some random exit polls/election day polls for other ballots

[nodealbrexiteer]

I have also collected information over the years for these.

Couple of earlier ones to start with-ITN/ORC EEC referendum 1975 68.3% stay in(actual 67.2%).

BBC 1979 European election-GB forecast con 60, Lab 18, Oth 0(actual Con 60, Lab 17, SNP 1. the 3 Northern Ireland seats were done under the STV system and weren't forecast)-source of information for forecast was turnout figures made available before the count

[nodealbrexiteer]

ITN/ORC 1979 Scottish devolution referendum Yes 57%, No 43%(actual Yes 51.6%, No 48.4%)(not a Yes to the assembly as the Yes had to represent 40% of the entire electorate and the figures was 32.9%)

[nodealbrexiteer]

BBC/NOP 1989 European election exit poll: GB shares Lab 44%, Con 32% ,Green 14%,SLD 6%. GB seats(actual result in brackets) Lab 51(45), Con 26(32), SNP 1(1).

Still trying to dig out what I have on the 1984 european elections for BBC and ITN, ITN's 1979 Welsh devolution referendum and any others

[nodealbrexiteer]

What I do know is in the 1984 European Elections ITN used a Harris exit poll and BBC a Marplan last minute poll on election eve and day similar to the Gallup poll in the 1983 general election. THe BBC/NOP 1989 Euros exit poll was part of their exercise to get exit polling right in time for the 1992 general election following the change of methodology from the 1987 poll fail

[nodealbrexiteer]

(I think from memory ITN/Harris 1984 Euros forecast for seats(still trying to find shares) was Con 44(actual 45),LAb 33(32),SNP 1(1))

[nodealbrexiteer]

Feel free to chip in any other forecasts you have-I collected less information from these other elections that were held after 1990

[nodealbrexiteer]

FROM AN EMAIL IN 2008 FROM PROFESSOR MICHAEL THRASHER ON SKY NEWS' ELECTION NIGHT FORECASTING TECHNIQUE:



you're quite right: sky's election night forecast method has not been published - unlike the BBC and ITN - but these forecasts are based on exit poll data whereas Sky does not have an exit poll to work with. Therefore, we wait for the first election results to come in and from then on we adjust the votes in each constituency to take account of the real votes. of course, this is not done as a simple national figure, i.e. if Labour is down five points in the first constituency we don't simply take five points off Labour everywhere. Instead we adjust the seats either on a regional basis and/or on some definition of whether or not the seat is significant in some way (e.g. a by-election gain in the previous parliament, an Independent, the capmaign has been notable for some reason etc. This makes the forecast an art as well as a science - but this is also true for the way in which exit poll data are finessed into an election forecast. This is quite a different exercise to that done by both BBC and ITN initially but at some point during the evening those organisations too will begin to project from the actual results rather than their exit poll data. in 1992 they both waited too long before dumping the exit poll data in favour of the actual results and Sky called the election first. By contrast, in 2005 their exit poll data was good and their interpretation/extrapolation from it excellent witht eh consequence that they called a 66 seat majority some time before we got there at Sky. you win some, you lose some. hope this helps.

[Harry Hayfield]

Jul 23, 2019 11:34:07 GMT nodealbrexiteer said:

I have also collected information over the years for these.

Couple of earlier ones to start with-ITN/ORC EEC referendum 1975 68.3% stay in(actual 67.2%).

BBC 1979 European election-GB forecast con 60, Lab 18, Oth 0(actual Con 60, Lab 17, SNP 1. the 3 Northern Ireland seats were done under the STV system and weren't forecast)-source of information for forecast was turnout figures made available before the count

The BBC published an estimate of the Euros in 1979 based on the general election result (based on the constituencies that had declared) and I believe came to a roughly similar figure although they admitted that "Highlands and Islands could be a law unto itself!"

[nodealbrexiteer]

yes i saw the European election projection on the recent BBC Parliament repeat but can't remember the exact figures but i'm sure someone can locate it on youtube

[Harry Hayfield]

Aug 11, 2019 19:52:01 GMT nodealbrexiteer said:

yes i saw the European election projection on the recent BBC Parliament repeat but can't remember the exact figures but i'm sure someone can locate it on youtube

This is why I downloaded all the BBC elections from YouTube when I found out that I couldn't transfer my recordings onto a disc when my DVD player / recorder packed up in 2017. The forecast is made at 7.15am after 510 results and comes out as follows:

Conservatives 47 MEP's
Labour 30 MEP's
NI Unionists 2 MEP's
NI Nationalists 1 MEP
No results from Highlands and Islands to base a forecast on

So even the BBC wasn't expecting as many Conservative MEP's as they got suggesting that Labour voters simply refused to vote at the Euros (now, where have we heard that before, eh?)

[nodealbrexiteer]

Aug 12, 2019 12:44:40 GMT Harry Hayfield said:

Aug 11, 2019 19:52:01 GMT nodealbrexiteer said:

yes i saw the European election projection on the recent BBC Parliament repeat but can't remember the exact figures but i'm sure someone can locate it on youtube

This is why I downloaded all the BBC elections from YouTube when I found out that I couldn't transfer my recordings onto a disc when my DVD player / recorder packed up in 2017. The forecast is made at 7.15am after 510 results and comes out as follows:

Conservatives 47 MEP's
Labour 30 MEP's
NI Unionists 2 MEP's
NI Nationalists 1 MEP
No results from Highlands and Islands to base a forecast on

So even the BBC wasn't expecting as many Conservative MEP's as they got suggesting that Labour voters simply refused to vote at the Euros (now, where have we heard that before, eh?)

the euro constituencies were non overlapping aggregates of Westminster constituencies so the forecasters had an idea of who may have won them on the votes of the general election-the turnout based pre results forecasts was based on the following assumptions:
someone who abstained in the general election would abstain again and someone who voted for party would abstain or vote for the same party.Look for the following article
Forecasting the British Election to the European Parliament Vol. 11, No. 2 (Apr., 1981), pp. 235-245 for more details.

