As we await the European election results tomorrow night, Lord Ashcroft has released something very interesting – a collection of constituency polls, focused on many of the main Conservative-Labour marginal seats that will have a substantial impact on the next UK general election. The full results which Lord Ashcroft has posted are here.
Most of these marginal seats, of course, are in England. But one of those that were polled is my own seat, Cardiff North. The poll conducted here was of 1000 voters, by phone, with sampling from the end of March until April 6.
There are a number of different numbers posted for voting intention. Two main questions were asked: one, a standard ‘If there was a general election tomorrow, which party would you vote for?’ and a second that specifically prompted respondents to be ‘thinking about your own constituency’. Fortunately, for simplicity of interpretation at least, these don’t make much difference. The numbers reported (in %) are:
Plaid Cymru: 7
Plaid Cymru: 7
There was also an interesting question about levels of local contact – asking people whether any of the parties had “contacted you over the last few weeks” by various means (including leafletting and canvassing); here, the Conservatives had the edge (26% to 18%) on the percentage of people reporting having been contacted.
So it would seem that Mari Williams of Labour is currently favourite for the much-coveted status of being my next MP, ahead of the Tories’ Craig Williams. In this respect, Cardiff North seems in line with the picture elsewhere; the poll showed Labour generally leading in these key marginal seats, with the constituency voting intention questions averaging a 6.5% swing from the Conservatives to Labour from 2010 – a swing that would be sufficient to give Labour a clear parliamentary majority.
It is maybe worth saying that constituency polls don’t necessarily have a great record. This seems counter-intuitive – surely a poll has a better chance of being right among a population of 60-90,000 people than among one of some 60 million? You might think so, but in practice constituency polls have often been pretty poor. (Although there are exceptions, such as my Aberystwyth students in 2007, who did a constituency poll of Llanelli for the Asembly election, and got the four parties all within 1% of the final result!). Constituency polls often have a relatively small sample (although Lord Ashcroft’s do not) and are frequently conducted some time before an election. Moreover, national polls often get things pretty close to the actual election result by, in effect, ‘averaging out’ the various eccentric local swings in particular constituencies.
However, Lord Ashcroft’s efforts at polling marginal seats are much more detailed than just about anything anyone has ever tried before. We are still a long way from the general election. But his work provides strong evidence that, at least in early April this year, Labour was performing disproportionately strongly in the key seats that both they and the Conservatives need to win next year.