There is just the one exit poll these days at British elections. MORI and NOP used to do seperate ones, now they carry it out jointly on behalf of BBC and ITN (and for the first time at this election, Sky), so all three channels will have the same one.

It is carried out at around 130 polling stations, and they conduct about 16,500 interviews. They try and use the same polling stations at each election (though changes in wards and polling districts sometimes make it impossible) so that direct changes from the previous election can be drawn. 107 polling stations will be the same ones as last time, with an extra 23 new ones, including some new ones in LD-v-Lab seats which were previously underrepresented. Unlike US exit polls there are no questions about why people voted, it’s just who they voted for.

Interviewers stop every nth person coming out the polling station, and give them a mock ballot paper to fill in, if someone refuses they are not replaced by another person. Every hour the papers are collected and phoned back to HQ, where they are weighted for differential response rates and crunched by people like John Curtice, Rob Ford, Clive Payne and Steve Fisher (if you were watching the BBC’s campaign show last night, Steve was the chap demolishing the myth of bad weather helping the Tories!). The first result comes out at 10pm on the dot, with a final projection at 11pm or so.

The aim of the exit poll is to predict the seat totals, not the share of the vote, and the team will try to work out if there are different shifts in support in different types of seat. The call is based on a probability of each seat going one way or the other, all summed up to make a seat total.

In terms of past accuracy, the exit poll last time got the Labour majority exactly right (though they were slightly off with Conservative and Lib Dem seats). Unless something goes terribly wrong, we should have a broad idea of the result a couple of minutes after 10 o’clock.


53 Responses to “About tonight’s exit poll”

1 2
  1. @ Anthony

    I think I’m right in saying that the early postal voters are taken into account by the pollsters in advance of election day, but will the exit pollsters have a formula whereby they take postal votes into account before they publish their predictions?

  2. Interesting that the polls have all converged around similar numbers in the last few days of the campaign after quite a bit of movement in the middle.

    Am I the only one that thinks they are going to be broadly proved correct in terms of share of the vote?

    Cons on around 36%, Labour on around 28% and Lib Dems on around 27%.

    Even there though a difference of 1% in any of the parties vote could make a drastic difference to the outcome.

    Also what the share of the vote doesn’t tell us is what the swing will be in the marginals – will it be larger than the national swing by around 2% or something less?

    Will there be a smaller regional swing in some areas to save more Labour MP’s and larger regional swings in other areas that will oust them?

    Even if the exit poll confirms the share of the vote from the opinion poll I am not sure it will give us an indication of all of these factors – especially in seats where the result may be very close.

  3. If they HAVEN’T got their heads around the postal votes this could be the least accurate poll in the entire election…

  4. I didnt find Steve Fisher’s explanation very convincing. What ultimately drives turnout is the closest of the outcome. It boils down to will my vote effect the result. Hence we see high turnouts in 1992 and 1974 and lower ones when the result is certain, all other things equal.

  5. Wild stab in the dark as to election result, for laughs:

    Con 32.5 Lab 29.5 LD 30

    Con and Lab to be close in terms of seat share, Con slightly ahead, LDs with around 100 seats, a bit lower.

    I’m rather skeptical of pollster figures and UNS this year, it’s an odd election.

  6. GRINCHY

    Yes – there was a good little summary of this on one of the news shows last night.

    Basically there is no correlation between high turnout and good weather.

    There is no correlation between high turnout and Labour doing well.

    These are just myths.

    The times when turnout has been most depressed is when the result seemed obvious beforehand – when the election result is closer turnout has been higher.

  7. I think, in terms of seats, there will be a small Conservative majority (10+) which could be boosted by Ulster Unionist seats.
    Anthony – Does this site have a holiday after to-day?

  8. I found the following link on the Ipson-MORI site which describes the methodology used for the exit poll (a Powerpoint by Steve Fisher)

    h t t p://www.ipsos-mori.com/researchspecialisms/socialresearch/specareas/politics/generalelection2010/theexitpoll.aspx

    No mention of adjustjment for postal votes that I can see.

