Completing the set
Posted on
September 03, 2013 by
Rev. Stuart Campbell
And just to finish off our in-depth study into respective coverage of the recent YouGov and Panelbase polls in the Scottish media, here’s the Scottish Daily Express.
Reporting of the YouGov poll (giving the No camp a 30% lead) is at the top, and the Express’ coverage of the Panelbase poll (putting Yes narrowly in front) is below.
The second article is much more interesting and accurate Stu!
That second piece looks exactly the same as on line The Scotsman’s login page.
Good of the Express to provide a thoughtful and insightful analysis of the two polls, before contrasting them and also putting them in the context of the debate without stooping to any overexcited language that furthers an agenda.
Top work. I expect to see more.
Reading nothing is still far more accurate than anything you will find in the vast majority of any of the tabloids or online news sources.
Operation Fear is turning into Operation Panic.
Can we not have a picture of Kerry Gill in that empty box with his fingers in his ears shouting ‘la-la-la I’m not listening’?
The Express is a spoof newspaper, it makes the Daily Mail look measured and composed. A lot of people think the Express is staffed by actual people, it isn’t though, it consists of an old 286 PC – its only task is to generate headlines after words from a limited list have been uploaded via 3.6inch floppy disk consisting of:
Diana
Immigrants
Diana
Cancer
Britain
Diana
Separation
Diana
Immigrants
The EU
Diana
Plaid trousers (Teflon coated to prevent staining.) Special offer.
Diana
I <heart> Yougov.
As noted previously, the Yougov poll is actually ~37% Yes / ~47% No. On the higher side for No within variance most likely, but still 5% swing territory. This can be easily calculated by re-weighting to Holyrood (as opposed to the flawed UKGE 2010 weighting). By doing so, you then match all the other online polls (ICM, Angus Reid, Comres, panelbase).
I plan to go back and do it all the way to 2007 when yougov’s method stopped working due to the rise of the SNP / fall of Labour in Scotland (yes, this began back in 2003 – 2011 was simply the culmination of an 8 year process). I suspect I shall get a perfect match with all the other polls for a referendum tomorrow (the recent panelbase was for 2014 in contrast).
In the meantime, if Yougov can keep convincing the No campaign they’re going to win by a mile, whereby the same tactics are continued, them I’m a happy man.
Rev : Cant see the second story from express at my end – has it gone ?
Thought the Scotlandshire debate last night came across pretty good – not much support for union in evidence from the audience and J Baillie up to her usual guff, but kept in check by a very able Jamie Hepburn.
On Scotland Tonight ; John Curtice couldn’t wait to cast doubt & aspertions via his leading question scenario (?) of Panelbase Poll. He was champing at the bit. He has shown is hand on this matter no credability left at all surely.
Does Scotland posess no other pollsters ??? (Scotlandshire & STV keep using this guy why is he the only one)
LOL. I was looking at the second picture for about ten seconds waiting for it to load before I twigged.
I think the article on the Panelbase poll is probably one of the most honest and balanced articles I have ever read in the Unionist press.
I just can’t put my finger on why I should think that.
The Emperors new clothes look absolutely spiffing.
Sorry to go O/T here Stu and apologies if someone else has already mentioned it but the BBC Scotland site is allowing comments at the moment.
link to bbc.co.uk
Enjoy! 😆
When the polls look bad for YES, the MSM report such.
When the opposite occurs they either ignore or smear the professional polling companies. They don’t like it up ’em!
oh dear, I’m a bit slow today – took me far longer than I care to admit to get this!
Yep, journalistic cowardice. They simply haven’t the intestinal fortitude to print reports on back to back contradictory polls. Most especially when one slaps their own narrative in the kisser.
Fearties.
Pa_broon74
“HAS THE HOUSE PRICE CRASH MADE PROPERTY PRICES OBESE?” etc
link to qwghlm.co.uk
Always worth a repost
BBC comments open on the Salmon story today.
Had my say! Much sense from the YES much bile and hatred from the NO’s, they equate smearing AS with rational debate!
Folk watching the comments roll over will surely raise an eyebrow and wonder.
