A little over a year ago I asked the following question:
Has the day come when election polls are nearly always right?
Famously, in the last US presidential election, Nate Silver correctly predicted the winner of all 50 states and the District of Columbia. His prediction for the election before that was correct for 49 out of 50 states.
Both times, I had hoped it would turn out otherwise. My hopes had been a little higher than they should have been because of the residual glow from the Shy Tory factor, first exhibited to a dramatic extent in the 1992 UK general election and still apparent, though in lesser degree, for several elections after that. I had known about that factor in my guts before that election, from listening to people on the tube, and had correctly guessed the final result would be more Conservative than the polls claimed. As the results came in I did not rejoice that the Government would be Conservative, but I did rejoice that the Chattering Classes had been confounded, their bubble burst, their conversational hegemony broken open and their flary-nostrilled noses put out of joint. Yeah.
Unfortunately not-yeah since then. I haven’t eaten a hearty post-election breakfast with schadenfreude sauce about the polls for many a year now. George Bush winning in 2004 was splendid fun, of course, but it was no great surprise to anyone who had been paying attention. The polls had given him a consistent small lead for months before the election.
Betteridge’s law of headlines strikes again. That day had not come. The polls in the General Election of 2015 were wrong, wrong, wrongety-wrong, wrongbert, wrongble and wrong.*
As was I, but least I had the nous to put in a question mark.
So, elections just got interesting again. Goody! But none of the articles I have yet seen adequately explain why the Shy Tory effect was successfully allowed for by the pollsters in the UK General Elections between 1992 and 2015, only to burst forth again now, nor why political polling in the US has generally managed to factor in Shy Republicans just fine. Except for the 2014 midterms.
The one place where the UK polling companies did fairly well this time round was Scotland, although they still underestimated the scale of the SNP’s triumph. Wishful thinking led me to suppose that the estimates being chucked around of 48 seats for the SNP were exaggerated; in the event they were too cautious. Going back to 2011, it is part of Scottish Nationalist mythology that the victory of the SNP in the Holyrood election of 2011 was completely unpredicted by the polls. However the very last polls were quite close to the actual result when it came to the constituency vote, but much less close when it came to the regional vote in the Scottish Parliament’s semi-proportional voting system. Probably the polls recorded a shift of opinion in the last few weeks of the campaign, which is all you can ask of them. When you think about it, polls cannot predict anything; the people who look at them do that. The final polls for the Scottish referendum were out by a not-bad 5% or so, in the usual direction of underestimating the small-c conservative side.
All in all, a British or American polling company attempting to sell its wares to interested political parties or news organizations on May 6th 2015 could have made a fair case that they were on top of the Shy Tory problem. So what happened on May 7th? What will happen on November 8th 2016, and will we have any idea beforehand?
*This is funny but nothing to do with this post. Americans and people under 50: don’t ask.
Sorry, but I have to call bullshit on all of that as far as the UK is concerned (and possibly elsewhere).
When the polls predict a hung parliament with Labour in the lead and the actual result is almost the opposite (a slight majority for the Tories), then the polls are bullshit.
You can describe it as the “Shy Tory” phenomena or whatever you like, but you called it one way and it went the other means that either (a) you weren’t listening or (b) you were being lied to.
Neither of these options is glorious.
OT but if you want an anti-QotD, try this one:
“Liberalism dominates British politics and is at the heart of the two major parties ”
http://www.telegraph.co.uk/news/general-election-2015/politics-blog/11594561/This-was-not-the-strange-death-of-Liberal-England-just-implosion-of-the-Lib-Dems.html
I have seldom read anything made more furious.
If you demonize a section of society they will lie to pollsters and your poll will be useless.
Which part of this is hard to understand?
On the polling thing.
There is a truism:
Never ask your partner whether he loves you just before you’re going to have sex. The likeliest thing you’re going to learn is how much he wants to have sex.
It has wider applications.
Roue le Jour,
“Which part of this is hard to understand?”
Why this effect was strong in 1992, but weak or nonexistent in 1997, 2001, 2005 and 2010, and came back in 2015. Alternatively, why this effect, if present, was successfully integrated into the sampling by polling companies in 1997, 2001, 2005 and 2010, but not in 2015.
“Why this effect was strong in 1992, …etc”
I refer you to my truism
All dialogues involve status negotiation. What you tell someone reflects what you want them to think of you. So it depends on the perceived value of the replies. Different times, different values
There are many links here, are NONE of them open in a new window, unlike other articles, making it a pain in the rear end to keep reading as I have to keep going back and reloading the damn page each time. Please be consistent Samizdata!
