Foggy
Non-Aligned
Er/ihn/ihm
Posts: 4,967 
|
Post by Foggy on Mar 9, 2018 1:19:13 GMT
I remember in 2015 when the BBC were reporting that Labour had apparently held some 10 seats in Scotland including East Renfrewshire and Glasgow North East and the Lib Dems were on track to hold Edinburgh West... In 2017 I believe they were briefing that Perth & North Perthshire was a conservative gain, otherwise I think they got everything else bang on. I think what caused confusion in 2017 was the BBC trying to use the exit poll figures to forecast individual seats, and to be fair they repeatedly put health warnings on their Scottish projections I presume because of a relatively small Scottish sample in the exit poll (which may also explain why they were badly off in North Wales initially, forecasting Labour losing Wrexham, Clwyd South and Ynys Mön). *Môn The regional swing model isn't reliable for predicting individual seats. In Wales you may well be right that a small sample size was the problem. In Scotland it was more of a problem that using 2015 as a base didn't help with clarity as to which party would be the main beneficiary of any SNP collapse, especially in the North East (where the Lib Dems had finished second in many seats that were ultimately won by the Conservatives). Nevertheless the exit poll for the last GE came remarkably close to getting the national seat totals spot on, which is its main purpose. Broadcasters shouldn't pretend that it can be extrapolated to point to a particular result in a single contest, and viewers shouldn't believe them when they do so... then again, they do have a few hours to fill with chatter before most of the results start pouring in! |
|
|
|
Post by timmullen1 on Mar 9, 2018 1:23:48 GMT
I think what caused confusion in 2017 was the BBC trying to use the exit poll figures to forecast individual seats, and to be fair they repeatedly put health warnings on their Scottish projections I presume because of a relatively small Scottish sample in the exit poll (which may also explain why they were badly off in North Wales initially, forecasting Labour losing Wrexham, Clwyd South and Ynys Mön). *Môn The regional swing model isn't reliable for predicting individual seats. In Wales you may well be right that a small sample size was the problem. In Scotland it was more of a problem that using 2015 as a base didn't help with clarity as to which party would be the main beneficiary of any SNP collapse, especially in the North East (where the Lib Dems had finished second in many seats that were ultimately won by the Conservatives). Nevertheless the exit poll for the last GE came remarkably close to getting the national seat totals spot on, which is its main purpose. Broadcasters shouldn't pretend that it can be extrapolated to point to a particular result in a single contest, and viewers shouldn't believe them when they do so... then again, they do have a few hours to fill with chatter before most of the results start pouring in! Sorry about Môn, fat finger syndrome; I should know better as we used to go on childhood holidays there for the first 19 years of my life. Aside from that, yes to your analysis, although if they were going to do that with the exit poll it does beg the question why they didn’t do a few seat specific polls as they did in (I think) October 1974 when they certainly successfully picked up Bob Cryer gaining Keighley for Labour. |
|
Foggy
Non-Aligned
Er/ihn/ihm
Posts: 4,967
|
Post by Foggy on Mar 9, 2018 1:33:21 GMT
*Môn The regional swing model isn't reliable for predicting individual seats. In Wales you may well be right that a small sample size was the problem. In Scotland it was more of a problem that using 2015 as a base didn't help with clarity as to which party would be the main beneficiary of any SNP collapse, especially in the North East (where the Lib Dems had finished second in many seats that were ultimately won by the Conservatives). Nevertheless the exit poll for the last GE came remarkably close to getting the national seat totals spot on, which is its main purpose. Broadcasters shouldn't pretend that it can be extrapolated to point to a particular result in a single contest, and viewers shouldn't believe them when they do so... then again, they do have a few hours to fill with chatter before most of the results start pouring in! Sorry about Môn, fat finger syndrome; I should know better as we used to go on childhood holidays there for the first 19 years of my life. Aside from that, yes to your analysis, although if they were going to do that with the exit poll it does beg the question why they didn’t do a few seat specific polls as they did in (I think) October 1974 when they certainly successfully picked up Bob Cryer gaining Keighley for Labour. No worries. Having honed my French and German skills just across the Menai Strait from there, I am quick to spot a rogue umlaut where a circumflex should be! That's not begging the question (  ), but I think by last year Lord Ashcroft's attempts at constituency-level polling from before the 2015 GE had convinced them that it wouldn't be worth the hassle. |
|
|
|
Post by conservativeestimate on Mar 9, 2018 2:25:13 GMT
The BBC predicted the Lib Dems would regain Gordon - they actually ended up with 11.6% of the vote...
