The weakness of a landslide is the amount of Tories likely to throw votes down the waste trap of UKIP… (and BNP)”

Total partisanship… at least say something about the poll!

The previous ComRes poll found 12% of people saying they voted Lib Dem in 2005. Once you exclude all the don’t knows, didn’t votes and so on, that works out at 21% of those who gave a 2005 recalled vote.

They deliberately don’t match the actual vote shares because of false recall. Panel studies have repeatedly shown that people do not accurately report to pollsters how they actually voted at the previous election, so weighting recalled vote to actual vote would actually give inaccurate results.

Lots more detail here

They are starting to look a little more consistent now, but there is still a fair bit of variation from polls carried out just a couple of days ago, with no obvious cause. Of course it is possible that YouGov had a rogue, but as YouGov tend to be the most reliable pollster I think it is more likely to be down to poll volatility continuing from the conference season.

On a separate point, it doesn’t do the Conservatives any good to be bashing the BNP. Doing it under cover of army officers who are well known Conservative supporters is better than politicians openly doing it, but I think it would be far more sensible for us politically (as well as being the right thing to do anyway) to admit that BNP voters have very real and legitimate concerns, and to address them.