Worcester's Weblog - The Political Triangle (c)
MORI chairman Sir Robert Worcester analyses the latest opinion poll data.
Over many years I have evolved a theory of who wins and loses British general elections, the "floating voters", and how they decide which party they will vote for. Every election is won and lost by these floating voters whose votes are up for sale, by the party which is most closely aligned to their perception of the parties' position on issues important to them, plus their understanding of the image attributes surrounding the leaders' image and the image of the parties which they are willing to consider voting for.
The image in my mind is comparable to the child's playing with iron filings and a magnet. I think of the voters as iron filings, and the parties' positions on the political triangle's issues, party image and leader image three-dimensional sheet of paper upon which the iron filings rest until the magnet is twisted, and some iron filings are repelled and some attracted by the force of the magnetic field.
Most voters vote on the values they hold and their traditional loyalties and habits, affected by many things such as their parents' voting behaviour, their class and housing type, their workmates and their family and friends. But the "floaters" are the ones that the parties can move, either because they have struck an electoral nerve on an issue, such as immigration, or on the leader image of not glitz and spin, but on a question of competence, or arrogance, or perceived hubris.
No poll can predict how any individual might vote, and we must recall lessons learned in many past elections, such as the finding from many panel studies we've done in the past when newspapers had money that between a quarter and a third of the electorate switch allegiance during a campaign, and at the last election, some eight percent said they'd made up their mind whether or not to vote, and for whom they would vote in the final 24 hours of the election campaign.
In this election, at least at the outset, there was a greater chance of volatility than in any of the last four elections. At the beginning of April, 41% of those who did express their intention to vote for a particular party said they might change their allegiance over the month between then and polling day on May 5th.
In our survey for the Financial Times 7-11 April, we updated the aggregate of the electorates' voting determinates, represented by the tetrahedron below. In this election, issues are thought important to 47% of the public, up sharply from the last two elections, as in 1997 41% and in 2001 42% thought issues an important determinant of how they would vote. The increase in salience of issues came pretty equally from the decreases in party and leader image values.
Today's New Poll: Populus on 39% Labour, 33% Conservative, 21% Liberal Democrats
The only poll on offer this evening is the start of the Populus rolling poll for ITV News and the Times, in which the pollsters, led by former Tory Central Office apparatchiks, will be producing a poll every day from now until election day, interviewing a few hundred electors by telephone each day, and aggregating over several days which started out yesterday evening on 1,400 interviews, with a voting intention spot in the middle of yesterday's NOP, MORI and themselves.
Each day from now on they will drop the 350 or so interviewed four days ago, add interviews with c. 350 new respondents, and report a new 1,350 to 1,450 aggregate result. Result? Cuts the costs of the polls both by splitting costs between the two clients, and allows them to claim a sample size of c. 1,400 people while only paying for interviews with c. 350 people each day. Populus is showing a six point Labour lead. The British Elections Survey people are also into the "rolly polly" game, and showed a one point Conservative lead in the Mail on Sunday. Can they both be right?
