Options
politicalbetting.com » Blog Archive » David Herdson wonders how much we can trust the referendum
politicalbetting.com » Blog Archive » David Herdson wonders how much we can trust the referendum polls
Knowledge, information and judgement:
0
This discussion has been closed.
Comments
Well people can change their minds, so I think it would be a bit unfair to compare a poll two weeks before the referendum to the results themselves.
The other thing that encourages me to believe the picture shown in the referendum polls will stick, and likewise to a lesser extent the UK 2015 polls, is that they've shown a very stable pattern, which we'd expect if people have mostly formed strong opinions that won't easily change. AV was trickier because most of the voters had very little idea what they were voting on, which makes them subject to late switches when they finally tune in.
I take the point that the pollsters don't get to practice referendum polls, but nah, I don't think yes is value.
http://www.telegraph.co.uk/finance/economics/11049834/Scottish-independence-may-trigger-immediate-capital-flight-warns-HSBC-chairman.html
http://www.telegraph.co.uk/finance/economics/11049180/Scottish-business-leaders-prepare-to-unite-against-independence.html
When there is a late swing, it has to be in a vote which people are paying attention to at quite some level, and there has to be some prompt for it. I don't see the 2011 AV referendum having either of those, so would put the disparity between the polls and the result down to error rather than late swing. (It also has to be physically possible: in that vote, turnout was only 42%, which means postal votes will have played a very significant part and most of these will have been cast before polls taken in the final week).
There are three factors for error in using polls as raw predictors: that the methodology's wrong, that the sample's unrepresentative (the MoE), and that people change their minds. My point is principally about the first one, which isn't to write off the other two but I had to keep the length of the piece reasonable.
Polls frequently are good, but that's precisely why I think we may be giving too much credence to them just now: they're good because they've had decades to hone their skills and methodologies and to eliminate errors. Even so, when something new comes along, that knocks askew their balancing calculations. Another example I didn't give was this May's Euro-election, where the last poll of Opinium, ComRes, TNS and Survation all put UKIP more than 3% above what they actually scored. Again, there might have been a late swing away, though as YouGov produced different figures at the same time which were borne out at the election (except for the Lib Dems, which they had too high), so probably not.
UKIP was a new factor in an election at that level of support and the pollsters were still feeling their way as to how to handle that situation. Likewise with Scotland. The simple fact that the polls have such a wide spread at the moment means that by definition, some must be wrong; the problem is we don't know which. However, history does show that there tends to be something of a 'group think' mentality among pollsters and when some do get it wrong, those that do usually all get it wrong in the same direction. It may be that they are *all* wrong in the same direction: I wouldn't rule out 62-38 or 50-50 as possible 'real' positions. Hence my belief that Yes represents value.
1. The polls are still showing too high a turnout. If you believed the responses of those who answer, turnout will be much nearer 90% than 80%. Now, it will be actually higher than officially stated due to deaths and multiple-registration of students and the like, but that only has so much impact. The unweighted results in YouGov's poll from 12-15 August, for example, found 85% saying they were 10/10 likely to vote and a further 7.3% giving themselves an 8/10 or 9/10 rating. Even after weighting, the 8+ score was 89.5%.
2. And those results may well be accurate for YouGov's panel, which by definition has a bias to those willing to get involved. That's one of my concerns: that in such a high-turnout election, pollsters may not be finding enough people who are usually totally disengaged by politics but will turn out this time, and that's a group which may behave differently from the rest, if only because they are new to the process. For the telephone pollsters, that's less of a direct issue but there may still be indirect effects which produce the same results (a politically disengaged person may easily be socially disengaged in other ways).
3. The pollsters have had lots of practice with 40% elections and 60% elections, have tuned their machines to deal with them and the best are very good. I would not be at all surprised if despite their conscious attempts to model an 80%-turnout referendum, the default position of the companies is to return to the comfort zone of their GE modelling in areas of uncertainty, of which there'll be many.
Betting value is when you and Edmund agree on something most of the rest of us hadn't thought of!
