Pollsters disagree on who is ahead in UK general election race

The political parties are not the only people on tenterhooks ahead of the election on May 7: the reputations of Britain's pollsters are also on the line.

For now, they cannot agree who is in the lead: on Monday, Lord Ashcroft put the Tories ahead by 6 points, suggesting a majority for David Cameron. Populus said Labour was 3 points in front - a win for Ed Miliband's party.

Most other polls point to a hung Parliament. So what explains the big differences between individual polls?

Part of the answer is the margin of error inherent in any survey: typically this is about 3 percentage points but because a huge number of polls are published in the run-up to an election, inevitably some will err by a wider margin.

"If I say it is 32-24 [for each main party] it could easily be the other way around and still be within that margin of error," said Ben Page, chief executive of Ipsos Mori.

"Humans are hard-wired to look for patterns . . . but within a small period you might see a trend that is actually random noise," said Will Jennings, an academic at Southampton University.

Then there is the difference between how pollsters make their calculations - the "house effect".

In Britain the industry is not partisan, unlike in the US: there are no "Conservative pollsters" or "Labour pollsters".

But a different result will be caused by methodological differences such as prompting respondents with party names, weighting, adjustment by likelihood to vote, sample sizes and dealing with "don't knows".

<

The tabular content relating to this article is not available to view. Apologies in advance for the inconvenience caused.

Instead the polling companies have been fairly consistent in terms of their own overall findings, he said.

Four pollsters (ICM, Lord Ashcroft, ComRes and Opinium) have had the Conservatives firmly in the lead for the past two months. Three (Panelbase, Populus and Mori) have had Labour ahead. Just two - YouGov and TNS - suggest that support is switching between the two main parties.

Only Survation has been pointing to one party surging past another in recent weeks: the Conservatives.

"I don't think that the polls are confusing, it is just that some are right and some are wrong," said Mr Pack. "We won't know which are which until polling day."

The FT's latest poll of polls suggest Labour and the Conservatives are in a deadlock: with Labour on 33.2 per cent and the Tories on 33.

This figure includes the most recent poll from every pollster, with older ones given less weight.

Yet one trend is emerging from the detail of the survey work: pollsters who use telephone polling are more likely to find the Tories ahead, by about 1 percentage point. Those using the internet to collect data typically put Labour one point ahead.

<>These results seem to fly in the face of the usual pattern where online polls find more support for the Tories. Some experts have suggested this is because there remains a social stigma to telling someone verbally you support the party, and this effect is removed in an anonymous online poll.

Alternatively the "shy Tory theory" may just be a myth, according to Tom Mludzinski, head of political polling at ComRes. "It is arguably just as unfashionable to say you're going to vote for Ed Miliband," he said.

It could be that telephone surveys find more Conservative voters because they are more likely to have landlines. But pollsters do weight the results accordingly.

Instead, some in the industry believe the reversal of this pattern may be because companies are aware of this bias and are overcompensating.

Some suggest the split is between the newer and the more established polling companies, with the former favouring Labour and the latter more likely to give leads to the Conservatives.

One new online company, Survation, suggests more support than others for the UK Independence party, at the expense of the Tories.

<>If history is a guide, the established companies are more likely to be right: the newer companies overestimated Ukip's vote share by about 4-5 points in the run up to last year's European election.

In 2010, Ipsos Mori and ICM correctly predicted the final result for Labour and the Conservatives: 29 to 36 per cent. ComRes and YouGov were only a point out. Yet all four overestimated support for the Lib Dems in a big way: by up to 5 points.

Mr Page said that face-to-face or phone polling was still more reliable than online polls, not least because only 85 per cent of the country has internet access.

A gap of a few percentage points might not have mattered when the result was clearly going one way, for example to Labour in 1997: this time it could make or break forecasters' reputations.

"Someone is going to have egg on their face but we don't know who," said Rob Ford, a political scientist at Manchester University.

"We just know that this is a very difficult election to poll."

© The Financial Times Limited 2015. All rights reserved.
FT and Financial Times are trademarks of the Financial Times Ltd.
Not to be redistributed, copied or modified in any way.
Euro2day.gr is solely responsible for providing this translation and the Financial Times Limited does not accept any liability for the accuracy or quality of the translation