|
Post by Andrew_S on May 27, 2017 19:17:46 GMT
The Tory honeymoon is going to be very very shortlived. The council elections next year could be very ugly. It would be unusual if the local election results were anything other than bad for a party that would by then have been in government for 8 years.
|
|
|
Post by Old Fashioned Leftie on May 27, 2017 19:57:28 GMT
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on May 27, 2017 21:13:49 GMT
ComRes always weight their findings "bigly" against Labour - apparently the original figures here show a Tory lead of less than 4 points. The presumption of them - and other pollsters - is that turnout patterns this time will be the same as two years ago. But what if they aren't? Indeed, the problem with this approach is that you could receive data indicating a deviation from past trend and deliberately ignore it because... it deviates from past trend. Which is actually fine if you're producing a forecast, but pollsters often deny that that is what they do...
|
|
|
Post by Devil Wincarnate on May 27, 2017 21:22:50 GMT
Nah, even if Brexit is as unpleasant as you think it wouldnt be in time for 2018. Arent you making the risk of thinking that everyone views the world like you do? Nope. 60 bn quid and rising budget deficit, council cuts and tax rises to come in November. Brexit is the icing on the cake. The social fabric is crumbling (school classes getting bigger, running out of cash) with more to come next year. There will be a need to raise revenue. CGT on house sales perhaps, raiding FSP lump sums? This is a binary ge. And cleverly framed as such. I would remind you that Farron thought it was binary, and the polls are not just biting him on the arse but chewing on his perineum.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Jun 3, 2017 17:14:33 GMT
|
|
|
Post by marksenior on Jun 3, 2017 17:42:38 GMT
It is all in the weightings . Look at Table 3 and adjust for the 10% Don't Knows , Con 43 Lab 38.5 LD 8 , almost identical figures for today's 2 polls before the pollsters manipulate sorry do their weighting adjustments . Which pollster if any are making the correct adjustments , we won't know till next week .
|
|
|
Post by gwynthegriff on Jun 3, 2017 17:48:52 GMT
It is all in the weightings . Look at Table 3 and adjust for the 10% Don't Knows , Con 43 Lab 38.5 LD 8 , almost identical figures for today's 2 polls before the pollsters manipulate sorry do their weighting adjustments . Which pollster if any are making the correct adjustments , we won't know till next week . We may not know even then. Polls are a snapshot, not a prediction. Actually, I'm increasingly of the mind that they're a waste of effort.
|
|
|
Post by Andrew_S on Jun 3, 2017 18:07:57 GMT
It is all in the weightings . Look at Table 3 and adjust for the 10% Don't Knows , Con 43 Lab 38.5 LD 8 , almost identical figures for today's 2 polls before the pollsters manipulate sorry do their weighting adjustments . Which pollster if any are making the correct adjustments , we won't know till next week . We may not know even then. Polls are a snapshot, not a prediction. Actually, I'm increasingly of the mind that they're a waste of effort. It's more and more difficult to get a representative sample these days. That's why there have to be so many adjustments to the raw data.
|
|
|
Post by curiousliberal on Jun 3, 2017 18:12:54 GMT
It's more and more difficult to get a representative sample these days. That's why there have to be so many adjustments to the raw data. Out of interest, why do you think this is the case?
|
|
|
Post by Davıd Boothroyd on Jun 3, 2017 18:22:09 GMT
On looking at unweighted polls and drawing conclusions from them, there was a great analogy by Lewis Baston today:
|
|
|
Post by marksenior on Jun 3, 2017 18:26:53 GMT
It is not the adjustments to the raw data to cater for sampling issues that is causing the variations . It is the in house pollsters adjustments for turnout etc. See the differences in this poll between Table 3 and Table 6 VI figures .
|
|
rocky
Non-Aligned
Posts: 122
|
Post by rocky on Jun 3, 2017 19:19:19 GMT
It is all in the weightings . Look at Table 3 and adjust for the 10% Don't Knows , Con 43 Lab 38.5 LD 8 , almost identical figures for today's 2 polls before the pollsters manipulate sorry do their weighting adjustments . Which pollster if any are making the correct adjustments , we won't know till next week . I think the thing we can all agree on is some pollsters will be in hot water next week. Maybe all of them. Not necessarily a bad thing. I prefer to consult my tea leaves. Yep either some pollsters are in trouble or the party strategists will be because no party is fighting the campaign the current polls point towards.
|
|
|
Post by John Chanin on Jun 3, 2017 19:19:45 GMT
It is not the adjustments to the raw data to cater for sampling issues that is causing the variations . It is the in house pollsters adjustments for turnout etc. See the differences in this poll between Table 3 and Table 6 VI figures . This needs repeating, although it is not quite as straightforward as Mark suggests. The sampling problems are quite subtle, and difficult to adjust for, as it's basically about young people who don't know anything and care less about politics, and are virtually unreachable. It seems that they are considerably more likely to vote Conservative than other young people, although their propensity to vote is also low. One can of course speculate why this is so, but that is a matter for a different discussion. YouGov have compensated for this through adjustments to their sample weights, while leaving propensity to vote as self reported as before (there are adjustments according to whether they actually voted last time, but these aren't new). ComRes have chosen to ignore self-reported propensity to vote, and are simply assuming that different age groups will vote in the same proportions as last time. Basically this is crap, as we know that turnout has varied very substantially from one election to another amongst young people, and more so than older groups. There is also an unacknowledged problem that a significant proportion of young people are registered in 2 places (probably around 10%), which will depress reported turnout in constituencies, while not of course depressing turnout as seen by polling companies. This 10% is probably more sophisticated(ie current students not living at home), and more likely to vote where it matters most. One of the lessons of this campaign is surely that more young people are motivated to vote this time than in 2015. So ComRes are on a bad bet in my view.
|
|
|
Post by marksenior on Jun 7, 2017 18:33:30 GMT
Final Comres
Con 44 -3 Lab 34 -1 LDem 9 plus 1 UKIP 5 plus 1 Green 2 plus 1
|
|
|
Post by Devil Wincarnate on Jun 7, 2017 18:48:28 GMT
Comedy Results probably living up to their name.
|
|
|
Post by Pete Whitehead on Jun 7, 2017 18:51:36 GMT
It's similar to the ICM one - looks plausible enough
|
|
wallington
Green
The Pride of Croydon 2022 award winner
Posts: 1,246
|
Post by wallington on Jun 7, 2017 19:40:54 GMT
I would be very surprised if the Tories didn't win by around 10 points tomorrow. Anything less really is a pretty terrible result. I think Corbyn has done very well given the circumstances and Theresa May has been pretty horrific. If anything this campaign has shown her to be pretty poor.
|
|
|
Post by Adam in Stroud on Jun 7, 2017 19:54:45 GMT
I would be very surprised if the Tories didn't win by around 10 points tomorrow. Anything less really is a pretty terrible result. I think Corbyn has done very well given the circumstances and Theresa May has been pretty horrific. If anything this campaign has shown her to be pretty poor. Indeed. And I shall be going to the loo for a pretty in a minute.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Jun 11, 2017 23:59:19 GMT
ComRes were always smugly insistent that their post 2015 "cook the numbers to downgrade Labour" policy would work out on the night, so ha ha ha really.
|
|
The Bishop
Labour
Down With Factionalism!
Posts: 36,552
|
Post by The Bishop on Jun 12, 2017 10:03:51 GMT
Some of us who have maintained for years now that ComedyResults are a heap of s***e can feel smugly vindicated
|
|