Opinium have put out their final poll of the election campaign. Their final figures are CON 43%(nc), LAB 36%(-1), LDEM 8%(+2), UKIP 5%(nc). Fieldwork was between Sunday and today, so it is also the first poll we’ve seen with fieldwork conducted after the terrorist attack in London Bridge, though there’s no significant change from Opinium’s previous poll at the weekend. Full details are here.

This is the first poll to be badged as a final poll. Survation put out their final poll for Good Morning Britain this morning (showing topline figures of CON 42%, LAB 40%, LDEM 6%, UKIP 3% – tabs here) but final poll for GMB is not necessarily the same as final poll. We shall, no doubt, have a flurry of final polls from ICM, ComRes, YouGov, Panelbase, Kantar and others tomorrow, MORI on Thursday.

I mentioned yesterday that Ipsos MORI normally do their final poll on polling day itself, and got a few comments about whether this is allowed. Yes – it is. The rules on election day is that you cannot publish an exit poll (or any poll that’s based on the opinions of people who have already voted) while polls are still open. It’s perfectly fine to publish polls conducted before polls opened. MORI do their fieldwork on Monday, Tuesday or Wednesday like everyone else’s final poll… they publish on Thursday because they partner with the Evening Standard who publish around midday, rather than overnight.

1,166 Responses to “Opinium’s final poll: CON 43%, LAB 36%, LD 8%, UKIP 5%”

1 17 18 19 20 21 24
  1. Social media is as big an echo chamber as you make it. It can be any kind of political affiliation, any kind of quackery, any kind of scientific, etc, etc. You can make it such that the only non-friend stuff is about mathematics.

    So no, social media is not left leaning – your acquaintances are, and the left leaning ones are more active, and in general their posts (not necessarily political) are the ones that you click on or like. That’s it. You made it.

  2. @Ken

    Is that the London evening standard? First time in London for a while, and I noticed that headline splashed on my free paper.

    The way it was worded, do you think it will ring home with the average London commuter? On the face of it, I think it might…depends how much wait that paper carries

  3. Does anyone know if TOH has made a prediction yet? Not seen him for a while but then I don’t read every page as too many to go through.

    I’m interested as he was spot on last time and no one else was near.

  4. @AlbertTatlock

    That’s a silly response.

    End of reply.

  5. Laszlo

    Thanks for that.

    Depending how many people were used as data to fit the model (this is different to the 50,000 used to “score” it for each strata) the number of variables which “should” be used to fit the model with vary.

    For a start we’d expect to see 650 indicator variables for constituency. We have no idea if some of these are “pooled” into similarly behaving constituencies once accounting for the other variables.

  6. “So, I actually think that the echo chamber of Facebook and Twitter is actually making the Shy Tory thing even more of a thing this time around”

    But wasn’t the conclusion last time round that the Shy Tory thing wasn’t a (significant) factor?

  7. @ Aaron

    Other landslide years are available – are you sure you didn’t mean 1997?


  8. I think Caroline Lucas is likely to be the recipient of quite a few Labour voters, let alone former Lib Dems (who have no candidate). I can’t see it going anywhere else.

  9. Edge cases such as Brighton Pavilion are extreme instances of the (now oft-repeated) fact that this modelling approach *doesn’t* predict each seat in isolation. However, getting the seat (what seems to be) this badly wrong does rather call into question Ashcroft’s overall predictions

  10. EXILEINYORKS – Some people take the stance of “I’m not telling any-one, if I tell no-one, no-one can find out, judge or attack”. They may even say any old party to just answer the question without seeming rude by saying no..

    It’s almost like they put a political iron wall up…

  11. @ChrisLane

    Ted Dexter was also a very talented golfer, and a legendary big hitter.

    Not so great as a cricket commentator, though.

  12. Wish my boyfriend was a shy Tory! Keeps telling me about the evils of the 70s!

  13. Re the left-wing, metropolitan-elite, echo chamber that I keep reading about, are some people forced onto it against their wishes?

    I’m mildly curious as it obviously troubles a lot of people, yet at the same time they seem far less concerned about the influence of one individual – Murdoch or Dacre for example.

    I ask because, whilst I would regard myself as LOC, my own facebook account is limited to friends and family only and I don’t visit anything else online** apart from a guitar forum and this one.

    I feel as though I’m missing something.

    ** Ooooh………. apart from the Russian Ladies that Crossbat11 recommended of course. And, naturally, I gave my apologies and let myself out.

  14. @Robert Newark

    I believe TOH did, but not sure what. However, he was probably only spot-on because of personal bias – if the polls have corrected for it, he should be about 7pts above the actual gap!

    (Of course, he could also have a very sophisticated model that is completely objective, in which case I bow to his superior analysis)

  15. ‘Osborne endorses party he was always going to endorse’ shocker

    Quite how we’ve avoided having UN inspectors monitoring this election is a mystery to me

  16. Caroline is 1/10 with William Hill, I think the constituency should be renamed Brighton Green Pavilion.

  17. RichardW

    I had no idea myself, but had to look it up!