Even the standard polls show an increasing percentage of Labour 2011 voters (the hard core) voting Yes - the last figure I saw was 37% and it has been going up every time it is polled.
This "huddling" might reduce the commercial risk for the pollsters but it does not necessarily make them more accurate as the consistent errors with UKIP and Boris showed.
That all said there has not been a non leading question poll in favour of Yes throughout the campaign. A Yes vote now would be the worst disaster in the history of polling. Not as great a disaster as it would be for Scotland of course.
With regards to the Scottish referendum, the "Yes" side will probably be more inclined to vote, while the "No" voters will be more apathetic about it.
This is mainly based on conversations on the subject, and comes with the usual caveats.
I think that's right. But as EiT points out , number 2 and probably number 3 are not the case for the Indyref (and probably not for the General Election), because the same general picture is shown by one poll after another. One poll can easily have a biased sample. The probability of numerous polls nearly all having the same biased sample is negliglble. If the methodology is correct, then there is a stable though slightly reduced majority for No and Labour has a stable though slightly reduced lead and a very stable voting share (37-38).
So is the methodology wrong? To base this on more than vague unease (or vague hope), we need to have some plausible ideas on how it might be wrong. Like EiT, I think it's unlikely for a high-turnout referendum - if pretty much everyone on the register except for the dead and the terminally indifferent is going to vote, a major uncertainty is entirely removed. For the GE, I do have my doubts on weighting, which may be understating UKIP and overstating the LibDems because pollsters (especially ICM and Populus) assume a drift back to previous voting, which may not happen for either party.
Late shifts are always possible, of course. The second Indyref debate is an uncertainty, and there are still several events between now and May which can shake up the GE. But in general, if a professional pollster tells you something 20 or 50 times, the default is to believe him.
Suggest people read the comments.
The main argument that was getting through was how remote London is to where you live. Many of the places we visited were several hours from Glasgow and Edinburgh let alone anywhere further south, and the emotional pull was that the south of England already feels like a foreign country to you, so it might as well be.
The weakness of the YES Scotland material was that it assumed that as soon as there was a Government in Edinburgh, it would stop making decisions you disagreed with. Magic money tree stuff from start to finish (carefully written to avoid making any pledges about the things you don't like about London) and a reminder of how hated the Tories are hated and this is your one chance to remove them from Scotland on every leaflet.
Interestingly enough I took a photo of a notice about currency in the display of a YES Scotland shop which not only had a Plan B but also a Plan C for currency - not sure this was on message! For info it was
Plan A - currency union
Plan B - sterlingization as per Isle of Man, Jersey
Plan C - own currency
My worry was that the greater passion was on the YES side. If the turnout really is into the 80%s then this will only help the NO side based on what I've seen. If the turnout doesn't reach these levels (and I've seen many people commit to voting who are too busy washing their hair and watching TV on polling day) then the union may well be in trouble.
One way or another this is not going to be a happy ship.
I should be a definite "No" vote given my beliefs, but even I am finding it hard to muster up much enthusiasm.
Populus is more in line with Yougov etc, whereas ICM is currently a long way low on the UKIP score -
My personal view is that the final ICM will be closer on the Lab - Con battle where the methodological assumptions probably hold true than it will be on the LD-UKIP vote share %s.
Nowadays, he is what ever he thinks is going to win, which, unfortunately is now a major problem for him and the SNP. Nobody really trusts him now.
If only the SNP had put forward a positive case - not carried on and on about being rid of Westminster and the Tories. (There is a Panda joke, but I'm pretty sure you'll have heard it before......) People heard 'the English'......
On topic - excellent article Mr Herdson - tho I do not necessarily agree with SO that the YESNPers certainty is a function of the likely outcome, more the result of years of talking among themselves (when not abusing others, in some cases).......Comical James goes Pop being a prime example.....
Yes, you are right, more and more people are living beyond their means
"Food poverty: Experts issue malnutrition health warning"
http://www.bbc.co.uk/news/health-28883892
Oddly enough, "gout" the "disease of kings" is also on the rise.
Join the dots.