OK I missed a ‘d’ cant edit !
Ian’s (latest) Smartarsery neatly put to bed by the eloquent BTL comments
link to ianssmart.blogspot.co.uk
O/T
I just noticed a ‘Hen Broon’ posting on the BBC article comments section which is a bit flame’ish. Now I understand the dubiety which was cast against the nome de plume when I first posted here.
That particular ‘Hen’ is not me although I did once post a reply on an other BBC Vic Galloway article asking what had happened to a particular Scottish act as ‘Henn’ informing the readers of what each of the members were doing now. That was a few years ago though.
As a matter of interest, if I should decide to change screen name is ‘Henn Broon’ locked to the email addy I’ve supplied for this comment section or is it a simple case of being able to just change it?
I’ve just realised.
The way yougov weighting works is that the higher the support for the SNP (at Holyrood) is, the higher the No and lower the Yes share will be due to the flawed weighting method.
There’s almost a perfect correlation going back to pre-2007 when the problem first emerged. That’s bloody hilarious.
People, pray that the Next Yougov poll shows N even higher! That will mean the SNP are on a roll.
Basically, as support for the SNP has risen, so the flaw has been accentuated.
I wonder how much money has been wasted by the pro-union campaign for Y/N polls by Yougov. LOL.
Man that’s so funny.
@scottish skier
GREAT analysis; Pure respect from me. LOL too.
Not quite the full set (admittedly this is from the Uncle Tam section of the London press):
link to blogs.spectator.co.uk
Scottish Skier
interesting stuff – an article from yourself with a fully detailed breakdown and methodology would be very interesting!
scottish skier: “i’ve just realised”. Is it possible for you to give a fuller description of what you mean about the weighting, for non-statisticians? It would be really useful to know.
scottish_skier says: 3 September, 2013 at 1:09 pm
I’ve just realised.
The way yougov weighting works is that the higher the support for the SNP (at Holyrood) is, the higher the No and lower the Yes share will be due to the flawed weighting method.
So how do you correct for that, and do you think other pollsters are making the same weighting mistake?
Rev, you got a response from the British media eventually.
As I predicted on these pages a few weeks ago, there would be silence from the Establishment for a short period while they reflected upon the results of your Panelbase survey.
The Unionist media has obviously colluded with each other & thus has decided to publish the results of their own surveys as a counteroffensive measure. This was expected because they are unable to pick holes in the results of your own survey.
The degree to which they have distorted & fabricated the results of their own commissioned surveys suggests to me that they are deeply unsettled. We can only expect more lies, distortions & comic style scaremongering rubbish until at least September next year.
And the only reasonable conclusion that can be arrived at is that there is an emerging groundswell of support for independence that the British media are now struggling to hide. I predict their tactics to include more time focusing on the messenger (i.e. you & we, your readers), when we can expect to be labelled with all manner of dreadful names in a puerile attempt to demoralise & undermine the movement.
Opinion polls are effective WMDs (weapons of mass distraction) that will be used increasingly by both sides in the final year to the referendum. What BT-MSM did not expect was a crowd-funded WMD to be launched whilst they were away on their holidays. You have really put a spanner in the works, Rev! I initially thought that another crowd-funded poll was a bit too soon to launch. However, I now realize that, with all the furore caused by the first one, a crowd-funded poll launched at regular intervals during the next year could seriously unhinge the BT-MSM campaign.
Perhaps I have misunderstood something about the nature of “bookies’ odds”, but it seems to me that there is somting strange about this page:
link to oddschecker.com
All the companies quoted show NO to be the clear favourite, but the diagram illustrating the most popular bet shows that the money is going on YES. In a two-horse race, should this not mean that the bookies, nomatter what their opinions, would make YES the favourite?
Arbroath 1320 says:
“Sorry to go O/T here Stu and apologies if someone else has already mentioned it but the BBC Scotland site is allowing comments at the moment.
link to bbc.co.uk”
There’s obviously a unionist team working on the comments trying to “out-rate” the pro independence comments. I am convinced many who post in favour of the union are paid activists.
