Athough ahead on votes te Tories would trail on seats
Unfortunately I was tied up in a meeting when Lord Ashcroft’s first weekly phone poll came out and am only now just catching up.
Looking at the methodology this seems to be following ICM or how Populus used to do it when they were pollster for the Times. The sample is past vote weighted and there is a spiral of silence adjustment. It looks sound.
The only caveat I’ve got is that of the 1,001 people interviewed only 487 expressed a voting intention. This is low and does increase the margin of error.
Hopefully we’ll get the chance to compare these numbers with the ICM May poll which is due out tonight or tomorrow.
The big news is that at last we have a poll showing a crossover. Feeding the numbers into the Electoral Calculus seat predictor Labour still comes out on top. This has CON 294, LAB 310, LD 18, NAT 9, OTH 19, UKIP 0.
Those figures almost certainly flatter the Tories because they assume 23 gains from the Lib Dems. If they got half of that they would be doing very well. My calculation taking into account the strong LD incumbency effect is that the Tories would be on about 280 seats.