What the pollsters think will happen at the general election

by Stephen Tall on January 24, 2010

Forget data sets, interquartile ranges and margin of error. The Guardian recently reported the collective wisdom of the wet-fingers-in-the-air of the UK’s pollsters, who met this past week “to refine their methods ahead of the election, and ended with off-the-cuff predictions for the final result.”

And here’s what they came up with:

Statisticians from most of Britain’s main polling companies attended the session, organised jointly by the British Polling Council and the National Centre for Research Methods.

Four of them were brave enough to come up with predicted vote shares for the main parties. Put together they average a shade under 40% for the Tories, just over 30% for Labour and 21% for the Liberal Democrats.

In terms of seats, one estimate suggests those figures would leave Cameron 10 short of a majority. That would make Nick Clegg, with about 53 seats, the most powerful Liberal leader since Lloyd George. It would also leave the Tories – hoping for a majority – reeling. But Labour would have suffered most, with a loss of 105 seats on election night and just 251 surviving MPs, against 316 Conservatives.

Most pollsters think the Tories will do better than that. Two stood apart from this consensus and skewed the overall average: Nick Moon, from GfK NOP, suggested the Tory lead would be a tight 8%. Ipsos Mori called an even closer result: 36% for the Conservatives and 32% for Labour – which if it happened would leave Gordon Brown clinging on as a minority prime minister.

Everyone else – including statisticians from YouGov, Populus, ICM and Strathclyde University – thought Cameron will get his majority, but only just.

The 40-30-21 split seems pretty plausible to me. It would represent a 6.5% swing from Labour to the Tories; and a 4% swing from the Lib Dems to the Tories. However, if the Lib Dems do poll 21%, I’d be surprised to see the party slip back to 53 seats, even if that’s what the national swing might indicate. Let’s recall that the Lib Dems won 52 seats in 2001 with 17% of the national vote.

And with the Tories reportedly giving up on marginal Lib Dem seats, such as Cheadle, and with Labour highly vulnerable in 12 seats where swings to the Lib Dems of less than 5% are needed, it seems more than possible that, even if the pollsters are right and the party’s national poll rating slips back slightly compared with 2005, the Lib Dems may well still end up making a net gain of seats in 2010.