X_Sticks says:
Many who post in favour of the union on the BBC are uninformed and are probably not eligible to vote in the first place. I am heartened by the pro yes commentators, who in the main, stick to the facts or base their argument on equity and fairness.
Smearing a person (AS) or (SNP) is the hallmark of someone who has not a lot to say and will lead to the inquisitive don’t knows opting for the yes side when reading.
I am amazed that the comments section has been open so long!
Your doing good missionary work X-sticks.
@:0}
@ X-Sticks –
You’re not far wrong matey, comments allowed on the alex Salmond piece but no comments allowed on the piece about Osbourne claiming our taxes will rise if we have an oil fund similar to Norway. The EBC has lost all sense of fair play. No shock there though.
I have posted three comments re Salmond topic on the BBC website, the likes go up to 11 and higher, the next time I look they are in the minus figures…..me thinks a Britnat plot is afoot!!
O/T and on an earlier post re the Better Together shop/office opening up on Union Stree in Aberdeen
Best of luck to them handing out Better Together/Union Flags on a Thursday, Friday and Saturday night up that part of town….would love to film that one haha carnage 🙂
Must remember that when the BBC open comments on a Scottish article it is solely for the benefit of all deranged unionists in the UK to pour bile into it, this is to ‘inflate’ the negativity against predominately Scottish-based pro-Indy/SNP support. You will see that this article and comments open has been advertised on the main news web page as to alert all and sundry to the pro-union cause ie Tory, BNP, UKIP, NF, EDL, DUP and Labour activists waiting for the call = One Nation Better Together. No thanks!
@AberdeenLoon…
I have posted three comments re Salmond topic on the BBC website, the likes go up to 11 and higher, the next time I look they are in the minus figures…..me thinks a Britnat plot is afoot!
That’s happened to me recently too. A fair comment getting the plusses only for it to be in the minuses after about three hours or so.
I had a comment deleted because I was called an idiot by some Britnat, I proved his comment wrong and suggested that he was the idiot and got deleted 🙁
So how do you correct for that, and do you think other pollsters are making the same weighting mistake?
Other pollsters don’t make this mistake. None weight by Westminster past vote for Scotland and that’s where the flaw lies.
The flaw in yougov stands out like a sore thumb in the tables. I may not be able to fully correct for it because Yougov are secretive and only provide the bare minimum of detail. In this case, the client has also been freaked by the final voting intention numbers (excellent for the SNP I’d predict from other polls) so has asked for these to be blanked out and this is a further hindrance.
What I can say with a huge degree of confidence is that Yougov is bollocks, has been since 2007 (up to 10-15% too high for No and 5-10% too low for Yes) and a look at weighted and unweighed ‘Westminster Vote 2010’ figures should show people why (or at least where the problem starts before being exacerbated). A check with unweighted Scottish constituency vote supports how bad the former is in terms of invalidating the poll completely.
In the meantime, a prize for someone who can spot what is glaringly wrong.
Tables: http://tinyurl.com/qyualck
Tip. Those of you who support the SNP but have been voting Labour tactically for GEs really need to get your story straight…
Hi all have been reading wings along with other excellent sources of information that is rarely available in MSM. Am really excited about way things are progressing and concur that Scottish skier should write article with relevant links so a proper analysis of the damn lies and statistics is public knowledge. There are numerous experts in this field and those who work for organisations who depend on their reputation for future business.
The shenanigans of our ( so called) press is bought already although ironically less are buying their nonsense. People that live here deserve the facts and are not getting them directly.
Nobody likes being lied to that is the fact that will be their downfall.
Thanks to all who contribute and keep me slightly saner
scottish_skier says:“The way yougov weighting works is that the higher the support for the SNP (at Holyrood) is, the higher the No and lower the Yes share will be due to the flawed weighting method.”
Are you able to delineate this in a way that would be accessible to the functionally innumerate among us (which may be to say, most of us)?
@scottishskier
If you’ve not seen it the Alex Massie take on things, sure you’ll find it amusing!
link to blogs.spectator.co.uk
“If you’ve not seen it the Alex Massie take on things, sure you’ll find it amusing!”
I had, and it’s so disappointingly stupid I’d already taken him off the links column as a result. He’ll be desolate, I’m sure.