  9. An exit poll that doesn’t ask useful questions about why people vote. Does it ask about gender/age/income etc?

    If not this seems rather a waste of human resources given the result will be known six hours later!

  10. 36-29-27 is the final poll from Evening Standard according to their website

  11. Did the pollsters have access to postal votes cast already in this election? If so, would this explain how they all seemed to come into line last night with a small increase in Tory fortunes? And, for that matter, are all hovering on hung parliament territiry of around 300 seats to the Tories?

  12. Thanks Mike N.

  13. EOIN

    You are going to bust …badly.

    The old Polling and election prediction Models dont work too well when there is a strong 3rd party.

    Its NOT 2005 or 1997.

    Your insistence that the old methods and historical precedence would work in 2010…IMO dont get you any points.

    Frankly… you come across as someone who once they latch onto an idea… you dont let go no matter what the actual data shows

  14. Matthew – Nick Moon of NOP’s definition of an exit poll is an extremely expensive way of finding out the election result of couple of hours early :)

    Dave K – nope, pollsters definitely don’t!

    Dan the man – in 2005 the team looked at what the opinion polls had said about people with postal votes and decided that it didn’t make any difference. I expect they will do the same this time and either ignore them, or make an adjustment to account for them, depending on whether they feel it is necessary.

  15. Steve Fisher’s demolition wasn’t actually that convincing. Maybe he dumbed it down for TV, but you’d also need to control for the average turnout at the time (so 70% turnout in 2010 can be regarded as high but would have been dreadfully low in 1950). As such you’d need to maybe take change from the last election or it as an average of the turnout in the preceding and following elections.

  16. I think that the exit poll will be an excellent way of analysing what happened with the postal votes.
    Assuming that the exit poll will be a very accurate way of predicting what happens at the polling stations, any major variation of this against the actual result will indictae the distortion of the postal votes.

    Could be a great way to identify where election fraud could be occurring…

  17. @WXDavid,

    I wish you had been with us for longer than just this last week. You would have learned a lot about the etiquette of exchange. Alas, I fear it may be another five years before you advance in that department.

    Nevermind.

  18. Mike N – actually most of the postal votes have already been opened, but haven’t been counted.

    Local authorities deal with postal votes like this. For those who haven’t got one, postal votes these days consist of a ballot paper, which you put in an envelope and then put into a second envelope along with a form that needs your DOB and signature.

    When they arrive they are opened, and the form and inner envelope taken out. They are then verified – the local authority compare the signature and DOB with the ones on the postal vote application (done by scanning them onto a computer which automatically checks them, with the returning officer adjudicating on ones the computer can’t cope with). Once they are verified, the inner envelope is opened and the ballot papers removed and the number of papers is counted and verified.

    All of this can be done before the count (and will have been happening all of this week, and at the end of last week). At the count these papers are mixed up with the votes cast on the day after they have been verified, and only then are they counted by party.

    The verification of postal votes can be attended by the agents, and while the papers are not sorted by party at the time, agents can get a fair idea of how they split up by keeping a close eye when they are being counted and keeping a tally (especially if they are on light coloured paper you can see through!). Publishing this data however is illegal – it is what Kerry McCarthy got reported to the police for during the campaign.

  19. An Mailleach – I think Steve dumbed it down a LOT for TV. He is a smart cookie.

  20. If the Tories get 290+ seats, I can see David Cameron being the next PM under a minority or majority Tory government. Anything less than 290 seats and Labour may be able to form a coalition with the Libs.

  21. Grinchy – yes and closeness of the race is another factor one would need to control for.

    One thing you can see is that high turnout helps challengers.

  22. Anthony
    It strikes me that if won’t says are shy Tories or shy labour, they are just as likely to be the same when exiting the polling station. So missing them out as you describe, repeats the same error (were it present) in the original poll.

  23. @MATT

    “If the Tories get 290+ seats, I can see David Cameron being the next PM”

    I can’t. I think he needs to get to 300. Anything under 300 and Brown could well do a deal with Clegg

    Having said that I am predicting the Tories win 310-315 seats at the moment so it probably won’t come to that

  24. I have to say the turnout seemed pretty low in my polling station this lunch time – in fact, it was empty. Not saying it won’t pick up later, but I think the turnout will probably be between 65-67%. I doubt it’ll be any higher than that.