The Speccie reads the runes on the streets of Edinburgh:
Only in Scotland do you realise how well Cameron’s indifference to the campaign has played out. The SNP had hoped all along for a clumsy intervention from No. 10 that would restore Salmond’s position as a glamorous outlaw opposed to London’s evil schemers. By keeping quiet, Cameron has turned Salmond into the very thing he once opposed. The First Minister represents authority. He is now a vested interest. He has become the self-serving power base that seeks nothing but its own survival.
http://www.spectator.co.uk/arts/arts-feature/9292112/salmond-has-already-lost-if-the-edinburgh-festival-is-anything-to-go-by/
A 3 year campaign, a country split in to two tribes and simmering grievance all around.
Bodes well for the great nation builder.
"The fact is we're living way beyond our means and voters in the UK seem to simply not get this."
This is so true and could well get worse. A parent (GB) gives children sweeties to bribe them to like him, so when the other parent stops the sweeties, there are howls of anguish - even though both parents have insufficient income, are nigh bankrupt and expenditure has to be cut severely - now.
Of course the first parent (GB) and his friendly advisors (EdM, EdB etc) say it is not our fault and blame the second parent for unnecessary parsimony, whilst at the same time they are responsible for having brought too many new foster-children into the family, which has increased the family costs.
BTW, has EdB stopped his flat-lining gestures at PMQs?
You looking forward to a right wing revival then Malcolm?
A new blue tartan Scotland?
Be careful of the "gout", to much rich food and alchohol brings it on you know?
It has been building up for many years so another few made little difference, other than to highlight further the direction the UK is taking is the opposite to what Scotland wants.
Odin manage it, which was why the warriors tended to favour Thor, Odin was way to sneaky and clever.
http://www.adamsmith.org/blog/money-banking/an-independent-scotland-should-use-the-pound-without-permission-from-ruk/
Panama has VERY sound banks - it has to, as there is no lender of last resort. The only tiny wee problem is that Scotland is a magic money tree lefty jam tomorrow entitlement spendy utopia. So in all likelihood the very catastrophe Scottish business is lining up to warn of would indeed come to pass.
Doesn't Panama also have a small interest in merchant shipping to help their economy along?
Doesn't bode well for the Outers in a future referendum on British EU membership too, given Out's abject failure to win business to the cause.
On the other hand, those champagne socialists..............
http://www.telegraph.co.uk/news/worldnews/middleeast/iraq/11049953/Net-closes-on-Jihadi-John-as-London-pair-probed.html
Professor attacks SNP on health record and accuses Yes of spreading fear about NHS
Professor Alan Rodger, who helped transform the Beatson West of Scotland Cancer Centre into one of the best facilities of its kind, also launched a scathing attack on the SNP's health record.
The former oncologist said Nicola Sturgeon and her successor as health secretary Alex Neil were more interested in hiding bad news and attacking critics than improving the service, and said the Yes camp had hypocritically adopted the scaremongering tactics it has repeatedly accused the No side of using.
http://www.heraldscotland.com/news/health/professor-attacks-snp-on-health-record-and-accuses-yes-of-spreading-fear-about-nhs.25113817
Champagne causes rickets, TB, scarlet fever and malnutrition?
Gosh, you learn something new each day.
http://www.telegraph.co.uk/news/worldnews/middleeast/syria/11049663/British-woman-vows-to-become-first-female-to-behead-western-prisoner-in-Syria.html
I imagine that they could balance the budget of all Scotland's money just as well as they balance the pocket money budget, ie perfectly. The mammon boys who love to borrow are 500 miles south, busy filling their own pockets for the rainy days ahead as well. Better Together right enough.
Perhaps Scotland could become a "flag of convenience" or maybe a tax haven for the worlds rich?
Have you seen the sporting estate subsidies? A good investment opportunity for those with spare cash.
It would be an astonishing failure if all the polls were wrong.
Do the pollsters weight geographically? It may well be that No polls better in Edinburgh and the borders, and Yes in the North East. Indeed it is likely that the geographic variation causes a certain discontent with the result afterwards, If one region was dragged into an outcome it strongly voted against.
