The poll that caught me by surprise

My apologies for not getting a thread up earlier but the latest Guardian ICM poll has taken me by surprise – I wasn’t expecting it until next Tuesday which would have followed the paper’s normal pattern.

Also the fieldwork took place at the end of the half term week in many places when a lot of people are away. Pollsters tend to avoid such periods because they have thrown up odd results in the past.

The shares are with changes on the last ICM poll CON 37% (nc): LAB 34% (+2): LD21% (nc). So good for Labour and the Lib Dems but disappointing for the Tories.

For a commons seat spread-betting gambler like me the current polls are a nightmare. They range from producing just enough seats for the Tories to form a majority government to just enough for Gordon to hang on to power. I’m passing at the moment.

The ICM survey was carried out at the end of a week when the Commons wasn’t sitting and when there was not that much political news about. We didn’t see much of the Tory leader in the days running up to the fieldwork on Friday. The poll also missed the Northern Rock news.

One factor about ICM is the way it calculates the likelihood to vote. Those rating themselves as 7/10 are given the same weight as those who say they are 10/10. Other pollsters either only take the top group or weight responses in line with the answer. So a 7/10 score would be worth 70% of a 10/10 one. This is probably helping Labour compared with other firms.

Mike Smithson