Is there the evidence to make the toughest call in British politics?

This is Deborah Mattinson, joint chief executive of Opinion Leader Research, a pollster for the Labour party for the past two decades and the person who will advise Brown on whether to go for an early general election.

If Brown does decide to go in October it will be because Mattinson is convinced not only that Labour cannot lose but that Gord will almost certainly top Tony’s 2005 performance.

That’s a big ask and Deborah would not be doing her job unless she was challenging every minor aspect of all the polls in order to provide the best advice. And a key element for Labour is how certain it is that those who say they will vote for the party will actually do so.

Take for instance the worrying line of data in last week’s ICM poll. To the non-judgemental question – “Some people have said they would not vote in a new General Election, while others would go and vote at their polling station. I would like to know how certain it is that you would actually go and vote in a general election?” a total of 375 respondents naming a party and saying they were “certain to vote” split CON 141: LAB 131: LD 63: OTH 40.

Labour got its 6% lead in that survey from the responses of those who were less than 100% certain. There was a similar pattern in the last Communicate Research survey which had the Tories enjoying a substantial lead amongst the “certains”.

But YouGov, the pollster which has been showing substantially bigger Labour leads than the other firms, does not adjust its voting intention figures to take this element into account.

If I was Deborah the YouGov approach would make me a bit nervous for in an era of 60% general election turnouts the likelihood that electors will vote is, surely, a critical measure?

The only poll in today’s papers is from YouGov in the Sunday Times and is restricted to Scotland. To questions about about Westminster voting intention it found with changes on the 2005 general election result – LAB 40% (nc): SNP 31% (+13%): CON 14% (-2): LD 11% (-12%). So while last week’s GB-wide YouGov poll reported that Labour was up by a sixth on 2005 the internet pollster that does not adjust for turnout found that there’s a standstill in Scotland.

There should be the Guardian’s August ICM poll out this week and it’s likely that Communicate Research will produce its monthly survey for the Independent. The latter, like ICM, has a fierce turnout filter.

In my betting I’ve returned again to the Spreadfair “length of Gordon’s first term” spread market. Last week I closed down my 82 week buy trade. Last night I bought at 80 weeks which takes us into early 2009.

Mike Smithson