F1: bloody weird. Chilton now *is* driving P2 and the rest of the weekend for Marussia, not Rossi, who is only driving P1 after being told (and us all being told) he'd drive for the whole weekend.
You mean like IDS Malcolm? ;-)
(I wonder how his "reset" is coming along?)
Con: 301
Lab: 293
LD: 28
http://www.bbc.com/news/uk-scotland-glasgow-west-28880994
"I believe that Panama does not have unemployment benefit (JSA), working tax credits or child benefits. "
Sounds like your idea of Utopia?
http://web.worldbank.org/WBSITE/EXTERNAL/TOPICS/EXTPOVERTY/EXTPA/0,,contentMDK:20207841~menuPK:443285~pagePK:148956~piPK:216618~theSitePK:430367,00.html
Con .......... 301 (-2 seats)
Lab .......... 293 (+1 seat)
LD ..............28 (+2 seats)
Others .........28 (-1 seat)
Total ......... 650
Surprising imho, given the Tories' poor showing this week with the YouGov/Sun polls.
"The disparities in the distribution of assets and economic opportunity reflect Panama's uniquely dualistic pattern of development. Panama's privileged geographic location and its monetary regime anchored in the use of the US dollar as legal tender have fostered its comparative advantage in services, which contribute over three quarters of GDP and generate two-thirds of employment in Panama. These strategic factors have also spurred the rapid development of internationally-oriented, modern, dynamic service enclaves, including the Canal Zone, the Colón Free Zone, and the International Banking Center. While these enclaves generate large shares of GDP, they create little employment (3 percent of the labor force) or fiscal revenue. Moreover, they inject negative spillovers into the economy due to the huge differentials between wages paid in the enclaves, particularly the Canal Zone (which is subject to the U.S. labor code) and the rest of the economy.
Indeed, factor markets have been segmented by policies that drive up the cost of labor relative to capital. Panama's labor market is characterized by a multiplicity of policy regimes, with separate regimes for the private sector, the public sector, the Panama Canal Commission, and the Export Processing Zones. These regimes have created large wage differentials between workers in the Panama Canal Commission, public sector employees, and those employed in the rest of the economy.
Labor-market interventions not only hamper growth, but also have a direct link to poverty. By increasing the relative price of labor — thus, reducing demand for the poor's most abundant asset — these distortions have swelled the ranks of the unemployed and encouraged informality. Moreover, while they benefit those who work in formal sector jobs, the resulting segmentation of the labor market can put a heavy toll on informal sector workers by reducing their wages, making it difficult for the working poor to grow out of poverty through their own labor. Indeed, the LSMS reveals that the poor in Panama do not benefit from such distortions, but may be hurt by them: (i) the majority of the working poor receive wages that are below the official minimum wage; (ii) they do not receive the "mandated" fringe benefits; (iii) the majority are employed in the informal sector, where wages are lower and employment terms less favorable; and (iv) the urban poor are hurt by high rates of open unemployment and the rural working poor appear to be underemployed."
Further, I would argue that if the polls one week before an election are more accurate than those two weeks before an election then the simplest interpretation is that there has been a change of opinion during the latter period of the campaign. This is far from being the sort of hand-waving late swing that pollsters have appealed to in the past when their final polls prove to be inaccurate.
You noticed that?
Bit like London and the South East really?
The giant throbbing pulsating 5000 pound gorilla in the corner of all politics in the developed democracies is that we can't afford our welfare states. And something that can't go on won't. The choice we face is how to make it affordable. And to be as fair as we can while we all retrench. We need to say 'this is how much we can spend because this is how much we can take in tax and not shrink the economy - and that's all we can spend. Now how should we spend it?'. Anything else is dishonest jam tomorrowism.
How were corporation tax receipts this year, compared to the increases in VAT and National insurance receipts?
There is no reason why going forward they cannot run the economy such that borrowing is kept to what is necessary over a period to balance annual fluctuations.
If the troughers were as good at collecting taxes from their chums as they are at persecuting the poor then we would not have a structural deficit. Plenty of money in the UK it is just where it is going that is the issue.
http://youtu.be/-dRuPPSKNhE