The polling feast continues..

Normally this is a time of year when we have to put up with a polling-free period. In August 2006 just two national surveys were carried out and we had to wait until the final week of the month before we got the figures.

This year, with a new prime minister and change in the air, the polls are coming thick and fast and on top of the Sun’s Ipsos-Mori survey, the YouGov poll for the Sunday Times reported last night, we now have an ICM poll for the Sunday Mirror.

This has with changes on the last ICM poll two and a half weeks ago CON 33% (+1): LAB 39%(+1): LD 18%(-2). So not much change there except a confirmation of the seepage of LD support to Labour.

What’s really striking about this latest crop of surveys is that we have three polls which use very different methodologies and all are showing that Labour’s lead is consolidating

The Mori headline figures only include those “100% certain to vote; ICM only include those who say they are 70% certain or more and then apply a pro rata weighting; while YouGov do not make any adjustment for likelihood to turnout. ICM and YouGov weight their findings to ensure a politically representative sample – Mori don’t.

Without the turnout adjustment Mori would have had a 13% lead rather than the 6% that we saw. We haven’t got the detail yet from the latest ICM poll but in their last survey the turnout adjustment reduced Labour’s lead from 11% to the published 6%.

Figures that I like to look at are those that show how those who said they voted in 2005 are now thinking. For as I’ve been arguing recently those who claim a voting record should be given a bigger weighting than those who don’t.

In a discussion here nine days ago the boss of ICM and leading figure in the UK polling industry, Nick Sparrow, made this point: “Labour are now picking up more votes from those who say they didn’t vote last time. In our last Guardian poll 46 (number) people who said they did not vote in 2005 now intend to vote Labour against 19 for the Conservatives. Compare that to (say) March. Then 26 non voters in 2005 opted for the Conservatives and 23 opted for Labour..Historically non voters last time have declared greater support for Labour than the Conservatives and one of the features of the more recent polls up to June was that these people were showing for the first time a greater enthusiasm for the Conservatives.”

In my betting I have increased my spread position on Labour seats. I’ve had a couple of transactions at an average of 311.75. So for every seat above that I’m £32 up – for every seat below I’m £32 down. So if the election went according to YouGov with the Baxter seat calculation Labour would get 90 more seats than my trade and I would make nearly £3,000. Somehow I don’t think it’s going to be that easy. I’ve got out of my spread position on how long Gordon will have as PM before the general election.

Mike Smithson