Everything That Pollsters Could Get Wrong in U.K.'s ElectionBy
Turnout, samples and Brexit all create potential traps
‘We won’t know until election day if they’ve fixed it’
In 2010, they couldn’t conceive of a coalition government. Five years later, they didn’t think the Conservatives would win a majority. Few made the right call on the Brexit referendum. So what could go wrong in 2017?
As they prepare for the June 8 election, humbled political analysts point to the judgment calls they have to make that could affect how accurately they predict the margin of victory given the double-digit lead Prime Minister Theresa May’s Tories have over the opposition Labour Party.
In France, the pollsters got it spot on for Sunday’s first-round presidential election. But in the U.K., with its first-past-the-post system, they could miss the underlying story even though the headline result of a Tory win can be correct. Will voter fatigue keep people at home? Has Brexit redrawn political divisions? Which historical rules should be ignored -- and which still hold? Here are some of the dangers that lie in waiting.
Talking to The Wrong People
The probe into the 2015 polling disaster concluded that the main cause of the error was unrepresentative samples: Pollsters spoke to too many Labour supporters and too few Tories. But it warned this problem wouldn’t be easy to fix.
“The pollsters have made methodological adjustments,” said Will Jennings, politics professor at Southampton University, and one of the members of the inquiry panel. “We won’t know until election day if they’ve fixed it or over-fixed it.”
Hidden Power of ‘Can’t Be Bothered’
After the 2015 election and the 2016 referendum, “even people whose job it is to get excited about elections are fed up of elections,” said Robert Ford, professor of politics at Manchester University.
He pointed to the 2001 election, another vote held in June where people felt they knew the outcome in advance. Turnout was 59 percent, the lowest since 1918. “Who stays home then becomes really critical for what happens,” he explained. “A big swing from Tory to ‘can’t be bothered’ would be possible, and not something we’re thinking about.”
Not That History Lesson
Turnout was one of the things that misled polling companies with Brexit, according to Joe Twyman, head of political polling at YouGov Plc. “We were working on the assumption that habitual non-voters would not turn out,” he said. “And they did.”
That’s the problem with assuming that rules that have applied in the past will still apply. In 2015, it was hard for experts to believe polls predicting that Labour would lose all but one of its seats in Scotland, one of its historic heartlands. Equally, the Liberal Democrats had in previous elections demonstrated an ability to hang onto seats they already held. So it was believed that they would again.
But other electoral rules that people chose to ignore held true: Labour had consistently polled badly on questions of which party had the best leader or would be best at economic management, and those once again proved an excellent predictor of the result.
Missing The Detailed Picture
It wasn’t just the opinion polls that were caught out in 2015. The exit poll, a huge survey of how people had actually voted that is held up as the gold standard of prediction, initially predicted the Conservatives would be 10 seats short of a majority.
While it was closer than anyone else, it couldn’t get the individual battles in each district that would get the Tories an extra 15 seats. It isn’t just what proportion of the vote a party gets that matters, it’s where it gets it. The Scottish National Party came fifth in terms of national vote share, but with all its votes in Scotland, it came third in terms of seats in Parliament.
The Brexit Effect
In Scotland, the 2014 independence referendum had the effect of realigning voters. The most important issue was now whether they identified as nationalists or unionists. Has Brexit done the same in the rest of the U.K.? If so, the Tories could pick up seats in the north of England that have always been Labour, but now identify as “Leave,” and the Liberal Democrats could take prosperous seats around London that feel “Remain.”
Polls Could Affect The Outcome
“It’s a situation we haven’t been in before,” Twyman said. “A 20-point lead going into a campaign doesn’t happen. How that affects people, how that affects the dynamics of the campaign is unknown.”
It’s already having an impact. Labour candidates are publicly disowning their leader. Will that help them or hurt them? As the Tories attack Jeremy Corbyn, will it repel voters, or reassure them that he’s so unlikely to win they can safely vote for a Labour candidate they like?
With Many Unknowns, One Certainty
If May maintains her lead up to June 8 and then Labour wins, “then polling is stuffed,” said Twyman. “No individual final poll in the history of post-war British polling has ever been more than 6 percent out. A miss of that scale would be the end.”