“Alongside Labour, the Liberal Democrats and UKIP, the big losers are the pollsters. It was like 1992 all over again, but worse. The current result in no way resembles what they were predicting 24 hours ago. From an academic perspective there are a couple of reasons why this might be.
1) Shy Tories – people who were always going to vote Tory but didn’t want to admit it to pollsters. This is perhaps the trickiest thing to deal with as a pollster as there are very few ways to accurately measure the opinions of people who are determined to lie to you.
2) The data might have been right but the modelling was wrong. Most polling models have been based on previous elections and the rise of new parties might mean a fundamental rethink of how this is measured.
3) Problems with polling. Polls are usually done by either phone calls to landlines or the internet. Both have their issues in terms of sampling, less people now have landlines and internet usage is often skewed. This is something that will have to be carefully rethought. Possibly bigger samples from more constituencies will be the answer.
4) The final possibility is that the polls were right but there was a massive shift in the last 24 hours of the campaign. This is a real possibility. We know that lots of UKIP voters were former Conservatives. In the dying hours they might have looked at the close polls and decided that they preferred the guarantee of a Conservative government and an EU referendum over the risk of Ed Miliband and Labour. As a result they all suddenly decided to go home to the Conservatives as Cameron had been urging for months.
“Regardless of all of this. The polling companies will now have to devote huge amounts of time, effort and resources to improving their methods. What’s the point of a pollster who can’t even predict a general election?”