Populus poll of marginals
We still haven't seen a post-election Populus voting intention poll (though to answer Mike Smithson's question here, I understand they are still doing them, they are just having a quiet period following the election), but Lord Ashcroft has commissioned them to do some polling in marginal seats, with some interesting findings.
Firstly, in Conservative -vs- Labour marginals the Conservative vote is largely unchanged from the general election, but the Liberal Democrat vote has dropped to the benefit of Labour, this means on a uniform swing Labour would gain about 28 seats from the Conservatives (though these would be seats that the Conservatives gained at the last election, so in practice the Tories would be helped by the incumbency bonus of the new MPs).
In Conservative -vs- Lib Dem marginals the Liberal Democrat vote has collapsed towards the Labour party, presumably partially as Labour voters who previously voted tactically for the Liberal Democrats cease to do so (thought it would be interested to know how the questions were worded, since unless prompted people don't necessarily consider the tactical situation in their own constitency when answering voting intention polls). On a uniform swing, this would give the Conservatives about 30 seats from the Lib Dems. The poll did not cover Lab -vs- Lib Dem marginals.
Populus then asked how people would vote under AV. Exactly how they asked this is unclear from Lord Ashcroft's report, but the ultimate effect is that the Conservatives hold onto an extra 12 seats in Con -v- Lab marginals (implying that Lib Dems and others' second preferences broke in the Conservatives favour), and the Lib Dems hold onto an extra 11 seats in Con -v- LD marginals. Once again, we have no indication of what would happen in LD -v- Lab seats.
UPDATE: The full tables are now available on Lord Ashcroft's website (see the links at the bottom of this document). A few things worth pointing out. First, the AV questions were done in much the same way as the YouGov polls on it - people had AV explained briefly to them, then asked how they would cast their first and second preferences under that system (as opposed to making the often false assumption that people would cast their first prefences under AV in the same way as their vote in FPTP).
Secondly, this is the first AV polling I've seen that asked about third preferences. So far only 32% of respondents actually give a third preference (and a third of those were to minor parties who it is unlikely to benefit). At present of course it really isn't comething respondents will have given any thought to, so it won't necessarily bear any relation to how people would actually cast their other preferences, but it's something that would need to be taken into account if AV did come to pass.
Thirdly, the voting intention questions for FPTP were just the standard Populus VI question. That should be okay in the Labour -v- Con seats, but as I've said before, I'm not sure how good marginal seat polls in Lib Dem marginals they. Even when the Lib Dems are riding high in the polls they tend to show the Lib Dems doing badly and Labour gaining, probably because some people give their voting intention as their real first preference, rather than their local tactical one. My suspicion is that the FPTP position in those Con -v- LD seats may be unduly negative.
