Do his polling advisers need to delve deeper?

In trying to explain how Gordon Brown and Labour have got into their current mess there appears to have been one driving force – the massive Labour poll leads that at one stage touched 14%.

The party believed that everything was going its way and, inevitably, there was pressure to convert the apparent surge in support into a fourth successive Labour victory.

Yet were the poll leads ever as big as they seemed? Should the party have probed a bit deeper to assess what was really happening?

Over the past three months along with one or two other poll watchers, notably Mark Senior, I have been pointing out that a significant part of Labour’s poll lead had come from people who said that they did not, for whatever reason, bother to vote at the 2005 general election.

And I have also argued that the views you should give most weight to are those that at least tell the pollsters that they did turn out last time. Those with anything more than the minimal` experience of campaigning will tell you that the best guide as to whether somebody will vote is what they did last time.

Surely 2005 non-voters are less likely to turn out next time than those who did?

So in last week’s Populus surey which had Labour two points ahead just over three out of five intending to vote Labour had actually turned out in 2005. The Lib Dem proportion was about three quarters while the Tory figure was just underr four-fifths.

This has been a trend that has been seen throughout the summer. The Labour lead was based on the flimsiest of foundations.

My guess is that when we see the detailed ICM data from today’s survey it will show that a part of the shift to the Tories has come from the non-2005 voters.

Mike Smithson