@Rev
It was for @ss, just thought he might find it entertaining.
Unsurprising, however on a positive note does anyone else think we might have the basis for a new Scottish Independence Top Trumps?
Henn Broon says:
3 September, 2013 at 1:06 pm
“O/T
I just noticed a ‘Hen Broon’ posting on the BBC article comments section which is a bit flame’ish. Now I understand the dubiety which was cast against the nome de plume when I first posted here.”
What dubiety was that then Mr Troll? I can find no record of you or it. LMFAO
So let’s see about that Yougov.
Yougov use a panel so they don’t have to randomly phone people and hope they meet the demographic, phoning again and again until they get a decent sample. Instead, they know a bit about people on the panel so can e-mail target the demographic they’re after.
You can see this works well as before any weighting, they get this for Holyrood voting intention:
SNP 42% Lab 35% Tory 15% Lib 8%
That’s not far off current (conservative for the SNP) figures of
SNP 45% Lab 32% Tory 13% Lib 6%
Wahey. So unlikely that much weighing will be required! That’s great, as the less weighting the better.
So, how well does our target sample match up in terms of the last General election.
17% Say they voted Tory. Bang on!
22% Say they voted Lib Dem. That’s not far off 19% at all!
So, looks like without any pp weighting, our sample is already looking like an excellent demographic section matching past GE and current Holyrood intention.
Right, let’s move to the most important parties.
25% Say they voted SNP – that’s ~5-6% too high
36% Say they voted Labour – that’s ~5-6% too low
Hey, wait a minute. How can we match everything but these two?
Well, because people are telling porky pies or are so torn between the SNP and Labour they’re recalling wrong.
Yep, some people wanted to vote SNP but they voted labour. A lot of them. 12% of the voting electorate in fact. About half of these are admitting they (tactically) voted Labour, the other half, for whatever reason, say they voted SNP when they didn’t! It may be they voted Labour then decided to shift to SNP and that’s what they are recalling. That’s why you should use the most recent election for this; and that’s Holyrood 2011!
Anyway, clever guys at Yougov spot this and…
No they don’t. What they do is penalise people IDing as UKGE SNP voters and cut them down by 20% (5-6% of the electorate).
At the same time, they boost Labour ID’ers by 20% (5-6% of the electorate).
They then proceed with the Y/N and VI poll with a completely flawed base. The net result is in Y/N, Yes is hammered and N is given a huge boost. Also, if we could see Westminster VI from Yougov, we’d probably see the SNP somewhere close to Labour but in fact they’re considerably further out in front.
If the issue was with small parties, it would make little difference, but with the two which are by far the biggest, encompassing nearly 3/4 of the vote, a 20% weight change (5-6% of the electorate added to one party with it’s rival docked 5-6% creates a 10-12% difference) is massive; a 10% gap can become a 22% one and so on.
As per my earlier post, the more the floating Labour/SNP issue grew, so it appears did the discrepancy between Yougov and all the others. Hence my joke about a higher No meaning the SNP doing well 😉
So ignore yougov. Oh, and if you are polled by them, tell the truth! It’s SNP voters causing a bloody problem too.
Anyway, as per my earlier post, this is why Yougov have gone off like the lonely BT lady since 2007. That’s when the SNP/Labour battle really began. The SNP won in Holyrood, Labour held on for GE’s. They’ve lost the latter now.
Yougov will keep polling rubbish until they decide to pp weight by Holyrood.
Do this in a rough way for them, and you get something like panelbase. Conservatively 37Y/47 N or possibly 41Y / 43N…
scottish_skier says:
Aye right! Do you do on-line learning?
Statistical analysis was taught on that day I was off college doing a homer.
Slide rules were just being overtaken by calculators.
Mobiles were what you put above the bairns cot.
I’m glad your there to explain it all so well.
@:0}
Surely a ‘proper’ poll should contact folk by chance?
Phone 33.33% folk from the phone book at random.
Stop 33.33% folk in the street.
Email 33.33% folk at random.
Any poll only using only folk that want to be polled is surely suspect?
