jamie
Top Poster
Posts: 6,893
|
Post by jamie on Dec 7, 2019 22:23:20 GMT
At this rate some of the pollsters are going to have egg on their faces. There is a big difference between a 6 and 15 point lead - a small majority, or a landslide. If the Tories were around 10% ahead then that’s the sort of distribution you would expect. At least they are not obviously herding.
|
|
|
Opinium
Dec 7, 2019 22:43:54 GMT
via mobile
Post by andrew111 on Dec 7, 2019 22:43:54 GMT
At this rate some of the pollsters are going to have egg on their faces. There is a big difference between a 6 and 15 point lead - a small majority, or a landslide. If the Tories were around 10% ahead then that’s the sort of distribution you would expect. At least they are not obviously herding. No, it is way outside statistical error. That means there are systematic errors in some polls, but we don't know which. The full range from 3 to 18% (including statistical error) is therefore possible if we give all polls the same chance of being "correct"
|
|
The Bishop
Labour
Down With Factionalism!
Posts: 36,813
|
Post by The Bishop on Dec 8, 2019 10:26:40 GMT
The point is - and hard votes may of course prove this wrong in the coming week - this doesn't *feel* like a landslide election, a 1983 or a 1997.
In both those elections cheering crowds gathered spontaneously to greet the winner, in this one Johnson cancels visits because a handful of people want to protest against him (and could you imagine either vintage Thatcher or Blair dodging Andrew Neil, come to that?)
|
|
pl
Non-Aligned
Posts: 1,568
|
Opinium
Dec 8, 2019 10:50:50 GMT
via mobile
Post by pl on Dec 8, 2019 10:50:50 GMT
The point is - and hard votes may of course prove this wrong in the coming week - this doesn't *feel* like a landslide election, a 1983 or a 1997. In both those elections cheering crowds gathered spontaneously to greet the winner, in this one Johnson cancels visits because a handful of people want to protest against him (and could you imagine either vintage Thatcher or Blair dodging Andrew Neil, come to that?) I agree re landslide. I think the whole climate had changed since Thatcher/Blair though. Small protests get magnified by 24 hour rolling news and YouTube. Interviewers are also more combative now and trying to set the agenda. Also, 30 years ago one bad interview was forgotten, now it is endlessly replayed until the end of time.
|
|
|
Post by Merseymike on Dec 8, 2019 12:08:16 GMT
The point is - and hard votes may of course prove this wrong in the coming week - this doesn't *feel* like a landslide election, a 1983 or a 1997. In both those elections cheering crowds gathered spontaneously to greet the winner, in this one Johnson cancels visits because a handful of people want to protest against him (and could you imagine either vintage Thatcher or Blair dodging Andrew Neil, come to that?) I agree re landslide. I think the whole climate had changed since Thatcher/Blair though. Small protests get magnified by 24 hour rolling news and YouTube. Interviewers are also more combative now and trying to set the agenda. Also, 30 years ago one bad interview was forgotten, now it is endlessly replayed until the end of time. I think you are right, and of course it suggests that more and more campaigns will be based on dull, endlessly repreated slogans and people only being interviewed by those they perceive as sufficiently soft and/or sympathetic This goes across the parties. Personally, I've been underwhelmed by this election. I think its partially because so much of it has been about Brexit, and I think what has been said is such utter tosh that I find it hard to grasp how people can be so thick as to believe it (notably the idea that someohow the passing of the WA will get anything 'done' at all in any practical sense)
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Dec 8, 2019 12:17:43 GMT
I agree re landslide. I think the whole climate had changed since Thatcher/Blair though. Small protests get magnified by 24 hour rolling news and YouTube. Interviewers are also more combative now and trying to set the agenda. Also, 30 years ago one bad interview was forgotten, now it is endlessly replayed until the end of time. I think you are right, and of course it suggests that more and more campaigns will be based on dull, endlessly repreated slogans and people only being interviewed by those they perceive as sufficiently soft and/or sympathetic This goes across the parties. Personally, I've been underwhelmed by this election. I think its partially because so much of it has been about Brexit, and I think what has been said is such utter tosh that I find it hard to grasp how people can be so thick as to believe it (notably the idea that someohow the passing of the WA will get anything 'done' at all in any practical sense) I think that's broadly right, though the "tosh" has spanned the political spectrum. The campaign by all parties has been low-key, risk-averse and joyless, with a public mood of glum resignation rather than enthusiasm. EDIT I unaccountably left out the word "dishonest". Yes, that was an important omission.