  25. It does seem that it’s going to be:

    Con 36.5
    Lab 28

    But what will the Lib Dem vote be? If more of the undecideds vote Lib Dem than Lab or Con combined then Lab could have a terrible night, but Con might not be able to govern without Lib Dem support..

  26. “I can’t. I think he needs to get to 300. Anything under 300 and Brown could well do a deal with Clegg”

    The Tories have allegedly done a deal in NI, which will gain them about 8 seats or so. This would bring them up to 300, if not very close, if they get 290 tonight.

  27. Anthony W – thanks. I live and learn.

  28. @ Steve – “Anything under 300 and Brown could well do a deal with Clegg”

    Even if we ignore Clegg’s earlier assertions that he will support the leading party (and surely that must mean vote-share, not seats, given Clegg’s commitment to PR), the scenario you describe above will very much depend on how well or poorly Labour does. If it limps home in third place or a very poor second, I really don’t see Clegg keeping Mr. Unpopular in No. 10. If he does, today will have been the first and last time I voted Lib Dem.

  29. Howard – they have a quite counter-intuitive way of dealing with it. Before stopping someone coming out the polling station, the interviewer guesses how they will vote and notes it down. From past exit polls, they found this was a better predictor of people’s vote than age,gender, social class and so on.

    This means then can check to see if “Tory-looking”, “Labour-looking” or “Lib-Dem-looking” people are disproportionately refusing to take part. Last time they did not, and no such adjustment was necessary, but it is something that is checked.

    Sounds a bit bonkers of course, but – well, we’ll see. It’s probably a lot easier in polling stations that also have tellers from multiple parties, as you’d be able to guess based on people’s reactions to the tellers.

  30. @ Anthony Wells –

    thanks for that explanation of the rigmarole involving postal votes. From that, wont the cumulative knowledge of agent feedback up and down the country must be informing the parties and pollsters what’s roughly going on nationwide?

  31. Penddu makes an excellent point re identifying abnormal party bias (fraud) in postal voting by comparing exit polls with declared result. Will the exit poll tallies for individual constituencies be available on the night or available to anyone who wants to challenge the result in court I wonder?

  32. James Ludlow – if the offer is a full PR referendum within a year and a GE to follow within 6 months of that (if PR prevails) I can just about see Clegg supporting Brown.

    Being committed to PR isn’t going to make him support a PM who would never countnance PR in any form

    Bi IFs going on there though I know!

  33. John TT – if Labour puts that deal on the table, it’s serious desperado stuff. They’ve had 13 years to reform the voting system. I’ve heard of conversions on the Road to Defeat but that would really take the biscuit for unadulterated cynicism.

  34. @Steve

    If Clegg supports Brown with Cameron on 290 seats it’s safe to say the the Liberals are nothing more than the “Labour B Team” and will lose and credibility as an independent party.

    Clegg wants to REPLACE labour, not be seen to prop them up.

    If Lab and Con seats are equal he has a tough choice.

  35. Thanks Anthony. Time for a Howard anecdote. I told from 0700 until 0900 one year in a rural polling station. My Tory colleague, a retired headmaster, chuckled as a rather large young woman drew near. He warned me that we about to get a speech about privacy which we got. ‘She was SDP, now one of yours I assume’ he said. ‘She is of course a school teacher which presumably you guessed’, he chuckled. ‘I know my own kind’ he added.

    So there is somethimng in what you relate.

  36. @ Steve – in my constituency, the Conservatives are a poor third and have no hope of overturning the incumbent Labour MP. The Lib Dems are unlikely to either but there’s a chance, therefore I voted Lib Dem.

    I like both the Tories and the Lib Dems better than Labour though I can’t say I’m completely sold on any party right now.

  37. Dave K – it informs the political parties… it doesn’t inform the pollsters, since we haven’t got agents in every constituency and can’t attend counts anyway. It’s hard to get really useful stuff from them anyway, you count a couple of hundred ballot papers from tens of thousands, it’s a small sample and postal votes may not be that representative.