In view of the above, I’d guess pretty well everyone involved with polling coming from the liberal, progressive blob is going to weight the value of replies to them. Now give me an example of a surprise result for liberal progressivism & invalidate the thesis;)
Error 404 World Not Found
Try opening in new page/tab. It’s what your right click button’s for. Not to give the mouse symmetric eyes.
You’re trying to establish a correction factor where there no basis for calculation. You know conservatives are lying to you, but how can you calculate by how much?
This is not impossible. Like the Voight-Kampff test, it requires a number of questions, cross correlated. But “If there is an election tomorrow, how will you vote?” Is not going to hack it.
Election polling is a sub-“science” of political “science”, one of those social “sciences” that try to rule us all. It is a mis-application of statistics to try to determine who might win, or more importantly, how to tailor a campaign to get more people to vote for your candidate. It is a variation of commercial advertising and marketing techniques. In any case, it violates all sorts of statistical principles, starting with the fact that the objects being studied have free will, and can change their minds for reasons that cannot be modeled by the statistical method. So the error bars that they always talk about are rubbish.
There has been a strong movement in the US to do away with actual counting of people in the census, and to instead “adjust” the counted numbers based on statistical samples. It is claimed that it is more accurate and it does not undercount priveleged classifications (minorities, poor, etc), so that they are more likely to have fair representation and receive more money from the govt.
In the healthcare field, we have statistical methods being used to try to get the populace to eat less salt, to exercise more, eliminate fat from the diet, and to change their lifestyles in ways that benefit the government and health insurance companies by driving down their costs, but which absolutely cannot predict which individual people will benefit from these changes, and which ones will suffer harm. There might be an overall benefit to society from these nudges, but that benefit will only accrue to those who pay the bills and the (unidentifiable) individuals who benefit, while the (also unidentifiable) individuals who are worse off will not notice the difference. And all of us will suffer from the uncessant nagging to change our ways.
I think it is good that the failures of polling in elections like this should be discussed widely, to make sure that people do not start to believe that statistical sampling is really “truth”, and that it can be soley relied upon to make real decisions that impact people’s lives. Too often, the soft sciences are misusing statistics to accomplish social goals. Once it becomes embedded in people’s minds that this is acceptable and correct, there will be a great temptation to adjust the statistics even more, to account for desired outcomes, and we all know where that leads.
I wasn’t surprised. The implosion of the Febrile Deomrats was inevitable after their spectacular U-turn on tuition fees which showed that Clegg would do anything for love (unlike Meatloaf). God knows what else he promised Cameron. A night of mild BDSM with Miriam. And then Ed the Millipede unveiled his garden ornament which is now for sale on eBay due to pranksters.
People lie, especially to pollsters.
Pollsters generally have an agenda.
The only poll that counts is the one on election day.
We’ve had trouble with the networks calling elections based on tiny return levels. Like 2%.
C’mon guys. First is nice, but right is better.
Well I predicted my defeat – when I met people on Thursday the first words I said were “I have lost” – even though the count was not till Saturday.
Many factors can be blamed – bad leaflets, lack of canvas cards, whatever….
But the bottom line is simple.
People want politicians who tell them they can have the government spending they want – without any bad consequence.
The government spending to be paid for by magic spells – such as “government investment” (for example buying property at the peak of a property boom and expecting to make a profit to be spent on “public services”).
The public do not want people like me.
“Why do politicians not tell the truth” the scream goes.
“Because if we tell the truth we LOSE”.
That is the answer.
I voted Conservative in an SNP safe seat. Only a fool would openly admit to doing so to anyone, pollster or otherwise. Of course it didn’t make any difference. I should have just voted UKIP since they chime more with my conscience, but I was trying to vote tactically.
The bizarre situation we find ourselves in is that a majority of the British public hold political views that it is socially unacceptable to express in public for fear of ridicule, harrassment or ostracization. If this isn’t evidence in favour a left wing media controlling the narrative, I don’t know what is.
Natalie, is it possible the shy tory effect is growing stronger. In 1992 it was there, and the pollsters then tried to adjust for it in 1997 and after. However hate speech laws, and other ramping up of “you can’t say that” have appeared since then. This election, there were also more physical forms of intimidation: Farage and family were attacked in at a pub meal; a man was punched for having a UKIP sign in his garden; a tory candidate had her garden ruined and her car vandalised. In the indy referendum, natz invaded peoples gardens and smashed their posters with boots and knives – and if the owner repaired it, they came back again the next night. Several times, I heard people say, “I’m ‘No’ but I’m not putting a poster in my window – I’m afraid of a brick through it.” (For the record, I never saw such a smashed window though I did see lots of garden and field poster vandalism.) As regards the person, not property, I once saw the Glasgow kiss, and I heard from friends of some minor incidents – being spat on and the like. Could it be there’s more reason to be shy now than 20+ years ago – the pollsters are allowing for then, not for now.
