LONDON (Reuters) - According to opinion pollsters, British Prime Minister Theresa May is either set to win a resounding election victory on Thursday or to scrape home with her authority in tatters.
The pollsters, who failed to predict David Cameron’s unexpectedly strong victory in the 2015 election and then gave confused signals on the outcome of the 2016 Brexit vote, now seem unsure about how the 2017 election will turn out.
Why is there so much confusion?
The latest polls all show the Conservatives' lead narrowing, but the advantage ranges from 3 to 12 points. Voters are not giving different answers to the pollsters, but the polling companies cannot agree which voters will turn up on the day. (For an interactive graphic on the election: tmsnrt.rs/2pgjH8p)
“None of us are stupid, none of us are crazy, we all may be wrong in the end or some of us may be wrong, but it’s just a difficult problem,” said Ben Lauderdale, an associate professor at the London School of Economics who helped create YouGov’s new electoral model.
“The question is what is the alternative? It’s not clear there is one,” Lauderdale said.
His model, using a technique called multilevel regression and post-stratification, currently predicts May’s Conservatives failing to secure an overall majority.
The polls underestimated the Conservative vote by four percentage points and overestimated Labour by three points, leading to the widespread view that they “had got it wrong”, according the British Polling Council (BPC).
Reports put the error down to factors like “shy Tories” - Conservative voters who were reluctant to explain their preferences - but the pollsters said a bigger problem was accurately forecasting who would actually turn up to vote.
The polling companies have adjusted their models in various ways to reduce the influence of those who are politically engaged and increase the weighting of those less likely to vote.
“The difference in the polls in this election is easy to understand - it is almost wholly to do with how pollsters treat turnout,” YouGov pollster Anthony Wells said in a blog post.
“Generally speaking, the polls that continue to show a large Conservative lead are those who are basing their turnout models on the pattern of turnout in 2015. Those that show smaller leads are basing turnout on how likely people say they are to vote.”
Some 246,487 young people registered to vote on May 22, the last day to sign up before the election cut-off. That was up from 137,400 at the cut-off date before the 2015 vote, and those aged between 18 and 24 skew heavily towards Jeremy Corbyn’s Labour.
But just 43 percent of those in this age group voted at all last time, far below the percentage of older voters.
“These are strange times so we don’t know what will happen, but the Labour 40 percent share ... does include a lot of younger voters who have not voted before, people who don’t normally vote,” Ipsos MORI chief executive Ben Page said on Friday, after his company put the Conservatives on 45 percent and Labour on 40 percent.
“They are telling us they are going to vote this time but they have not done so before.”
The referendum on Britain’s EU membership galvanised a lot of older working class voters last year, bringing many people to the voting booth for the first time or after a long absence.
It is not clear whether these voters will keep the habit, and whether they will back Labour or switch to the Conservatives who have been more explicit in promising to deliver Brexit.
The polling companies changed their methodologies after the 2015 election to tackle the bias towards those who are politically engaged.
The puzzle in 2017 is whether the Brexit voters have remained engaged or not, and if so, which party will attract their votes.
Britain’s first-past-the-post electoral system does not make the pollster’s job any easier.
The national vote share is not a prediction of how the 650 seats in the House of Commons will be divided among the parties, because coming second or third in individual constituencies does not result in any seats.
“Political context matters a lot in UK elections,” Lauderdale said. “People who voted Labour last time, to give an example, behave differently according to what kind of constituency they are in.”
With the pollsters all testing so many variables and adjustments in this election, it is not surprising that there is a huge range of forecasts, Lauderdale said.
“As far as I can tell, the major differences I am seeing are not down to the raw data that the different pollsters are getting,” he said. “Different pollsters are envisaging a different set of people turning out.”
YouGov pollster Wells said that if younger voters turned out on election day, then May could win a much smaller majority than initially predicted or even lose her majority.
“The alternative is that all those young Corbynistas will prove a mirage and that some polls still contain too many of the sort of young people who vote, with the end result being that the Conservatives win a large or landslide majority,” he said.
Editing by Guy Faulconbridge and Tom Heneghan