Life on the margins
The deluge of polls published over the past week has pointed to one conclusion: the most the Conservatives can hope for in the coming election is a hung parliament. David Cameron and his party look likely to be roughly ten to 15 seats short of the 325 they require for a majority.
The two most significant recent surveys both focused on the marginals, and returned very similar results. The Channel 4 News/YouGov poll was carried out in 60 Labour-held seats requiring a swing of between 4 and 7 per cent for a Tory win: the survey found a swing of 6.5 per cent.
Meanwhile, the Populus poll for the Times covered 100 seats - those numbered 51 to 150 - on Labour's "most vulnerable to the Tories" list. The swing was 6.7 per cent.
Headline vote shares from polls of marginal seats often get reported as if they could be counted the same way as the results of standard national polls. But in fact they have different baselines. In 2005, for example, Labour had a 3 per cent vote lead; in the seats polled by Populus, that figure was 14 per cent. So the vote split suggested by the most recent poll, 38:38, would deliver 97 Labour seats to the Tories. According to the YouGov survey, the figure would be 95.
The remaining 20 or so seats needed for a majority would have to come from the Liberal Democrats. But the signs are that their incumbents will be tougher to shift - whatever the national polls suggest.
Extrapolating the Populus data to a national scale, we find that it is broadly in line with the latest ICM survey for News of the World, which had the Tories back at 40 per cent - 9 points ahead of Labour.
Just about the only thing that can be said with any certainty about the coming election is that there will be more opinion polls, published by more pollsters, than ever before.
At the last count, there were ten firms measuring public opinion - each using its own bespoke methodologies and mathematical approaches.
In addition to the five pollsters that were around in 2005 - Ipsos MORI, ICM, YouGov, Populus and ComRes, all of which publish surveys every month - we can now expect polls from five other firms, which are either new to UK political polling or only produce polls intermittently. These are Harris Interactive, Angus Reid, TRS-BRMB, BPIX and Opinium. This last firm entered the fray only this month, polling for the Daily Express.
But despite all the new voices, the numbers supplied by the pollsters that people are most familiar with will, inevitably, attract the most attention. Results from ICM and YouGov, in all likelihood, will still be the most discussed.
Follow my leader
It is easy to forget that pollsters ask questions about a lot more than just voting intentions.
But polls on other issues can also provide useful information on the likely outcome of elections. In the past, leaders' approval ratings in particular have proved to be a good indicator.
Ipsos MORI has been asking voters what they think of the party leaders since 1977, and this information came into its own at the 1992 general election. Judging on the basis of voting intentions, none of the pollsters foresaw John Major's victory. However, his personal approval ratings were consistently in the high 40s or low 50s - putting him well ahead of the Labour leader, Neil Kinnock.
Since last summer, there has been a noticeable improvement in Gordon Brown's standing, while Cameron's ratings have suffered a bit of a decline. There is still a marked gap between them, but if the current trend continues, who knows? The Prime Minister might still pull off a sensational result.
Mike Smithson is the editor of politicalbetting.com