|
|
|
Post by Adam in Stroud on Dec 8, 2019 12:23:33 GMT
The point is - and hard votes may of course prove this wrong in the coming week - this doesn't *feel* like a landslide election, a 1983 or a 1997. In both those elections cheering crowds gathered spontaneously to greet the winner, in this one Johnson cancels visits because a handful of people want to protest against him (and could you imagine either vintage Thatcher or Blair dodging Andrew Neil, come to that?) If there is a landslide, I don't think it will imply massive enthusiasm for either Johnson or his programme (such as it is.) His campaign is clearly minimalist, based on "get Brexit done" and vox pops consistently show us people saying they will vote for him on that basis, but not generally with much enthusiasm. The impression given is that they maintain the right to resume voting Labour as soon is it is "done" Of course once in power with a workable majority all sorts of opportunities to build support will arise and certainly a Brexit regarded as successful or even as just good enough to put the issue to bed will be one. Conversely a recession or a bodged or drawn-out Brexit will see support drain rapidly. But it could be a John Major-like govt, where support was lost within the first year but the govt dragged on for the full-term in the hope of things turning round.
|
|
|
Post by Merseymike on Dec 8, 2019 12:24:17 GMT
I think you are right, and of course it suggests that more and more campaigns will be based on dull, endlessly repreated slogans and people only being interviewed by those they perceive as sufficiently soft and/or sympathetic This goes across the parties. Personally, I've been underwhelmed by this election. I think its partially because so much of it has been about Brexit, and I think what has been said is such utter tosh that I find it hard to grasp how people can be so thick as to believe it (notably the idea that someohow the passing of the WA will get anything 'done' at all in any practical sense) I think that's broadly right, though the "tosh" has spanned the political spectrum. The campaign by all parties has been low-key, risk-averse and joyless, with a public mood of glum resignation rather than enthusiasm. I'm not disagreeing!
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Dec 8, 2019 12:29:43 GMT
The point is - and hard votes may of course prove this wrong in the coming week - this doesn't *feel* like a landslide election, a 1983 or a 1997. In both those elections cheering crowds gathered spontaneously to greet the winner, in this one Johnson cancels visits because a handful of people want to protest against him (and could you imagine either vintage Thatcher or Blair dodging Andrew Neil, come to that?) If there is a landslide, I don't think it will imply massive enthusiasm for either Johnson or his programme (such as it is.) His campaign is clearly minimalist, based on "get Brexit done" and vox pops consistently show us people saying they will vote for him on that basis, but not generally with much enthusiasm. The impression given is that they maintain the right to resume voting Labour as soon is it is "done" Of course once in power with a workable majority all sorts of opportunities to build support will arise and certainly a Brexit regarded as successful or even as just good enough to put the issue to bed will be one. Conversely a recession or a bodged or drawn-out Brexit will see support drain rapidly. But it could be a John Major-like govt, where support was lost within the first year but the govt dragged on for the full-term in the hope of things turning round. I don't altogether disagree. A decent majority would at least give the Conservatives a chance to make a go of it and reinvent themselves. However, the danger of a small majority is that it will be 1992 all over again - a victory that quickly became a suspended death sentence.
|
|
|
Post by Adam in Stroud on Dec 8, 2019 12:32:12 GMT
We're dangerously close to all agreeing on this thread atm. Probably means we're all talking bollocks and it'll be an Ein Gwlad majority govt. Anyone know what their Brexit policy is?
|
|
|
Post by Merseymike on Dec 8, 2019 12:43:12 GMT
Transfer the European Parliament and Commission to Blaenau Ffestiniog?
|
|
|
Post by finsobruce on Dec 8, 2019 13:10:37 GMT
Transfer the European Parliament and Commission to Blaenau Ffestiniog? They'd get slated for doing that.
|
|
pl
Non-Aligned
Posts: 1,568
|
Opinium
Dec 8, 2019 13:46:01 GMT
via mobile
jamie likes this
Post by pl on Dec 8, 2019 13:46:01 GMT
Transfer the European Parliament and Commission to Blaenau Ffestiniog? And Welsh to be added as a working language of the Commission alongside French and German, replacing English.
|
|
Vibe
Non-Aligned
Posts: 931
|
Post by Vibe on Dec 8, 2019 14:15:15 GMT
The point is - and hard votes may of course prove this wrong in the coming week - this doesn't *feel* like a landslide election, a 1983 or a 1997. In both those elections cheering crowds gathered spontaneously to greet the winner, in this one Johnson cancels visits because a handful of people want to protest against him (and could you imagine either vintage Thatcher or Blair dodging Andrew Neil, come to that?) A lot of people faced with a bad choice of Prime Minster are begrudgingly voting for one to stop the other, so we won't get cheering crowds. It's like do you want to your cat or dog to be run over!
