Election night

I will be working on the BBC’s election coverage tonight (behind the scenes, so no need to look out for me), so won’t be updating the site. I’ll be closing the comments when I head up to the studio in a couple of hours, discussion will be over on the backup site at http://ukpollingreport.wordpress.com. I’ll update when the TV coverage finally draws to an end about 5pm tomorrow. I’ll probably have a sleep first :)

Best of luck to all readers who are standing or working on campaigns. Have fun watching the results those who aren’t.

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.


Ipsos MORI’s final call poll is now out, and has topline figures of CON 36%(nc), LAB 29%(-1), LDEM 27%(+4). Very much in line with the other companies, all but two of whom have the parties within 1 point of CON 36%, LAB 28%, LDEM 27%.

Last night MORI also rang back 116 people interviewed on Tuesday to check for any last minute swing – and found no significant difference. That’s it until the exit poll!

Final predictions

Ipsos-MORI’s final poll will be in the Evening Standard tomorrow, but apart from that all the other final polls are out.

Populus 37 28 27 +9
ComRes 37 28 28 +9
Opinium 35 27 26 +8
ICM 36 28 26 +8
YouGov 35 28 28 +7
Angus Reid 36 24 29 +7
Harris 35 29 27 +6
TNS BMRB 33 27 29 +4

There you go. Yesterday I said I expected the final figures would be somewhere around where ICM, Populus and YouGov were (at the time CON 34-36, LAB 27-29, LDEM 27-29) and not much has changed that, though the average level of Conservative support seemed to have ticked up a bit, with most polls now showing them at 35-37. All the pollsters are now very close indeed, the majority of the polls are within 1 point of CON 36%, LAB 28%, LDEM 27% (the exceptions are Angus Reid and TNS BMRB).

Looking at the details of the polls, Populus had a sample of 2,500, conducted between yesterday afternoon and today. An interesting part there was that of the 13% of people who said they were voting tactically, half said they were voting Conservative, suggesting the Conservatives being the beneficaries of some tactical voting for once. 59% of people told Populus they were certain to vote, compared to 57% in their poll in 2005.

The Harris poll was conducted between the 29th April and 4th May, so for a final eve-of-election poll there was some very old fieldwork. No it wasn’t! Fieldwork was actually 4-5th May like everyone else.

Angus Reid provide some analysis of marginal seats in their poll. In 150 Labour held Conservative targets Angus Reid found the Conservatives up 5 points on 2005, Labour down 19. This equates to a somewhat implausible 12 point swing from Labour to the Conservatives in marginal seats, enough for a crushing landslide. In the 63 Liberal Democrat held seats they found support at CON 33%(+4), LAB 12%(-7), LDEM 48%(+2) – a slight swing from the Lib Dems to the Conservatives.

YouGov interviewed 6483, so a big old sample. They also carried out a second poll of Labour held Conservative target seats with a majority between 6% and 14% (the same seats YouGov polled a few months back for Channel 4). These showed a 7% swing from Labour to the Conservatives, evidence that the Conservatives continue to do better in marginal seats and echoing the sort of swing the MORI and Crosby/Textor marginal polls have shown in the last week.

ComRes’s final polls is already looking ahead to the aftermath of the election. They found 51% of people agreed with the statement that the party with the most seats should provide the PM, but 81% agreed that the party that received the most votes should provide the PM.

Anyway, enough of the polling – how will it translate into seats? On a uniform swing a split of 36/28/27 would translate into a seat distribution of Conservatives 297, Labour 245 and Lib Dems 76. There are a couple of departures from UNS that we can be pretty confident of though – firstly, Scotland will not follow this pattern. The polling suggests Labour’s support in Scotland is pretty solid and there is certainly no Conservative advance (and the SNP surge we were seeing a year or two back has also gone.)

I also expect the Conservatives to outperform the national swing in Labour held marginals – the seven percent swing we’ve seen in the YouGov, MORI and Crosby/Textor polls would take the Conservatives over 300 and close to a majority. What is much more difficult to do is predict what will happen in Lib Dem marginals. It seems likely the Conservative advance will be slightly offset by losing seats to the Liberal Democrats, but there have been a couple of straws in the wind like that ICM poll of Lib Dem marginals that suggest the Lib Dems are advancing more in Labour held marginals than Conservative ones. We don’t have the polling evidence to judge that – but whether the Lib Dems advance evenly, or do better against Labour, will be the difference between the Conservatives getting about 300 seats and getting up to 310 or more.

Anyway, I can’t delay it any longer: my guess is we are going to see the Conservatives between 300-310, Labour between 220-230, the Liberal Democrats between 80-90 (though I warn you, I may be a pollster, but my personal powers of election prediction are notoriously poor!)

For other pollster predictions, TNS have made a seat prediction of CON 292, LAB 204, LDEM 114; Peter Kellner’s personal prediction is CON 300-310, LAB 230-240, LDEM 75-85; Angus Reid have a prediction of CON 320-340, LAB 165-185, LDEM 105-120.

NB – for comments, please keep on using the backup site at http://ukpollingreport.wordpress.com – to my surprise the traffic still hasn’t died down even at 11 o’clock!

Final Polls

First out of the stalls is Opinium for the Express – CON 35%(+2), LAB 27%(-1), LDEM 26%(-1). Others are presumably around 12. I will update as the rest of the final polls come in tonight.

UPDATE: Secondly we have TNS BMRB – CON 33%(-1), LAB 27%(nc), LDEM 29%(-1)

UPDATE2: Populus’s final call is CON 37%(+1), LAB 28%(+1), LDEM 27%(-1)

UPDATE3: Plenty of polls coming in now, Angus Reid for Political Betting have figures of CON 36%(+1), LAB 24%(+1), LDEM 29%(nc). Labour remain in a poor third there, and unsurprisingly it looks as though Angus Reid will be showing the lowest level of support for them.

YouGov’s final poll for the Sun has topline figures of CON 35%(nc), LAB 28%(-2), LDEM 28%(+4). So a reverse of that drop in Lib Dem support yesterday, which in hindsight looks like sample error. These figures take YouGov back to what they were showing on Monday (and within a point or so, for the best part of a week). YouGov also did a final poll of Lab-Con marginal seats, showing a 7 point swing to the Tories. More on that later once we get all the final polls and I can crunch the figures properly.

UPDATE4: Harris in the Daily Mail seem to have figures of CON 35%(-1), LAB 29%(+3), LDEM 27%(-1) (I’ve also seen 35/29/29 quoted, but 35/29/27 is what the BBC are now saying). They really have come into line with the established pollsters. ICM and ComRes still to come, when Martin Boon stops teasing people on twitter ;)

UPDATE5: ICM’s poll has finally turned up (treat them mean and keep them keen!). Topline figures there are CON 36%(+3), LAB 28%(nc), LDEM 26%(-2). Conservatives back up three points after a drop in ICM’s previous poll. ComRes’s poll for ITV meanwhile has topline figures of CON 37%(nc), LAB 28%(-1), LDEM 28%(+2). Apart from tomorrow’s MORI poll in the Stanard, I think that’s it, there’s your polls. Proper voting starts at 7am. I’ll have a proper look at the eve-of-election polls to dig out any interesting nuggets later, and then get myself off the fence and make a prediction (all will be posted on the mirror site at ukpollingreport.wordpress.com too – there have been 350,000 visits to the site so far today, so do try and help keep it online!).

NB – Comments are closed to stop traffic crashing the site. Comments are open over at the backup site http://ukpollingreport.wordpress.com instead.