What about the 99% of folk in Scotland that will vote, but slam the phone down on cold callers, ignore spam emails, and walk past folk that try to stop them in the street?
I think I trust the yougov one on this would be nice to see a swing but I reckon last few weeks in the press haven’t been kind to us
I think I trust the yougov
I want to seem more of them. They serve an important purpose.
Surely a ‘proper’ poll should contact folk by chance?
Yes and no. All methods have their pro’s and cons.
Telephone polling by random digit dialling (MORI and TNS interview organising) has proven increasingly inaccurate due to people no longer using landlines. 20% of homes don’t have a landline and a signficant % don’t use the one they have other than for broadband. ICM are trying a landline-mobile approach although it’s early stage.
The rest are all going with online and getting good results. The number of people without an internet enabled device is reducing rapidly.
Yougov just weight stupidly as they have a quick way of doing things and don’t understand that it doesn’t work in Scotland because we have two parliaments with a strange swing vote between them; i.e. people voting for two parties which are in opposition to one another at the same time.
I’m trying my best to understand polling methods SS, but I can’t get my head around it.
It seems to me that polling companies have invented a business to satisfy their customers, make money, and have set their own rules.
If I live in an area with 1000 Yes voters, but only one No voter in the area registers with the polling company, then I live in a No voting area if they contact him/her?
Maybe i’m just thick, but I can’t understand the method.
Well I think yougov one has a better track record of being more reliable than one that is paying people to do it and commissioned by YES. It’s going to inherently ignore people in work, who dare I say tend to vote no (probably as they don’t want to change the status quo) so it’s nice to see but we need to do more!
“Well I think yougov one has a better track record of being more reliable than one that is paying people to do it and commissioned by YES. “
No it doesn’t. And YouGov pays people just the same as Panelbase does, and was commissioned by a No campaign group.
Pollsters simply try to gauge public opinion at a given instant. They have a hard job of it. They make mistakes. People lie to them. People tell them they plan to vote for X but then they vote for Y. People say they’ll vote then they don’t…
Panel polling works well. Far better than landline telephone now. I can’t think of a reason why people who support independence would be less likely to be registered with a panel. I’m registered with Yougov and have been polled by them. I’ve also had both MORI and TNS RDD phone me.
They collect different amounts of data in different ways. There are benefits and problems. They then analyse these in different ways and try to produce an answer. Again, the way this is done can affect the conclusion. Yougov make what seems like a possibly minor error which results in a huge one. MORI hit a more No demographic with landlines and have ‘Shy’ factor problems. TNS too. ICM, AR, panelbase online all show more sensible, less reactive results due to the nature of media the poll was conducted under.
I’d not pay a lot of attention to polls right now. They can’t predict what will happen, only give indications as to what might and that often requires delving deep and asking other questions rather than the obvious.
The class example of polls ‘getting it wrong’ was the 2011 election. They didn’t get it wrong at all. People were saying they planned to go out and vote Labour. However, they then went out and voted SNP. If the polls had delved deeper, they would have seen it. If Prof C was worth his salt he’d have seen it too. After all, the SNP were on up to 45% in 2009 and did nothing to deserve going to 30% in 2010-early 2011. All that happened was a proportion of Scots returned to type in a panic at the Tories coming back. The proportion that’s causing Yougov to report crap. They then saw sense and voted for the party they genuinely had come to respect in 2011 and for a referendum to escape the Tories. It took guts, but they went out and did it. These people will carry Yes.
They’re nervous right now and got cold feet after the elation of 2011 when yes went ahead. They’re now coming back to Yes and once they decide, that’ll be that.
I guess your right it’s probably not worth looking at polls, if we did that we’d have given up a while ago! Just try and keep getting the message out there.
I guess your right it’s probably not worth looking at polls
I mean they’re not worth taking literally and straight Y/N right now are really not a good indicator of what will happen – the 2011 huge swing being an example of why not to get too excited/depressed well in advance. All we can do is look at them all, seek long term trends, try to find common features.
Yougov actually match panelbase really well in terms of raw numbers (they see the same levels of support for indy in terms of party ID, the same levels of support for parties in holyrood etc). It’s just what yougov did with what they had that causes the discrepancy.
