YouGov have AV and FPTP neck and neck
YouGov's latest fortnightly tracker on voting intention in the referendum on the Alternative Vote shows a significant narrowing of the No campaign's lead, with the Yes and No campaigns now virtually neck and neck. In this week's poll 39% said they would vote in keep First Past the Post, 38% would vote to switch to the Alternative Vote. To rule out the chance that this is just an unusual sample or a rogue poll we re-ran the question last night and found the same result - a lead of just 1 point for the No campaign, with FPTP on 38% and AV on 37%. It looks as though there has been a genuine tightening of the race.
YouGov has been asking voting intention in the AV referendum regularly since June 2010. Initially it found a strong lead for the Yes campaign, but this declined throughout last year, largely due to Labour supporters, initially pro-AV, moving against it. By September last year the tracker was showing a consistent lead for the NO campaign, it peaked in November and has been pretty steady since then. The two YouGov polls in January both showed a nine point lead for the No campaign. This week's polling therefore represents not just a much lower lead for the No campaign, but a reversal of the previous trend.
Looking at the breakdown of support in this most recent poll Liberal Democrat supporters, always strongly in favour of AV, have become even more pro-AV, with 84% saying they would vote in favour of AV, Conservative supporters while still opposed to AV were also slightly less hostile, with the proportion of Tory voters saying they would back AV rising to 28%. It may be that the coverage of Labour opposition to the legislation setting up the referendum in the House of Lords has convinced some partisan Conservative and Lib Dem voters that AV itself can't be bad, or it may just be a sign that the Yes campaign is starting to have an effect upon public opinion.
UPDATE: There is a persistent myth about why YouGov tend to show results on the AV referendum than other companies that I've seen crop up again today. Basically there are two explanations put about, one which seems perfectly likely and one which has been tested, found to be wrong, but which refuses to die. This first is that YouGov preface the question with explanations of the two systems and that this makes a difference - in my view this is almost certainly the case, YouGov also tend to show lower don't knows suggesting telling people about FPTP and AV results in some people who would otherwise have said "don't know" saying "No". The second is that it's because YouGov's introductory text identifies the referendum as something the coalition is doing - we know this doesn't make a difference because YouGov have tested it using parallel surveys with one mentioning the coalition, one not, and otherwise identical wording. It made no significant difference.
