How did pollsters get this so wrong?
THE BROADCASTERS’ exit poll provoked a huge row last night – with Labour immediately claiming it was ‘wrong’.
The poll predicted a radically different result from that which had been predicted throughout the campaign, with the Tories having a significant lead and Labour nearly 100 seats behind.
And if it proves right, last night would be a disaster for all the polling companies who have attempted to gauge public opinion throughout the campaign.
They have shown the parties neck and neck, with only a few suggesting the Tories had a narrow lead. A separate YouGov post election poll last night continued to show the parties close together.
A BBC poll of polls, which accumulated results from all pollsters during the campaign, had put the Tories one point ahead of Labour on election day.
If the exit survey proves true, it would be the worst result for the pollsters since 1992. Then, in the run up to polling day, they predicted a hung Parliament with Labour ending up as the largest party. The phenomenon of so-called shy Tories – Conservative supporters who refuse to tell pollsters how they voted – was blamed.
In 1992, like last night, the exit poll showed a markedly more favourably result for the Tories. It suggested the Tories would be short by 25 seats in a hung parliament. By the end of the night the Tories were in even better shape and John Major finished with a majority of 21.
Former Labour spin doctor Alistair Campbell told BBC News the exit polls didn’t ‘feel right’. But he added: ‘If this is right, there’s going to be a pretty big look at the polling industry.’
Yesterday’s exit poll put the Tories on 316 and Labour well behind on 239. It suggested the Lib Dems will get ten MPs, the SNP 58, Plaid Cymru four, UKIP two and the Greens two. A Labour source continued to insist: ‘We are sceptical. It looks wrong to us.’
Shadow chancellor Ed Balls said the poll was an ‘outlier’ that was ‘completely at odds’ with those during the campaign. Even if it right, it puts David Cameron’s majority ‘right on a knife edge’, he said. The exit poll is a much bigger poll than those conducted during the campaign which normally involve around 1,400 people. It cost some £300,000 and involved interviews with 22,000 people in 141 locations in 133 seats. It was carried out by three different pollsters, GFK, NOP and Ipsos/Mori for BBC, Sky and ITV News.
Peter Kellner, president of YouGov, said the exit poll was much more likely to be right because you could be ‘absolutely certain’ that those questioned had voted. He told BBC News: ‘It’s a very, very big sample – bigger than any conventional sample.’
Another factor in polls getting it wrong is undecided voters. A Daily Mail ComRes poll at the end of April showed four in ten of those who planned to vote were totally undecided how they would vote.