The tables for the latest poll arn't up, but yesterdays (which had Con 37%, Lab 32%) was there.
YouGov weight by newspaper affiliation, and here's how the data looked for yesterday's poll:
Newspaper Type Uneighted Weighted
Express / Mail 370 236
Sun / Star 204 323
Mirror / Record 140 235
Guardian / Independent 128 59
FT / Times / Telegraph 173 140
Other Paper 181 184
No Paper 275 294
And here's what the Guardian reports daily newspaper circulation to be:
Express/Mail 2,790,884 (677,750 + 2,113,134)
Sun/Star 3,647,893 (2,862,935 + 784,958)
Mirror/Record 1,540,255 (1,225,502 + 314,753)
Guardian/Independent 487,480 (300,540 + 186,940)
FT(UK)/Times/Telegraph 1,314,825 (115,447 + 521,535 + 703,249)
Looking at the above, YouGov are right to scale down their Guardian/Independent respondants sharply. The problem seems to be in the tabloid section. They are not getting many raw Sun/Star readers, and getting way too many Express/Mail readers. And it's hard to work out why they've weighted the Mirror up.
I guess this is the problem with online polling. The types of people who read the Sun and Mirror will be construction workers, plumbers, front line people who can't log onto the net during work hours, and who probably can't be bothered to go online when they get home either, especially if there is sport or soaps on TV (or the pub to go to). In addition, you get activists such as BNP types who organise to sign up to all the online polling units - and they probably give their political affiliation as Labour, to tie in with their line, "I used to vote Labour, but..." Hence the reason YouGov ask them what newspaper they read (it's less likely they will fake that, and this may account for the number of Express/Mail respondants in the unweighted figures). You can see why YouGov are having a mare and weighting their respondants to correct for this.
My hunch (and that's all it is), is that the telephone pollsters will be more accurate, as these biases simply won't occur for them. But we shant find out till the general election. Labourites should assume that there is still a ton of work to do, and should not relax till the telephone pollsters show us closing the gap.