Is it all down to how the question is asked?
With a Labour leadership contesting looking more imminent there’s a lot more attention being put on polls showing how people would vote if the leading contender, Gordon Brown, was up against David Cameron.
Surpisingly there’s a big difference between the main pollsters. Taking all the surveys when this was asked since the new Tory leader took office these are the average changes in the CON-LAB margin comparing Cameron-Brown against the standard voting intention responses. The December surveys, of course, had Charles Kennedy as Lib Dem leader.
YouGov: the Tories averaging 2% better in relation to Labour when asked “The next general election will probably not be held until 2009 or 2010. Suppose the leaders of the main parties at that time are Gordon Brown (Labour), David Cameron (Conservative) and Ming Campbell (Liberal Democrat). How do you think you would vote at that election?”
ICM: the Tories averaging 1.75% better in relation to Labour when asked “Suppose the Conservative Party were to be led by David Cameron, Labour by Gordon Brown and the Liberal Democrats by Ming Campbell. If there were to be a General Election tomorrow how would you vote, would you vote Conservative, Labour, Liberal Democrat or for another party?”
Populus: the Tories averaging 5.7% better in relation to Labour when asked “If the choice at the next election is between Labour led by Gordon Brown, the Conservatives led by David Cameron and the Liberal Democrats led by Ming Campbell, which party would you vote for â€“ or would you vote for some other party or not vote at all?”
The difference between the three is the the number of sentences the pollsters use. YouGov, where respondents are at their keyboards inputting on a computer screen has three. Of the telephone pollsters ICM has two while Populus manages to get over the point in one.
In the final sentence the respondents listen to or read YouGov and ICM make no reference to the names of the possible party leaders. Populus does and this is why I believe there’s such a difference.
There’s no right or wrong here – I’ve just been trying to find an explanation for why the three produce the figures that they do.
I’ve made the point before here that the Tories seem to poll better the more they are presented as “Cameron’s Conservatives”. The Populus approach, surely, describes the party in a way that produces more linkage and that is why they appear to do better.
The Lib Dems and Labour should do everything to de-couple “Cameron” and “Conservative”. The Tories should do the opposite.
NOTE: A story in the Telegraph this morning about a YouGov poll showing big differences between party performances in the North and South is not a new poll. It is based on detailed data from last week’s survey showing the Tories 6% ahead.