The ICM telephone poll is now being carried out - I know that because they called me.
After answering the girl's questions, I started to ask her a few of my own for this blog. She put me through to a supervisor who explained that they are a consumer marketing firm who do political polls as a sideline. I knew that already, since ICM are the only polling outfit that got the 2011 AV referendum bang on.
He explained that the calls are made via an automatic dialler, and the girls then start by asking the person's age range. It seems that the company has enough answers already from people over 65, and are now only polling the 16 to 64 year olds. He went on to explain that the poll should be ready over the weekend, but I hope that it is leaked earlier, since this is the one that everyone is waiting for.
ICM are known to weight down by 50% the people who did not vote in the last election. However, they also asked about the previous UK general elections as well as the last Scottish one, which suggests that they are weighting for both in this poll. That may be a problem if turnout is very high next week since only 50% voted in the last Scottish elections.
The main objection I have with the ICM method is that it only covers people who have landlines, which excludes an awful lot of young people, those who live in shared houses, and the poor who rely on PAYG mobiles. That is a big chunk of the Scottish population, so an awful lot of weighting must be going on.
That means that the resulting poll may not be as accurate as the AV one of 2011.