Labour drop to third place
I've been speculating about it for a couple of months, but in the Sunday Telegraph tomorrow we finally see a poll (as suspected from ICM, who tend to give the Liberal Democrats their highest levels of support) putting Labout in third place. The topline figures, with changes from ICM's last poll, of CON 40%(+1), LAB 22%(-6), LDEM 25%(+5). The Lib Dems caught Labour as recently as 2003, after the Brent East by-election, but as far I can see one has to go back to 1987 to find them ahead of Labour.
The Lib Dem score contrasts wildly with Populus's yesterday - the two companies use very similar methodology. Populus's fieldwork is conducted by ICM, their weighting figures are very close, they carry out almost the same re-allocation of don't knows by past vote, the fieldwork dates for the two polls were the same. Possible differing approaches to polling the European election shouldn't make a difference, since Westminster voting intentions were asked first. There is a slight difference in the question that is asked, but my guess is that most of the difference between these polls must be down to sample error.
ICM also asked about European voting intention. Topline figures, with changes from ICM's last poll a week ago, are CON 29%(-1), LAB 17%(-7), LDEM 20%(+2), UKIP 10%(nc), Green 11%(+1), BNP 5%(+4). Again, we have a sharp contrast with Populus, who put UKIP second and the Lib Dems fourth. Mike Smithson is speculating that the difference might be down to ICM not prompting using the names of the minor parties, that would explain the difference in UKIP support - but ICM and Populus are showing broadly similar Green and BNP support. I guess we'll have to wait for the tables to see.
UPDATE: Darrell in my comments has looked through past polls more carefully than I - there was a single poll in 2004 that had Labour in third place too.