By the way, I spell natz with a z instead of an s above, not so much as others do to imply their similarity to a certain other party, but because the spell checker in this web widget otherwise insists on correcting it to hats. 🙂
What will happen in Nov of ’16? It is fairly obvious this far out. It will be a referendum on Prohibition. At least 8 States will have Prohibition on the ballot in ’16. And despite the unpopularity of the Democrats they could pull a 1932 and get in on the back of anti-Prohibition sentiment.
Barring a Rand Paul nomination of course.
What is delicious about the Prohibition stuff is that the Left is fighting State Power and the Right is working to maintain it.
I’m getting the Left to fight against State Power.
Which is another way of saying I’m double crossing them. Openly. And they LOVE it.
Well it isn’t just me. But I am in the thick of it.
And I give this to my lefty friends to ponder on: Every tax, every regulation comes with it an army of bureaucrats and behind that an army (with guns) of enforcers.
bloke in spain
May 9, 2015 at 1:55 pm
In America (aside from a libertarian contingent among Republicans) we have two Big Government Parties. What they differ on is the section of Government they want to enlarge.
“Political tags–such as royalist, communist, democrat, populist, fascist, liberal, conservative, and so forth–are never basic criteria. The human race divides politically into those who want people to be controlled and those who have no such desire. The former are idealists acting from highest motives for the greatest good of the greatest number. The latter are surly curmudgeons, suspicious and lacking in altruism. But they are more comfortable neighbors than the other sort.” – Robert A. Heinlein
Surely if the polls were always right that would influence whether people would actually vote? “my guy is going to win anyway, why bother voting?”
ISTM that the glee over the pre-election polls being off is misguided.
All of those who are actors in the election (including the voters) can more effectively pursue what they want to achieve if they know what the situation is.
Suppose a poll told me that my constituency was “safe”, either for the party I heartily dislike, or for the party I support. I actually like the local candidate, but either he has no chance or is a shoo-in. I therefore decide to cast a “protest vote” for a minor candidate with no chance of winning, just to “send a message” to the national party. Then I discover that the polls were wrong, the vote is very close, and the candidate I like loses to a candidate I loathe.
If the polls had been accurate, I would have known better. Should I be happy they were wrong?
Or suppose I am a political activist with a multi-part agenda. My organization is about to fight an election in order to get the power to enact this agenda, or at least part of it. Is it bad for us to know what part of our agenda is the most popular, so that we can campaign on that issue, win, and get something accomplished? Or for us to find out that a lot of the public is (mistakenly, in our view) alarmed about part of our agenda, and back off on it, while we try to correct the popular misapprehension with accurate information?
There is an attitude which says it is somehow ignoble to adapt to circumstances in a battle or controversy; that it is more honorable to lose completely than to compromise in any way. I don’t agree. And to adapt to circumstances means knowing what the circumstances are.
Natalie –
Mike McNally, over at PJMedia, makes an interesting case for why the Shy Tory thing made a re-appearance this election.
I’m not so sure the polls were really wrong. People forget that they had a margin of error of about 3%. The results were at those extremes.
Tim Evans – known to many Samizdatistas – has been predicting a Tory victory for months. His reasoning, insofar as I understand it, goes like this: David Cameron is a strategist. His strategy has been to politically approach his opponents. In the case of Miliband it would be to say something left-wing. In the case of Farage it would be to indulge in some mild immigrant bashing. The result would be that in order to demonstrate they are different from Cameron his opponents become more extreme: Miliband more left-wing, Farage more nationalist. Thus we get to the situation in the general election where Cameron looks sane while all his opponents look like loonies.
That is my understanding of what Tim has been saying and if I am wrong I apologise.
Suffice to say I think it’s hogwash. My theory is that the English did not and do not want to be governed by the SNP and the moment that looked like a possibility they made damn sure they weren’t.
To reuse slightly edited version of my comment posted on a different article:
Ed Miliband was a comically bad candidate, a Wallace without a Gromit. Cameron won because Labour did all the heavy lifting for him. Indeed the Tories won because Miliband told the truth about what Labour wanted to do, and thus scared the crap out of so many people. I know several folks who were aching to vote UKIP but because they were in marginals, they voted Tory and practically puked on the ballot paper. Not so much “shy Tory” as “nauseated Tory” or perhaps “terrified Tory”… people who very rationally who saw the possibility of a return to the fucking 1970’s as just too dire to even contemplate.