|
|
|
Post by Robert Waller on Dec 8, 2019 14:37:21 GMT
If the Tories were around 10% ahead then that’s the sort of distribution you would expect. At least they are not obviously herding. No, it is way outside statistical error. That means there are systematic errors in some polls, but we don't know which. The full range from 3 to 18% (including statistical error) is therefore possible if we give all polls the same chance of being "correct" Not quite sure what Andrew is saying, but maybe worth pointing out explicitly that the 'margin of error' (based on sample size) is for individual party vote shares so has to be greatly increased for the 'lead'. For example, with a MoE of 3% + or -, a Conservative share of say 42% means 39 - 45%, Labour say 32% means 29 - 35, so the lead would be somewhere between 4% and 16% (this is one of the reasons why it is so ridiculous to make a lot of apparent small changes in lead in any individual poll). However, as I'm sure I've said many times before, the margin of error is in my view fallacious as there are other sources of error as well such as weighting technique and inaccurate sampling. (Therefore Andrew is right in his implication that some polling companies are likely to be using more inaccurate methodologies when it comes to the final 'predictions' than others). It's not an exact science, there is a lot of fuzziness ... or perhaps it is therefore like the science of quantum physics ...!
|
|
Vibe
Non-Aligned
Posts: 931
|
Opinium
Dec 8, 2019 14:55:56 GMT
via mobile
Post by Vibe on Dec 8, 2019 14:55:56 GMT
No, it is way outside statistical error. That means there are systematic errors in some polls, but we don't know which. The full range from 3 to 18% (including statistical error) is therefore possible if we give all polls the same chance of being "correct" Not quite sure what Andrew is saying, but maybe worth pointing out explicitly that the 'margin of error' (based on sample size) is for individual party vote shares so has to be greatly increased for the 'lead'. For example, with a MoE of 3% + or -, a Conservative share of say 42% means 39 - 45%, Labour say 32% means 29 - 35, so the lead would be somewhere between 4% and 16% (this is one of the reasons why it is so ridiculous to make a lot of apparent small changes in lead in any individual poll). However, as I'm sure I've said many times before, the margin of error is in my view fallacious as there are other sources of error as well such as weighting technique and inaccurate sampling. (Therefore Andrew is right in his implication that some polling companies are likely to be using more inaccurate methodologies when it comes to the final 'predictions' than others). It's not an exact science, there is a lot of fuzziness ... or perhaps it is therefore like the science of quantum physics ...! Pretty poor if a pollster can't be a bit more accurate than the lead is somewhere between 4% and 16%. Are they really that imprecise?
|
|
|
Post by Robert Waller on Dec 8, 2019 15:55:16 GMT
They'd say that the 'true figure' would be more likely to be in the middle of the range. (On the other hand they should admit a 5 per cent chance it's outside even this range), I wouldn't say so much that it's poor as that it's a statistical fact. What's poor in my view is the over-analysing of and over-reaction to single polls! On the other hand , take lots of polls together and that's the best guide to the state of the race we shall get .
|
|
Chris from Brum
Lib Dem
What I need is a strong drink and a peer group.
Posts: 9,241
|
Opinium
Dec 8, 2019 18:24:44 GMT
via mobile
jamie likes this
Post by Chris from Brum on Dec 8, 2019 18:24:44 GMT
Transfer the European Parliament and Commission to Blaenau Ffestiniog? And Welsh to be added as a working language of the Commission alongside French and German, replacing English. They'd have to get that past the Irish, the Maltese and possibly the Cypriots first.
|
|
|
Opinium
Dec 8, 2019 23:30:31 GMT
via mobile
Post by andrew111 on Dec 8, 2019 23:30:31 GMT
No, it is way outside statistical error. That means there are systematic errors in some polls, but we don't know which. The full range from 3 to 18% (including statistical error) is therefore possible if we give all polls the same chance of being "correct" Not quite sure what Andrew is saying, but maybe worth pointing out explicitly that the 'margin of error' (based on sample size) is for individual party vote shares so has to be greatly increased for the 'lead'. For example, with a MoE of 3% + or -, a Conservative share of say 42% means 39 - 45%, Labour say 32% means 29 - 35, so the lead would be somewhere between 4% and 16% (this is one of the reasons why it is so ridiculous to make a lot of apparent small changes in lead in any individual poll). However, as I'm sure I've said many times before, the margin of error is in my view fallacious as there are other sources of error as well such as weighting technique and inaccurate sampling. (Therefore Andrew is right in his implication that some polling companies are likely to be using more inaccurate methodologies when it comes to the final 'predictions' than others). It's not an exact science, there is a lot of fuzziness ... or perhaps it is therefore like the science of quantum physics ...! The error of differences is normally calculated as the square root of the sum of the squares, so if the error in a single measurement is 3% the error in the difference would be 4.2%. It may be a little more complicated in this case since we are not talking about independent measurements. However you certainly can't just assume the error of difference is +/- 6% as you do. The maximum range shoukd be around 8.5%, and since we are talking 95% confidence intervals there is only a 5% likelihood of getting that range between two successive polls. Anyway I think we agree that the large range in Tory lead is due to variations in methodology. (and we don't know which if any are correct)
|
|
|
Post by Robert Waller on Dec 11, 2019 13:26:02 GMT
Westminster voting intention:
CON: 45% (-1) LAB: 33% (+2) LDEM: 12% (-1)
via @opiniumresearch , 10 - 11 Dec Chgs. w/ 06 Dec
|
|