    MarkH – they aren’t, and I think it is one polling station per constituency, so you couldn’t really judge anyway. It could just be a different sort of swing in that polling district (more to the point, most allegations of electoral fraud in the UK seem to be based around postal votes, and by definition exit polls don’t interview postal voters).

  38. @WXDAVID

    The data from the Mori marginals says undecideds lean 15% Con, 23% Lab, 15% Lib Dem on the data table. I’ve never seen what you are quoting.

  39. Ref: Postal votes.

    It may be worth knowing that a lot of people, if my experience this morning is anything to go by, are bringing their postal votes TO the polling station in order to vote.

    There is also a HUGE turnout here so far. Of course, I don’t know if it will last but it’s huge. People are making all sorts of efforts to get here to vote including one lady who was about to give birth and several who could hardly walk.

    If nothing else, people are defiitely enthused. BTW, I was only asking for poll numbers but several people volunteered their preference. They were ALL conservative voters. That tells you nothing except that Conservative voters are extremely keen to vote and feel very strongly. The polling station was in a 3-way mixed ward.

  40. Brett – what colour rostette were/are you wearing?

  41. James Ludlow – reform of the voting system started with the Lords – and they were devilishly difficult to shift – the current offer of PR referendum for Lords reform is the bnext step. Their Lordships probably think 13 years is a very short space of time to give up their birthright and thant of their future descendants.

    It is not such a great leap to move from AV to an AV or PR option, and doesn’t mean Brown would be being cynical in any new way.

    His project has been to destroy the Tory party as represented by Cameron and his ilk. An offer as I described it presumably becomes less and less cynical the narrower the gap in seats.

    One thing is unavoidable – GB’s days as PM are numbered (if they exist at all)

  42. @ALAN
    “Clegg wants to REPLACE labour, not be seen to prop them up.”

    That is virtually impossible under the current electoral system, so if Brown offers him electoral reform then he is hardly likely to say no is he?

  43. correction:

    undecided leanings in the marginals:

    Con: 13%

    Lab: 21%

    LibDem: 15%

    http://www.ipsos-mori.com/Assets/Docs/Polls/Reuters_Wave_5_Tables.pdf

    page 8

  44. @ JohnTT – great, a PM with a “project to destroy”. Not that I’m surprised.

    Another reason it’s hard to imagine Clegg doing a deal with Brown is that Brown’s not exactly collegiate in his dealings with others. Lord knows how he’d cope with having to run things past a popular pipsqueak like Clegg. And the Mandelson/Balls conniving that would go on …

    Anyway, I think I’ll worry about that if it happens, to spare myself any more stress than absolutely necessary.

  45. “Tory looking people”
    “Labour looking people”

    In Irleand when denoting Protestant or Catholic, this is relatively easy. But pray do tell, what is the visual difference between Labour and Tories?

    :)

  46. I’m already quite excited. Might have election fever. There’s a lot of it about. http://wp.me/lZtQ

  47. @ Eoin – “But pray do tell, what is the visual difference between Labour and Tories?”

    Easy. Labour voters all have flat caps and whippets. They arrive on foot and smell of brown ale.

    Tory voters all sport monocles. They arrive in 4WDs and smell of horses.

  48. James Ludlow – i have to add that that was my choice of words – he’d probably put it completely differently

    So there’s no need to be surprised or not surprised – my own personal take. I’ve never met any of them so I’m guessing.

    Being Collegiate was the one majoir thing that Brown tried to do in his campaign , with constant references to his team, and handing the mike over to them.

    You’re right not to worry – crimes against the electorate by the elctorate are quite rare, so don’t have nightmares :)

    I don’t think it had much effect on the voters, but if he is returned as caretaker PM supported by Clegg, that’s one aspect where he would have to walk the walk without any question.

  49. @JAmes

    I see.

    So their shyness/bashfulness is in vain then.

  50. Because of the leaders debates and massive interest from first time voters, I still think with the error percentage 3 per cent, anything can happen !
    Huge veiwing figures tonight I predict for result programes across the networks.

1 2