You can get excited or depressed as appropriate about Y/N within the month or two ahead of voting day. They they will start to converge towards the final result.
Except Yougov, it will be showing 90% No while the rest have Yes storming to victory…
😉
I don’t understand what yougov does wrong really are you saying they use general elections to weight them which I guess would be seen as the best indicator as they’re seen by the majority of people as the elections that count most hence a trunout of 60% to 50% and shows real voting intentions as opposed holyrood.
pa broon says
“Plaid trousers (Teflon coated to prevent staining.) Special offer.”
Really? did you see how much?
arbroath 1320 says
“BBC Scotland site is allowing comments at the moment.”
Thanks for the heads up Arbroath, just been on there and gave them a piece of my mind, boy it felt good, and I didn’t give the Bastards the opportunity to moderate me, so my posts still up, I see there are plenty though that have been removed by the moderators, and the level of knuckle dragging debate from the almost 100% English comments are the ones, but you can bet your arse the BBC will turn it into a reason to close the comments again and call it cybernat abuse, but I’ll tell you the reasoned ones are almost 100% yes campaigners
HenBroon says:
What dubiety was that then Mr Troll? I can find no record of you or it. LMFAO.
link to wingsoverscotland.com
Well the last time I posted here with this Screen Name I was called a Troll and it seems I have been again. Incidentally I performed under the stage name of Hen Broon in the 80’s did you too or did the name just suit yer fancy? It doesn’t change the fact that the Hen I mentioned flamed the thread over there on the BBC site or that this Henn is a long time independence supporter provided you get yer jollies son.
I don’t mind changing my screen name and is really what I was making inquiries about in my last post here in the first place. Not wishing to appear a clone or flamer isn’t exactly the modus operandi of a Troll now is it?
I don’t understand what yougov does wrong really are you saying they use general elections to weight them
Yes. Problem is people don’t tell them the truth about what they voted for. Tories do, Libs do, but there an up to 12% SNP+Lab (voted for both) portion who don’t. The main issue is people who voted Labour saying they voted SNP. These are people voting SNP at Holyrood, tempted to vote SNP for Westminster, but then vote Labour tactically. However, when asked, a lot say they voted ‘SNP’. It may be that they are recalling wrongly or feel shame that they did not have the courage to vote SNP and reverted to their old ways. It may be they thought they were being asked what they voted in the last Scottish General Election…
Whatever it is, have a look at the tables and you can see it. The targeted unweighted sample almost perfectly matches Holyrood VI for all parties and past V for 2010 for the Tories and Libs. This should only require minor weighting to create the weighted sample. However, past UK SNP vote is far too high and Labour far too low (total up to 12% of the electorate wrong response)… Yougov just correct by weighting and there comes the f-up.
You can see it in independence support matched with Westminster VI. Double the number of Labour voters for Westminster support indy than for Holyrood. Eh? Nope, these are SNP supporters voting labour tactically.
It’s not specifically Yougov’s fault and I understand they have noted there could be a problem. However, if you past vote weight, you should really do it for Holyrood as it is always the most recent election (thus lower voter recall mistakes) and it is PR with no tactical FPTP voting as per westminster.
I’ve never come across a case where a pollster weighted on 2 elections back rather than the most recent. Chuck in the complex, evolving politica dyamics in Scotland where people are voting for parties at UK level they don’t like and who are the bitter enemy of the party they do support and you have a recipe for disaster.
Panelbase follow global standards and weight to the most recent election; Holyrood.
Yougov are completely out on their own. Panelbase match nicely with ICM, AR, Comres etc when similar question is used. These are all online. MORI and TNS match reasonably; these involve Landline and/or face to face. They’re more biased to no as a result. Yougov matches no other poll and has sat at the same silly values for 6 years now while the rest rise and fall in response to events.
Scott, I think what SS is saying is that the raw data YouGov collect are pretty well OK, but they then add weightings to these figures which lead to their headline results being wrong. They are in effect assuming that the respondents are lying to them in some way, because the responses don’t match what they think they should be, so they try to “correct” what doesn’t need correcting.