[Davıd Boothroyd]

Chris Cook and Mary Francis's 'The First European Elections - A Handbook and Guide' includes quick analyses of the constituencies' likely voting, in Chapter 9.

[nodealbrexiteer]

pdfslide.net/documents/statistical-methods-for-election-forecasting-in-the-united-kingdom-197090.html

pdfslide.net/documents/the-polls-have-a-lot-to-answer-for.html

2 articles above can be viewed/ I'll try and put some more links to all the exit poll articles I found

[nodealbrexiteer]

www.jstor.org/stable/2345211?seq=1#page_scan_tab_contents

1974 elections above-you can register on JSTOR and read 6 articles for free over a set period (BBC with some minor commentary on ITN)

www.jstor.org/stable/2680579?seq=1#page_scan_tab_contents

1997 election above (BBC)

www.jstor.org/stable/2987852?seq=1#page_scan_tab_contents

1983 election above (BBC with brief nod to ITN)

www.jstor.org/stable/30135085?seq=1#page_scan_tab_contents

2005 election above-joint exit poll

www.jstor.org/stable/40925394?seq=1#page_scan_tab_contents

2010 election above-joint exit poll

eprints.lse.ac.uk/83556/1/Kuha_Focus%20%202017%20exit%20poll_2017.pdf

2017 election exit poll above

eprints.ncrm.ac.uk/1582/1/Fisher_Kuha_exitpolltalk_ncrm1.pdf

2010 election exit poll again above


www.ipsos.com/ipsos-mori/en-uk/1997-itnmori-general-election-exit-poll

ITN 1997 Mori exit poll above

[nodealbrexiteer]

www.nuffield.ox.ac.uk/politics/papers/2002/w8/forecast.pdf

plus the above

www.jstor.org/stable/193585?seq=1#page_scan_tab_contents

1979 European election above BBC

[nodealbrexiteer]

I should make a couple of corrections to my general election exit polls messages from the 'interesting electoral facts' thread-The October 1974 Keighley poll reference should read a rare Tory overestimate in this election(not underestimate) and the 1997 ITN/MORI poll was conducted in Tory seats with a notional 1992 lead of up to 20%(not 10%)-apologies

[nodealbrexiteer]

The BBC/October 1974 Scotland (computer based) prediction based on their exit poll is interesting:


Con 15.6(actual 16),Lab 38.5(40),Lib 3.4(4), SNP 13.5(11)-remembering that the forecasts are sum of the probabilities of each party winning each seat in Scotland hence the forecasts are not whole numbers!-this is from the article below:


www.jstor.org/stable/2345211?read-now=1&seq=21#page_scan_tab_contents

[nodealbrexiteer]

Aug 28, 2019 21:50:45 GMT nodealbrexiteer said:

The BBC/October 1974 Scotland (computer based) prediction based on their exit poll is interesting:


Con 15.6(actual 16),Lab 38.5(40),Lib 3.4(4), SNP 13.5(11)-remembering that the forecasts are sum of the probabilities of each party winning each seat in Scotland hence the forecasts are not whole numbers!-this is from the article below:


www.jstor.org/stable/2345211?read-now=1&seq=21#page_scan_tab_contents

The same article suggest the first Liberal forecast was 14.3 and Plaid 1.9 so if you round Lib to 14, assume 12 Northern Ireland and round SNP and PLaid up to 14 and 2 I think the computer based exit poll forecast(rather than the non computer based one used on air) is indeed close to or is the one displayed on the prediction board after 2 results of 332 Lab, Con 261, Lib 14, Other 28! If only we could find the pre results one for February 1974..!

[nodealbrexiteer]

Welsh Assembly 1999:(from The Guardian)

A BBC Wales exit poll of 4,000 voters tonight projected that Labour would take between 28 and 32 of the 60 Welsh Assembly seats.
Plaid Cymru was forecast to win between 13 and 17 seats, with the Conservatives between seven and 11 and the Liberal Democrats between four and eight.

Results of today's election will not be declared until tomorrow, after counting begins across Wales at 9.30am.

But tonight's BBC poll clearly makes the Welsh Nationalists the main threat to Labour and could deny Labour an overall majority in the Assembly.

Details of the BBC poll were released as soon as polling stations closed at 10pm after the historic first vote for the Welsh Assembly.

In the poll of 4,000 electors, Labour was given 45% in the first past the post constituency seats, with Plaid Cymru on 26%, Conservatives 14% and Lib Dems 12%.












In the regional "top up" list seats, Labour had 41% support, with Plaid Cymru on 28%. Conservatives and Lib Dems were equally supported at 13%.

[nodealbrexiteer]

Scottish Parliament poll 1999(BBC/ICM)-from YouTube video:

First past the post vote:42% lab, 30% SNP, 13% Con, 11% Lib Dem 1st vote.2nd vote:Lab 38/SNP 31/Con 13/Lib dem 11-vote share.


seats:Lab 55-61, SNP 41-47, Con 11-17,Lib Dem 10-16,Oth 0.

The ICM polls were more like the on the day polls that the BBC used for the 1983 and 1987 General elections rather than actual exit polls