  18. LASZLO – Interesting point.. I read an article that investigated how “likes” feed the coding to tailor you news feed and slowly whittle down the friends that appear on your feed…

    So… I stopped liking things. And have done for about 2 1/2 years now. So, the echo chamber I have witnessed would not be caused by my liking habbits. Maybe commenting, but certainly not liking..

  19. Robin

    Fair dues to Ashcroft at least he given it to Labour and not the Tories .. cant say hes being biased anyway even if hes got it wrong with the Greens

  20. @ ALBERTTATLOCK – Ynys Mon 7-1 for CON so thanks I have placed a bit on that. The PC contented seats get a bit messed as they have quite a niche market and any Welsh regional swing will hence miss seat by seat specifics (same applies in LDEM seats across whole UK)
    Without sounding UKIP obsessed (I’m not – my model is)!!
    UKIP are standing but they also did get 14.7% in 2015 so with heavy tactical voting on their part it’s certainly possible and at 7-1 with some local knowlegde why not – good to spread the bets around!


  21. @ Aarron

    “EXILEINYORKS – Some people take the stance of “I’m not telling any-one, if I tell no-one, no-one can find out, judge or attack”. They may even say any old party to just answer the question without seeming rude by saying no..

    It’s almost like they put a political iron wall up…”

    That post is an evidence free zone…

  22. @Albert

    Indeed, congratulations to Ashcroft for only getting it appallingly wrong, rather than catastrophically wrong. :)

  23. NEIL…..Yes, The London Evening Standard, George Osborne recently took over as editor and has had a few snipes at his old paymasters in Westminster.

  24. Unfair comment on TOH who explained the way he read the polling evidence at the time very clearly and it made sense.

    [Well, for me anyway, after the event…………………]

    I’ll be interested in his view of this one as he is anecdote free. No nuts either.

  25. AAron

    You could be right. I haven’t done an analysis, but one of my aliases started to drop people to the very bottom of the scroll if I hadn’t done anything about their posts.

    there is another interesting thing in FB’s algorithm. It actually registers how many people saw your posts. There use to be an app that could bring it up, but it ceased to work about 18 months ago. Officially FB doesn’t tell you this so you don’t know that all your friends saw your post and none of them liked it.

  26. Markw
    Even with 5 momentums in each car that doesn’t add up to 1000. Whatever a momentum is.

  27. “To the recreational mathematicians out there, the Welsh seat projection is rather neat. All the seat totals are powers of three.”

    Nice. If there are any recreational mathematicians out there, and you’re fed up with polls and politics, I can heartily recommend “Project Euler” – just let google find it for you. No adverts, no trolling, just pure maths and computing. Something to do between the exit poll and the first results coming in.

  28. NEIL

    In Dr Mibbles’ defence…

    S/he is right. The only significant differentiator I can see is in youth turnout.

    Same as 2015 (43%): Con majority.
    Higher than 2010 (44%): Hung parliament, Cons largest party.
    Anywhere near 1992 (75.4%!): Anyone’s game. Labour minority govt. or Lab/SNP coalition most likely.

    Please could someone correct me if I’m missing something? (Assuming pollsters’ sampling is correct; no point debating that as we can’t know until it’s too late.)

    The last two elections youth turnout has been low. There’s a paper (posted earlier: https://www.opendemocracy.net/ourkingdom/sarah-willis/how-ukip-might-save-generation-from-apathy) that suggests this is because of a lack of distinct options rather than distraction or laziness. I’ve no idea how true this is so won’t make a prediction on that basis.

    However, I do think it unfair to damn Dr Mibbles for harping on about it – unless our underlying data is wrong, it’s the one factor that matters this election: will the young vote? Will they, won’t they? It’s like an episode of Friends over here.

    Certainly more relevant than the constant, unverified, unquantified whispers and hearsays people are otherwise posting.

    Kester (aka Red Kestrel)

  29. @ STADIUS – some people have speculated that the whole 2015 polling “error” was mostly due to “Shy Tory”. Long detailed write-up plus look at previous years here:

    NB The leadership gap and other soft issues (if we ignore the recent local elections) suggest around a 6-7% CON lead. Many speculate the companies with a 6-7% CON lead have fiddled the numbers to happen to arrive at that level.
    NB2 Recent local elections were actually a great example of Shy Tory factor as well but that was ages ago when May was way ahead in leadership opinion, etc. People thought small CON gains and it ended up being 500 or something – who cares about Local Elections!
    NB3 Are former “Loud Liberals” now “Shouty Socialists”!?!? Bagsy picking that name before Friday!
    NB4 YouGov should certainly have achieved scaring the Times reader “complacent CON”. CON seats at 302, Corbyn as PM with strings pulled by Sturgeon – you really could not make that up… or could you :)

  30. Kelal,

    probably need an enquiry into labours party expenses after this with all these events .. must be costing a fortune and are the candidates declaring it on their personal expenses if they get up on stage with Corbyn I wonder ? The electoral commission need to look into it. I assume hes using double decker buses or something to ferry these cheerleaders around the place

  31. MARTIN L

    Why would young people lie in a private online poll about their voting intention?

    there was no evidence of a shy tory phenomenon in 2015, it was a sampling error – why would there suddenly be one now?