|
|
|
|
Post by timmullen1 on Mar 9, 2018 2:27:51 GMT
Sorry about Môn, fat finger syndrome; I should know better as we used to go on childhood holidays there for the first 19 years of my life. Aside from that, yes to your analysis, although if they were going to do that with the exit poll it does beg the question why they didn’t do a few seat specific polls as they did in (I think) October 1974 when they certainly successfully picked up Bob Cryer gaining Keighley for Labour. No worries. Having honed my French and German skills just across the Menai Strait from there, I am quick to spot a rogue umlaut where a circumflex should be! That's not begging the question ( ), but I think by last year Lord Ashcroft's attempts at constituency-level polling from before the 2015 GE had convinced them that it wouldn't be worth the hassle. I think one of the problems with the Ashcroft constituency polls was most of them were done way before polling day and thus missed any late movement to or from a Party. A constituency based exit poll should(!) overcome that. |
|
nodealbrexiteer
Forum Regular
non aligned favour no deal brexit!
Posts: 506
|
Post by nodealbrexiteer on Sept 13, 2020 15:30:53 GMT
*Môn The regional swing model isn't reliable for predicting individual seats. In Wales you may well be right that a small sample size was the problem. In Scotland it was more of a problem that using 2015 as a base didn't help with clarity as to which party would be the main beneficiary of any SNP collapse, especially in the North East (where the Lib Dems had finished second in many seats that were ultimately won by the Conservatives). Nevertheless the exit poll for the last GE came remarkably close to getting the national seat totals spot on, which is its main purpose. Broadcasters shouldn't pretend that it can be extrapolated to point to a particular result in a single contest, and viewers shouldn't believe them when they do so... then again, they do have a few hours to fill with chatter before most of the results start pouring in! Sorry about Môn, fat finger syndrome; I should know better as we used to go on childhood holidays there for the first 19 years of my life. Aside from that, yes to your analysis, although if they were going to do that with the exit poll it does beg the question why they didn’t do a few seat specific polls as they did in (I think) October 1974 when they certainly successfully picked up Bob Cryer gaining Keighley for Labour. Sorry to be pedantic but the BBC forecast Keighley as a hold for Bob Cryer in the Marplan exit poll in October 1974. I guess the seat specific polls stopped after 1979 as the record was scratchy after the 1970 Gravesend poll and with greater 3rd party votes and more diverse swings they weren't deemed as useful |
|
|
|
Post by heslingtonian on Sept 17, 2020 21:39:57 GMT
The BBC predicted the Lib Dems would regain Gordon - they actually ended up with 11.6% of the vote... Talking of Gordon, the BBC claimed it would be the only Conservative seat in Scotland for much of the 1997 election broadcast. |
|
|
|
Post by bjornhattan on Sept 17, 2020 21:57:31 GMT
One rumour that I wish had been true from the EU referendum - there was talk of even larger Leave votes in some areas than we actually got. One of the BBC's reporters in the Midlands claimed that Leave were on 80% in one of the Warwickshire districts (either Nuneaton and Bedworth or North Warwickshire, can't recall which). Sadly, these extraordinary figures never came to pass.
|
|
Harry Hayfield
Green
has voted for the Green Party for the first time at a Westminster General Election
Posts: 2,003
|
Post by Harry Hayfield on Sept 18, 2020 7:54:56 GMT
BBC Exit Poll 2019 was full of them: Lab HOLD Blyth Valley, Lib Dem GAIN Sheffield, Hallam from Lab, Lib Dem GAIN Ceredigion from Plaid, Lab GAIN Cities of London and Westminster from Con.
|
|
Harry Hayfield
Green
has voted for the Green Party for the first time at a Westminster General Election
Posts: 2,003
|
Post by Harry Hayfield on Sept 18, 2020 7:59:12 GMT
The BBC predicted the Lib Dems would regain Gordon - they actually ended up with 11.6% of the vote... Talking of Gordon, the BBC claimed it would be the only Conservative seat in Scotland for much of the 1997 election broadcast. That's because the boundary changes turned it from a Lib Dem marginal in 1992 (Lib Dem 37.4%, Con 37.0%) into one of the safest Conservative seats in Scotland (Con 48%, Lib Dem 27%) |
|
The Bishop
Labour
Down With Factionalism!
Posts: 26,761
|
Post by The Bishop on Sept 18, 2020 9:58:22 GMT
Though the actual result made it clear that the notional 1992 outcome had over-egged the Tory lead there - one of a handful of instances where they were significantly out.
|
|
|
|
Post by iainbhx on Sept 18, 2020 13:28:56 GMT
Though the actual result made it clear that the notional 1992 outcome had over-egged the Tory lead there - one of a handful of instances where they were significantly out. When you get significant numbers of Indos or only partial slates, reallocating local results gets tricksy. From what I remember in 1997, we thought Gordon was probably Tory in 1992, but not by that much. Wasn't Basildon one of the others? |
|
), but I think by last year Lord Ashcroft's attempts at constituency-level polling from before the 2015 GE had convinced them that it wouldn't be worth the hassle.
