|
By:
Thank Meadow,
http://survation.com/wp-content/uploads/2017/06/Final-GMB-GE2017-V-Tables-020617TOCH-1c1d0h3-microage.pdf So in this case on page 5 they have 1049 people likely to vote. 577 males and 472 females. This adds up to 1049 so I see how you can weight the votes to get more females than males to represent the actual likely demographics of voters. What I dont understand is the unweighted total for leave/remain being 363/563. This only adds up to 926 in the sample. Where are the other 123 people from the unweighted total sample? How do you weight 2 samples of different sizes? The weighted total is 1066 but the sum of weighted totals for leave/remain is 919. I am new to this and don't understand why they are not set to the same weighted total sample sizes and how the weighting are merged to get the overall polling result. Any explanation will be helpful for this specific poll. |
|
By:
What I dont understand is the unweighted total for leave/remain being 363/563. This only adds up to 926 in the sample. Where are the other 123 people from the unweighted total sample?
What makes you think the 123 people voted in the EU referendum? |
|
By:
He wasn't using political allegiance he was using canvassing returbs which were completely accurate,nothing like the polls they were clueless,i've been a labour voter all my life bar green in council elections and liberal tactical vote
and have never bacled labour. |
|
By:
Not saying anyone in particular, just a general comment.
|
|
By:
So why dont they have a weighted total to 1066?
Why are the weighted totals after the adjustment lower and how do you then merge that weighting with the other data? What are the actual calculations from the raw data to total weighted results when the weighted totals are different? |
|
By:
If they weight the EU vote and only 926 of the 1066 voted in the referendum why would they (or how can they) weight the folks who didn't vote in it?
|
|
By:
So you are saying they did not vote in the referendum?
We are supposed to take a poll seriously that suggestions 12% of people who intend to vote in this election did not vote in the referendum? In the referendum 72.2% of the electorate voted. In the last general election it was 66.1%. So it does not make sense. New voters signing up will not represent 12%+likely lower turnout difference. I am not convinced at all Labour will get 1% less votes than the Tories on Thursday based on this poll. |
|
By:
going at a bit of a tangent i would want my sample nearer to 52/48 in EU
vote to deviate so far from that and weight back is asking for trouble , imo |
|
By:
Also taking out the likely voter and going to the raw data they have 377 leave and 575 remain.
This totals 952 of people in 1103 have said how they voted in the referendum. This is 86% of the sample. 72.2% people voted in the referendum so this sample is not representative of the population. We then add up the raw data for 2015 votes. THIS INCLUDES 'DID NOT VOTE'. The total raw number in the survey is 1028. The total people asked is 1103. So where are the other 75 people from the raw data if did not vote is included as option? |
|
By:
one qualifying question for polls may be ...do you vote?
so more voters than standard in sample as dont vote is often termination of interview |
|
By:
True Donny but then it would not very well represent the likely finally numbers on the day.
The raw numbers dont add up even when they include a 'did not vote' option for 2015 election. So I cannot work out how they do the weighting even for that. Its like they pulled the results out of thin air. Would really like to see how the raw calculations actually work. |
|
By:
Insider Trader, why does the EU referendum voting weighting have a direct bearing on somehow favouring Labour?
Put simply, you do not accept the poll because it is not what you want it to be. |
|
By:
|
|
By:
Yes, this "certain to vote" weighting is the difference in 18-24 age groups from Comres to Survation.
|
|
By:
|
|
By:
On the rise of the Labour vote in England, perhaps the tuition fee issue is having a direct effect on young voters and their parents/guardians. In Scotland we already have free tuition fees so the surge is not as dramatic?
You could well be right on the Scottish turnout. |
|
By:
|
|
By:
|
|
By:
If you are a parent /grandparent/guardian of a child who is likely to be beginning university this year or next year then that £9k per year you will save under the Labour free tuition fees plan must be tempting.
If you are the family of elderly relatives who may need professional care in the foreseeable future, the thought of losing your inheritance through the "dementia tax" must be deeply worrying. On UKIP, initially I was seeing more than half of UKIP switchers intending to vote Tory and Labour in single figures. Now it seems it's around 35% Tory, 20% Labour, 25% UKIP, 20% Don't know. These are factors in my mind why Labour are closing the gap. |
|
By:
Yougov Scotland poll...
