Look how the ICM “churn” numbers have changed
| ICM: Oct 8 2009 | CON 2009 | LAB 2009 | LD 2009 |
|---|---|---|---|
| Voted CON in 2005 | 94% | 0% | 2% |
| Voted LAB in 2005 | 20% | 62% | 11% |
| Voted LD in 2005 | 14% | 12% | 68% |
| ICM: March 11 2010 | CON 2010 | LAB 2010 | LD 2010 |
|---|---|---|---|
| Voted CON in 2005 | 88% | 2% | 6% |
| Voted LAB in 2005 | 15% | 70% | 9% |
| Voted LD in 2005 | 10% | 7% | 75% |
Does this tell the story of the squeeze in the polls?
One of the great polling innovations since the 2005 election has been the regular publication, first by ICM and then the other phone pollsters, of the past vote responses linked to current voting intention.
The result is that we can see very easily how the votes are changing and how the parties are succeeding at retaining the groups that should be their bankable targets – those that actually voted for them last time.
Looking at the top row of data, taken from ICM’s NOTW poll last October, we see that Cameron’s party was retaining 94% of its 2005 voters and was also winning one in five Labour voters from the last election as well as 14% of Lib Dem ones. The main loss was a small slippage to the Lib Dems.
Compare that with the bottom table taken from today ICM poll for the Sunday Telegraph. The retention rate is down to 88% and the party, although continuing to be a net gainer from the other two, is not doing as well as it was.
All this comes at a time when the Labour and Lib Dem retention rates have increased sharply.
Even at the March levels I find it hard to square the detailed ICM numbers with the YouGov Labour share showing the party down by just two points since the last election.