we know what’s happened, you shouldn’t be making party political points about this awful fire. Just my view. Way, way too early.

  26. @BANTAMS

    I’m not sure free university tuition is that left-field. After all, no-one argues with free education for 4-18 year olds.

    With 50%+ of young people going into higher education, and that figure rising, and our economy needing more qualified people (especially if we can’t import as much talent because of Brexit), then free tuition not only seems good for those that get it, but also for wider economic growth.

    The government gets increased tax revenue from higher earners and the businesses that benefit from those skilled employees helping generate more revenue, and our public services benefit as well (more teachers / doctors / nurses etc).

  27. About the shift right as you age: Historically, the working class was left wing and had shorter life expectancy. Possibly now that the former is less true (as per YouGov figures showing that classes all vote more or less the same) maybe in time we’ll see this effect (aging veering right) diminishing.

  28. On the point of education, there is definitely an argument to say that it is a public good, and spending public money on tuition fees is money well spent if it improves access to education.

    Also, makes party political sense for Labour when you look at the polling data.

    The other striking differentiation in voting was that the more qualifications people have the more likely they are to vote Labour or Lib Dem. (and the more likely they are to care about tuition fees and education policy in general)

  29. COLIN

    Thanks for the BPC link. No surprise that they don’t plan another enquiry but perhaps a little surprising that the downrating of younger voters turnout didn’t get a mention.

  30. “For the sake of factual accuracy, there were quite a lot of “green” policies introduced by Chris Huhne and Ed Davey from 2010-2015, arguably more than in the preceding 13 years of Labour government, but most have been undone 2015-2017 by the Tories….”


    Yes,Mathis is summat oft ignored. It’s not just about what policies are implemented by a party but how easy/likely it is they’ll be undone. Or even co-opted to not necessarily positive effect.

  31. @ Voice of Reason

    I wasn’t passing personal judgment on the tuition fee policy, I was making the point that the policy came unexpectedly from left field and skewed the debate as did the policy faux pas from the Tories.

    Anyone here who thinks their party is a shoe-in next GE could be in for a rude awakening.

  32. Alex

    Personally I think Labour trying desperately to make political capital out of this tragic event without any idea how or why this event happened is deplorable but so pridictable.
    No doubt momentum will have the internet alive with tweets about this. Of course if this had been a Labour ward the silence would have been deafening.

  33. Farron’s quit…

  34. Breaking news that Farron has stood down as Lib Dem leader…

  35. Hopefully he has an allotment…

  36. @Colin

    From da Beeb…

    “Former chancellor George Osborne has urged the government not to change its economic strategy after being left without a Commons majority.

    Mr Osborne’s newspaper editorial said a so-called “end to austerity” would lead to a “loss of economic credibility”.”

  37. Farron resigns.

  38. Farron gone. Time for Jo Swinson to step up I suspect.

  39. Tim Farron’s resignation statement:

    “The consequences of the focus on my faith is that I have found myself torn between living as a faithful Christian and serving as a political leader. A better, wiser person than me may have been able to deal with this more successfully, to have remained faithful to Christ while leading a political party in the current environment. To be a political leader – especially of a progressive, liberal party in 2017 – and to live as a committed Christian, to hold faithfully to the Bible’s teaching, has felt impossible for me.”

  40. with = within months, 4th paragr @ 6.36 pm

    “While the final polls were not ideal, the BPC does not feel there is a need for another formal inquiry;”
    Final polls’ average gave the lead at 8% when the actuality was 2%, with half the polls putting the lead in double figures, 10-13%.

    “the detailed findings of the Sturgis review are available here [which obviously I havn’t read] and provide a reference point for understanding the issues and challenges.”

    This reminds me of the sort of thing I used to read at work when the high-ups had made a screw up. The key phrase here is “issues and challenges.” which is habitually trotted out on these occasions, which at work always translated as “the severe mistakes and problems in our analysis/policy.”

  42. “To be a political leader – especially of a progressive, liberal party in 2017 – and to live as a committed Christian, to hold faithfully to the Bible’s teaching, has felt impossible for me.”

    I take this statement by Farron to mean that he felt his political support for LGBT rights was hypocritical as a fundamentalist Christian?

  43. I always thought Farron was ok. He is good for a one liner of which he had a few good ones in the election. Libs were squeezed and they still gained, if very slim.

  44. “The verb ‘SHOO’ is used for driving animals such as livestock,*** and there’s a clearer link to the earlier meaning of the noun as a fixed race: a chosen horse wins easily as they only had to shoo it in.”

    So, will posters please not write “SHOE IN anymore, ta v much.

    *** Rosie and Daisie do not require “shooing” as they are both perfectly trained.

  45. Also, while I’m at it, Sir Henry Paragraph invented the paragraph for the express purpose of making very, very, very, very long streams of words easier to digest. [Or “read”, if you prefer.]

    I tend to belt past all long posts that eschew their usage.

    So there.

  46. NickP – “Straight out of Victorian slum landlord history”

    Spending £8.6 million on refurbishments last year = “victorian slum landlord”?


    The ironic thing about this whole episode is that if they were penny-pinching and had not refurbished people would have lived, because the old cladding was fine.

    There is a lesson there – if it ain’t broke, don’t fix it.

  47. Rich

    I agree more to do with Corbyn losing the GE than any sense of real concern.

1 45 46 47 48 49