As Paul Marks says speaking about himself, he is not the face of the Tory party… because the face of the ‘Conservatives’ is that spoon faced pillock Cameron, who likes the EU and a Real Big State and and NHS and who has no discernible ideology beyond what Guy Herbert calls “governmentalism”. BlueLabour’s Dave Cameron won because unlike that RedTory Tony Blair, who was a marketing genius, Ed Miliband is just Red and moreover was honest about what his Really Bad Ideas were (and had far far worse ones than Blair).
Moreover the nightmare scenario of a Labour/SNP coalition looked very plausible (it was indeed the outcome Antoine Clarke & I were expecting), which would have taken us back to the era of Harold Wilson, with the added frisson of a bunch of Scottish Brownshirts also having power over England too.
The only thing I would say bravo to Cameron for is making sure he was not David Miliband (i.e having a rival in the party who could knife him)
Rich Rostrom writes in favour of tactical voting with FPTP, and his need for accurate polls to do that:
There is a better way. But the UK Referendum of 2011 on AV came down against it by 2:1.
Back then, I posted here on Samizdata, an analysis of AV showing it allowed voters to provide more information on their desires.
AV supports sending a message simultaneously with not splitting the vote (should that message not be a winning one). And, thank goodness, it does not require opinion polls to be correct.
What AV may well do, though the extent (and political direction) is by no means certain, is take seats away from the larger parties. That’s why they don’t like it, and that’s why the choice of electoral system should de-emphasise their involvement in the decision.
There is quite a big risk that now, with the problems of minor parties in the general election just past, there will be a change to the electoral system: but one that is worse that AV. For example party lists and/or multi-seat constituencies, both of which actually weaken democratic representation of the people, in favour of even more Westminster Village and party oligarchy that we currently have under FPTP.
Please all think about that, and see if ways can be found to prevent it happening.
Best regards
“they voted Tory and practically puked on the ballot paper. Not so much “shy Tory” as “nauseated Tory” or perhaps “terrified Tory”…”
I’ve been looking at UKIP results in the locals, where the UKIP candidate was strong in the Westminster contest. I’d hazard half million of them. But for their fears we could be looking at a very different result. In a FPTP, UKIP would be in the territory where they’d be picking up significant numbers of seats. And spoiling the day for a whole lot of Tories. Most likely we’d have a majority Militwat government, today.
Indeed. And that was simply too damn ghastly a notion for just enough UKIP sympathisers to contemplate… and they voted Tory.
Re Nate Silver, of “100% accuracy” fame, who took a stab at the UK Elections and got it spectacularly wrong, but is actually interesting in his Analysis of why he got it wrong.
Personally, I agree with Perry.
Rich Rostrom writes in defence of polls, “All of those who are actors in the election (including the voters) can more effectively pursue what they want to achieve if they know what the situation is.”
I agree. My glee was because a world where polls don’t always get it right is more interesting, and because in recent history the polls have underestimated opinions relatively congenial to me, not because I have anything against the effort being made to find out what people think.
One of the many, many threats to free speech in our time is the idea that polls should be forbidden in the period immediately before an election in case it “distorts” the way people vote. Asking people in polls what they think, making the answers public, and the public having the right to find out this information if they are interested – all these things are part of free speech.
This election, there were also more physical forms of intimidation: . . . Could it be there’s more reason to be shy now than 20+ years ago – the pollsters are allowing for then, not for now.
Interesting comment from Niall. And interestingly enough, the physical intimidation is aimed almost entirely at the Conservative Party and UKIP.
Where does it come from? I don’t know if anyone has tried to do an objective, comprehensive analysis of this abuse and intimidation, but it seems to me that it comes almost entirely from Labour and SNP supporters – or, to be more precise, from supporters of socialism and/or Scottish nationalism.
So why are UKIP and the Conservatives the only major parties to attract violence – and why are socialism and nationalism the ideologies that seem to be most likely to beget violence?
John, are you really asking “Are supporters of political ideologies that emphasise using force to reorder civil society more prone to use force to get what they want?”… I mean… who’d a thunked it? 😀
John and Perry: shortly before the referendum, someone tweeted, “Travelled from Bearsden to Stirling today. Vast majority of #nothanks banners damaged Yes banners no damage. Strange”. Like Perry, my reaction was to think, no, not strange at all.
Natalie: your latest comment chimed with a thought I had after my last comment. With elections, we get to see when the polls were wrong. As their wrongness consistently favours politically correct parties over others, what discount should we apply to polls telling us that X % support some PC opinion?