  32. The YouGov model is not ‘ignoring’ UKIP in constituencies where they are not standing because, according to their Q&A:


    they use:

    […] interviews with registered voters from our panel, who are shown both the parties and candidates running in their particular seat.

    so those who voted UKIP last time will have to make the choice of what to do in the absence of a candidate anyway and there is no need to make any correction. Given they have the technology I would assume they use a similar model for theit ‘standard’ polls as well.

  33. (My point would be more compelling giving 18-34 turnout rates, which show greater difference between 2010 and 2015 IIRC. Apologies.)

  34. The reason I bang on about youth turnout is because it is THE SINGLE REASON why ICM/ComRes/Panelbase show such large leads for the Tories.

    And it’s the single most important factor in this election in determining if TM loses her majority or not.

  35. Regarding shy Tories and the internet – I have a suspicion that a lot of “shy Tories” are probably also less likely to agree to do a poll. I know the pollsters say they’ve addressed the over supply of “politically engaged” people but we’ll only know if that’s worked tomorrow…

  36. @ Albert Tatlock

    UKIP are running 377 candidates down from 624 in 2015, so does that make them a “stalking horse” for the Conservatives, or is it really that in a snap election the micro parties really have to scramble to find suitable candidates.

    Green 3 candidates Scotland, Wales 10 down from 40 in 2015, and Green 7 in Northern Ireland and 448 out of 532 in England, down from 573 overall in the UK in 2015.

    Of high concern to you should be the almost complete absence of all the micro-socialist, communist and other fringe left parties, at least I have found no mention of them.


    Earlier in this thread you said there were a combination of reasons why and admired that youth turnout alone wasn’t enough to explain the difference.

  38. *admitted

  39. Canada

    All those “left” fringe parties support Corbyn, so they don’t have candidates (they officially announced it).

  40. lots of talk about the about the turn out of the younger voters and what impact they will have based on the Yougov model. It seems like Yougov have large numbers of them votes going into the Labour camp. Thats to be expected as Corbyn seems to have build a peoples champion image with that age group. Lots of people saying a high youth turnout will give Corbyn victory.

    However its worth mentioning that in the 2010 election votes in the 18-24 demographic broke down 29.5%(con)//31.3% (Lab)//29.8(LD)

    in 2015 according to Yougov the result in the 18-29 demographic went 32%(Con)//36%(Lab) 9% (LD/UKIP). Both times over the last elections the conservative have been only just behind in this age group.

    UKIP went from not being on the map to picking up 9% of that vote. Its believed that alot of UKIP voters will go Conservative this time. So will Corbyn have as bigger lead as people expect in this age brackett? Think someone posted a while back that in terms of numbers of registered votes they are only around 500,000 extra on the list.

  41. My bunions are aching.
    Is this significant?

  42. @ KESTER LEEK – 2015 UKIP almost 4mm voters and almost 13% of the vote (all for 1 seat, which they then handed back!). No disrespect to younger voters but I’d put money on 4mm ex-UKIP above a higher turnout from Corbynistas!

    Being a snap election we maybe haven’t had time to see the Clegg mania repeat but in 2010 exactly the same thing – young voters loved Clegg but LDEM under performed in the polls right at the end and on election day and then drastically were killed in coalition. Corbyn at least will probably be spared the need to be junior partner/gimp in a coalition – sorry Nicola!

  43. @ Albert Tatlock

    I lied the Workers Revolutionary Party are running 5 and the Social Democrats 6, but on the other hand I have seen no National Front, only 10 British National Party, 7 English Democrats and 31 Christian People Alliance.

    Maybe you could lend some of them the deposit next time?

  44. The obvious comment to Canada is that the commies don’t need to with Corbynetc running!

  45. @ Canada

    I believe TUSC decided not to stand against labour. I haven’t checked all constituencies.

  46. Kelal, I was mocking earlier suggestions of bussed in crowds. I had to pay first bus £2.40 last time when I went to see JC speak.

    Shocking, no free busses for JC rallies. Something has to be done.

  47. I see YouGov are now getting close to my prediction – the Conservatives at 302 seats rather than my guess of about 297.

    I continue to think it will be a hung parliament, with Conservatives the largest party but Labour forming the (minority) government. It’s going to be a tight fit over on the other benches if that is the case.

    Polls have been somewhat static this week, but static at a place where Labour should do well. Of course, they may be wrong, but if so the turnout predictions will probably be the key, but it could be the much imagined ‘shy’ voter. (If there are ‘shy’ voters I would guess it would be that they don’t want to take part in opinion polls at all.)

  48. PNG

    They are not commies, just as Corbyn isn’t (in my view, he is not even a socialist). They are various factions of the radicalised philistines.

    Anyhow, they officially announced that they won’t have candidates as they support Corbyn (and you were right on this).

  49. When are we expecting the first poll?

1 17 18 19 20 21 24