SNP 41% (-1) Con 26 (-3) Labour 25 (+6) LibDem 6 (nc) Fieldwork 1st-5th June. |
|
By:
Meadow X1
06 Jun 17 13:18 Joined: 07 Mar 05 | Topic/replies: 2,357 | Blogger: Meadow X1's blog Insider Trader, why does the EU referendum voting weighting have a direct bearing on somehow favouring Labour? Put simply, you do not accept the poll because it is not what you want it to be. ... No I am not saying that. I have not looked at the Comres one. I am sure there are flaws with that as well. I am looking for an explanation on the weighting calculations. I thought you would be the person who would know as you have made good predictions on the forum before. Did you work out the 2015 sample numbers when they include 'did not vote' the same only adds up to 1028 when the total raw sample is 1103? ... On your point : why does the EU referendum voting weighting have a direct bearing on somehow favouring Labour? That is my point we dont know how they do the calculations. Or if you do know you not sharing it with us. The numbers they provide come with no explanation. |
|
By:
CJ70 06 Jun 17 13:41
Jun 6, 2017 -- 11:58AM, thegiggilo wrote: I also remember that poster earlier cj saying several days in advance of last election that the labour vote was getting decimated by ukip from some canvassing that he had recieved,he's put up the 1.1 and 1.4 today,so i presume he's hearing the same again and he will be putting up bets against labour again.He was right as well as the polls didn't reflect anything that he knew from the doorstep.. I'm tied by what I can say as I'm actually working this time around. Many people who are canvassing are recording that the UKIP vote is being squeezed this time, mostly to Con but some blowback to Lab. Also pro-leave Lab voters are going Lab -> Con to complete the process of leaving the EU. If you add those issues to strong Lab voters refusing to vote for Corbyn you have a lot of previously safe Lab seats in play at pretty juicy prices. The only place I've genuinely heard optimism for Labour is within London. I suggest that demographics will play a bigger part of predicting those areas than following the polling.. As i suspected yet again the polls are total bollox.. |
|
By:
Insider, you seem to want to pick the tiniest holes in the Survation poll simply because it is not what you want to see. If you feel their methodology is flawed and you want the exact details to the decimal point you would be best speaking to a professional psephologist at Survation.
|
|
By:
Meadow if you dont understand how the numbers in that poll work thats fine. I dont and nor does anyone else on here. I was only asking if you knew why the unweighted sample totals didnt add up.
Looking at the Comres one and all their unweighted totals seem to add up. To state the obvious the key to this election is if young people vote this time. |
|
By:
|
|
By:
Its not a question of 'picking holes'. If that report was anything like accurate we would all be piling on Labour to win the most seats at 10.0
|
|
By:
It's not a report, it is a poll, a snapshot at that time.
|
|
By:
Ok if that poll was anything like accurate we would all be piling on Labour. The fact is the market does not trust it as a forecast of the future.
|
|
By:
I suspect that in reality the Tories are 6 to 8 points in front of Labour. Also there is a big chance that Labour is piling up votes in the wrong
seats while losing support in its typical working class heartland seats. |
|
By:
Spot on thats the exactly the way i've seen it since the start,just seems like common sense to me you don't need polls..
Meadow you are wrong about me wanting the polls to be wrong,i have a poster in my window with Dr laura davies on it.. |
|
By:
I think the gap is something like 42-38.
|
|
By:
I don't subscribe to the view that Labour is simply piling up votes in the wrong seats.
|
|
By:
Miliband got 30.4% of the vote. I would be astonished if Corbyn got less than that.
|
|
By:
Both should get higher with Lib Dems, UKIP and Greens dropping like flies. who will get the UKIP vote?
Are young uni types going to vote in numbers in the holidays? Are they going to just stack up votes in uni towns by postals or vote at their parents houses? I think alot will just miss the vote and complain after like they did with Brexit... but uni fees is a big motivator. |
|
By:
All the Uni's are still in term time....just about finishing exams now. I think the weather may be a bigger issue. It's forecast pretty wet for Thursday.
If only you could vote by text......Maybe Labour should put that in their next manifesto! |
|
By:
Didn't i read somewhere would take something like 80% of them to turn out to make a difference lol i'd be amaed iof 35% even knew what day it was never mind voted..
|
|
By:
If I said I would pay your £45k uni fees for the next five years would you vote for me?
|
|
By:
Problem is most uni seats and those with high levels of young voters are LD or Lab already. Lab should take Sheffield Hallam and Leeds North West from LD, but other seats are off limits.
Maybe Lab will get over Miliband in terms of percentage, but I can't see where they are going to pick up seats to beat 232? Maybe a couple off the SNP? Maybe a few off Con in Lab and Wales? Meanwhile Con will almost certainly have an increased vote share and Lab majoritys of 10k+ are in play. There's a lot of heavy lifting needed to be done by previous non-voters to get Lab anywhere near increasing their vote. |
|
By:
Yes. It is a very clever policy to win votes. Should have offered to pay off the debts of all the people who owe money for student loans as well. Would have got more votes of people up to in their 30s.
If you said you would give everyone a gold bar they would vote for you. |