Lying to Pollsters: The Counter-Weapon du Jour
Political pollsters are probably getting jittery as they head into the 2020 US Presidential election season and will again have to predict the outcome. There was, of course, much wailing and gnashing of teeth in the polling community after the complete polling botch of the US 2016 Presidential election that widely predicted a near-certain Hilary Clinton win. But they probably shouldn’t feel so conspicuously bad. That botch-job was part of a broader phenomenon.
The miscalls during the 2015 British elections were even worse. The majority of pollsters predicted an Ed Miliband-led Labor minority government and no pollster came close to predicting the 6.5 percentage-point win that produced a majority government for David Cameron’s Conservatives. And that followed closely the polling for the 2014 referendum on Scottish separation, which trended toward a narrow win for independence in the days leading up to the vote, but then a 9-point victory for remaining in the actual vote.
There have been lots of theories about why that as analytical techniques have gotten stronger, the polling predictive quality has headed dramatically south. Is it sampling techniques? Is it inability to contact some voter segments? Is it ‘shy Trumpers’ who are ashamed to say they are voting for Donald Trump? Nope. All of these reasons presume one thing about polling: passive cooperation. Instead, respondents are actively using the pollsters’ weapon against them by lying to them.
To understand, we have to look back through the history of weaponry. Throughout history, every effective weapon has begotten a counter-weapon. Swords begot armor. Charging horsemen begot caltrops. Shore artillery begot dreadnoughts. Nazi U-boats begot Allied sonar. ICBMs begot ABMs. The universal rule is that targets don’t just sit there and get devastated by the advent of a great weapon. In due course, they invent a counter-weapon.
As political polling rose from a cottage industry in the middle of the 20th century to a ubiquitous science in the late 20thcentury, it proved to be a powerful weapon wielded by politicians. They could figure out exactly what to say and how to say what they said to achieve their goal of winning political office. While this powerful new science could be used for beneficial purpose of better understanding constituents in order to serve them, it was also used cynically to manipulate voters with empty and inconsistent promises.
Did electorates around the world simply acquiesce to the firing of this new weapon? No. They got creative and invented an awesomely powerful and neutralizing counter weapon — on the scale of sonar. That is: we are not just going to stop your subs from sinking our ships. We are going to annihilate your entire sub fleet. Their counter weapon is lying to pollsters.
As is often the case with innovation, it happens where the need is highest. I first noticed the advent of this new counter weapon when it was used twice in quick succession in Chile and Nicaragua. In 1988, Chileans went to the polls in a plebiscite to determine whether military dictator Augusto Pinochet would have his reign extended another eight years or be forced to hold a democratic election. Pinochet held an overwhelming lead throughout — 11 points as of voting time. Yet he lost in a landslide — 55% to 43%.
The primary explanation at the time was that the electorate was fearful of answering the pollsters honestly in case the ‘pollster’ was actually a Pinochet representative in disguise. That explanation entirely underestimated the cleverness of the Chilean electorate. They knew the scoop. If they answered honestly and the polls went against him, Pinochet would have called off the plebiscite. They had to bait him into holding the plebiscite on account of his confidence that he would win. All that was required to guarantee their freedom from the dictator was to lie to the pollsters, then vote with their hearts — and they did both.
This was repeated in Nicaragua in 1990. The Sandinistas had overthrown the oppressive dictatorship, but in a classic case of ‘meet the new boss, same as the old boss’ they decided that ruling was kind of fun. However since their avowed purpose was to wrest the country from the clutches of dictatorship in order to replace it with a benevolent democracy, they were largely trapped into going through the troublesome ritual of running a democratic election. Sure enough, they did and in the lead up to the election, they led in the opinion polls by a wide-margin — 48% to 32% in the final poll. But they too lost by a landslide — 55% to 41% — and never had the pleasure of leading, except by force.
I’ve watched ever since as this innovation created out of the crucible of desperate times spread to more a more basic utilitarian use: getting politicians to do what you want rather what they want.
The Scots did it exquisitely in the 2014 referendum. They didn’t want to separate. But a ‘no’ answer to the pollsters wouldn’t have resulted in any benefit for Scotland. A ‘yes’ answer worked like a charm. It made Cameron desperate enough that he gave Scotland whatever it wanted to avert independence. Perfect. All it took was lying to the pollsters to accomplish what was otherwise out of reach.
Apparently, the Brits in their 2015 general election learned from the Scottish voters that Cameron could be massively manipulated. Overall, they didn’t trust Labor to manage fiscally but were mad at the Conservative for being so austere. They needed to get the Conservatives to act like liberals and all that took was a little lying to pollsters. They got him to promise completely un-Conservative policies including a big funding increase for the National Health Services, tax-free childcare, and a rail freeze for commuters. And the isolationists even squeezed a senseless, and in due course draconian, referendum on the European Union out of him by polling positively for UKIP. They kept up their pressure to the last day and managed to successfully manipulate their way to a kinder, gentler Cameron.
And in 2016, enough Wisconsin Trump voters convinced Hilary Clinton she didn’t have to visit their sure-thing-winner state a single time between the April primary and the November election by claiming they were voting for her, producing a 7-point swing between the RealClearPolitics final average and the actual Trump win. If the Blue Wall was your enemy, it sure didn’t take much to rope-a-dope your adversary and it was effective because your adversary thought you would simply sit there and let a weapon be freely used against you — for the first time in history.
I know lying is bad, but I must admit that this is an appealing if subversive counter weapon. One thing is for sure: as the counter-weapon becomes more ubiquitously used, political polls will go the way of the dreadnought and the U-boat.