So, if that one No voter in my area registers for a poll, and says No, the 1,000’s of Yes voters in my area are ignored, and my area is a No stronghold?
I’m still baffled.
Sorry, SS, I cross-posted.
Juteman, I really don’t think it’s anything like that at all.
I’m going through a dense moment. Ignore.
Averages for 1998 to 2009
Y = 44
N = 43
Then the referendum was going to happen and the volatility really took off.
Looks familiar though huh? Bit like the panelbase where people were finally asked what they’d do in 1 year on the real day, not tomorrow.
Yougov 2006/2007average
Y = 43
N = 42
Hmm, a great match.
Yougov 2009-10
Y = 28%
N = 57%
Just like that… Not changed within variance since.
Almost overnight N was way off the other polls by up to 15% with Yes lower by the same.
But then 2007 was the turning point that made 2011 almost inevitable. That was when Labour completely lost Scotland and 2011 became highly probable. The return of the Tories and Clegg in the Rose garden ensure it.
SS: Any chance of you putting a load of this stuff together in an article for us? It’s brilliant as comments, but if we could have it structured and laid out in a place it could be easily referred and linked to in future, it’d be invaluable whenever another iffy poll comes up, which I suspect will be an awful lot between now and this time next year.
They are in effect assuming that the respondents are lying to them in some way, because the responses don’t match what they think they should be, so they try to “correct” what doesn’t need correcting.
Actually, it should be obvious some respondents are not answering correctly. Also, anyone with any knowlege of Scottish politics should know about the odd SNP-Labour issue. People on here regularly admit to having voted Labour for Westminster in the past when they support the SNP.
However, Yougov just go ahead and weight it out not appreciating that the action is not correct things, but creating a flawed base. You should not be taking e.g. 6% off one party and adding to the opposition when everything else matches up suggesting no or little weighting should be needed. That’s a huge change that sticks out like a sore thumb and it can propagate down the process. A small mistake in the beginning can result in a big one in the end.
For a long time I’ve been perplexed by Yougov. I had a fair inkling this issue was the problem based on what they said they used for weighting. The Devo plus tables confirmed it. If you weight to a past election, you don’t need to show this – people know what the result was. This info is missing in a lot of Yougov tables. In the Devo Plus poll it was there for all to see. Yougov were GE weighting, not using the most recent election, and weighting out SNP by large margins in favour of Labour.
I’m not wahey! I never believed they were correct. Just glad I found out a solid reason why.
Any chance of you putting a load of this stuff together in an article for us?
I’ll see what I can do and no promises on timing (I spend enough time on here as it is!). Maybe before the end of the year I can do a summary of them all, good points and bad etc. Have like a day job and stuff 🙂
“Maybe before the end of the year I can do a summary of them all, good points and bad etc.”
Aye, no rush – we’ve a year yet 😀
But as you say, you spend a lot of time writing them as comments anyway, why not get them above the line where lots more folk will see them?
why not get them above the line where lots more folk will see them?
Aye, but I’ll need to make sure everything is accurate and referenced to a T. I’m a stickler for perfection when it comes to this. My normal media is like the Journal of Physical Chemistry etc. Polls a hobby.
‘Alex Salmond’s dream’ — this is the tack the BritNats are taking. This phrase has been their mantra – I know our poll is a bit iffy (however nice it is to see) but I suspect the BritNats have been aware of a change in voter mood — and this is their response; a personal attack.
SS – “I’ll see what I can do and no promises on timing (I spend enough time on here as it is!). Maybe before the end of the year I can do a summary of them all, good points and bad etc. Have like a day job and stuff”
I smell a FUNDRAISER coming on… 😛
I don’t see how that is going to help, unless he can take unpaid leave or something. Most professional jobs don’t really allow for that, because you just land your colleagues in it, because they can’t just find a trained and competent replacement sitting on the shelf.
If I find the time and go for it, a beer at some undefined point in the future would more than suffice. I’ll of course get the next round in.
😉
link to arcofprosperity.org
Seems he has a methodology and in March predicted 1st September as cross over point.
Graphs etc.