Last May, it will be recalled, Ipsos-MORI lanuched a major review of its methodology after its performance in the Ken-Boris battle for London.

The firm stopped publishing results for a couple of months before it announced big changes and this latest poll embraces measures to stop Labour being over-stated.

A major element of the reform was to identify and weight the public sector workers in their samples who have tendency to be much more pro-Labour and anti-Tory than the electorate as a whole. MORI looked back at lots of historical data and discovered that this group were being disproportionately represented in their samples and this was impacting on their results.

So in today’s poll MORI found that nearly a quarter of the people they spoke to were in the public sector worker category and their views had to be scaled back by more than half.

So if this poll had taken place before the Boris-Ken reforms then the top-line voting intention figures for those certain to vote would have been C39-L39-LD12.

This gives you an indication of just how stunning this latest survey is.

Stepping back it seems that each alternate poll from the different pollsters seems to be giving a very different picture.

We should be getting a new ICM poll within the next week and this should give us a good steer. This firm is the only one which takes specific measures to ensure a more accurate Lib Dem total and in previous surveys this has impeded on Labour’s numbers.

Mike Smithson