A prominent university lecturer has explained the dreadful predictions made in the run up to the general election last year.
Ahead of May’s vote, pundits and pollsters said the Conservative Party was, at best, headed to enter another coalition government… or even occupying the opposition benches.
But Professor John Curtice, the man behind the poll that most accurately predicted David Cameron's success, has published new research that says why survey companies got it so wrong. It’s because pollsters interviewed too many Labour supporters.
There were two major errors by pollsters. More time and energy should have been invested into finding and quizzing Tory voters, and they didn't properly identify which groups of people would abstain, and so disproportionally weighted their data.
Curtice appearing on the BBC's election night special
Snap polls, Curtice explains, are conducted over just two or three days, meaning pollsters themselves are more likely to interview those who able to be contacted most easily - either over the internet or via their phone.
Evidence from last year's British Social Attitudes survey suggests those who can be contacted most easily are less likely to be Conservative voters.
Of those who were contacted with greater ease – in this case agreed to be interviewed the first time an interviewer called - Labour enjoyed a clear lead of six points; Tory backers were eleven points ahead amongst those who were only interviewed after between three and six calls had been made.
Conservative supporters, though in the majority, were harder to find
Polling companies also failed to weight their data based on the likely turnout of different demographics properly, such as young voters.
The 2015 BSA survey showed 18 - 24-year-olds were around 30% less likely to vote than people aged 65 or more. But most polls anticipated a much smaller age gap than this.
Any tendency among surveyors to overestimate the turnout of younger voters, Curtice explained, meant there was a particularly strong risk that Labour support would be overestimated, leaving the poll drastically skewed.
The senior research fellow at NatCen blasted pollsters, saying: "A key lesson of the difficulties faced by the polls in the 2015 general election is that surveys not only need to ask the right questions but also the right people.
"The polls evidently came up short in that respect in 2015."
He went on to offer survey conductors advice, adding: "Random sampling, time-consuming and expensive though it may be, is more likely to produce a sample of people who are representative of Britain as a whole."