Voodoo Polls

Daily it seems we are confronted, confounded and sometimes infuriated by these phone(y) polls, fax polls, phooey polls and my collective name for them, 'Voodoo Polls'.

These are the polls that come in over the fax machine which promise to be collated and sent to No. 10, relevant House of Commons' committees, on various subjects, the handling of foot and mouth disease, the monarchy, the Euro and yes, the election, which aren't worth the paper they are printed on, and only serve to line the pockets of the unscrupulous people who stoop to this sleazy way of turning a buck.

Or the cheap and cheerful phone-in polls carried out by some newspapers and broadcast media which are really measures of the effectiveness of the pressure groups' ability to mobilise their people to ring in to 'fix' the result. I used to rail at George Ffitch when he ran LBC for doing them when he knew they had no validity whatsoever. I thought my case well made when one 'Desmond' wrote in to the Evening Standard to say how pleased he was that 'his side' won, having voted 157 times himself.

Or more fun, but hardly a reflection of real public opinion, the election marketing voodoo polls, such as Tesco's, or Bass's, or Thornton's, in 1997, or Sainsbury's so far this time, who set out their wares in red, blue and orange, or in the faces of the party leaders, and invite punters to show their party allegiance by buying the product of their (political) choice. Sometimes they accompany their spiel with the spurious claim that these wheezes are 'more accurate than the polls'.

To test their efficacy of measuring public opinion, we kept track of a few last time.

From Explaining Labour's Landslide (Worcester & Mortimore), Politico's

As Peter Snow would say: 'Just a bit of fun'. But most of the Voodoo Polls are not 'just a bit of fun', and it's when they are secondarily reported, with no health warnings that they become damaging to the democratic process, and journalists should watch them, report them for what they are, and not confuse themselves and the public that they represent true measures of public opinion.

OK, so far what's happened in this election? This is its seventh day, and what a roller coaster we've seen. One day there is a 20 point Labour lead, the next a 16 point Labour lead. That's what is quoted in the papers, and that's what's covered on the television. But what do the polls really say about what has happened in the election so far? The table below shows the answer: nothing.

As I have said so often, watch the share, not the gap. In fact, watch the share of the poll for the Conservatives. Four polls published on Sunday and Monday, from fieldwork carried out between Thursday and Sunday, and four polls done by three different polling organisations have the Conservatives at 31%/32%, say 32% plus or minus 1%, four polls have the Liberal Democrats at 14% plus or minus 1%, and four polls have the Labour Party at 49% plus or minus 2%. And these are polls of c. 1,000 people, from which you would expect a sampling tolerance of plus or minus 3%. In other words, nothing has happened.

Observers, as pointed out by Bruce Anderson, have all agreed that the past week has been a strong start for the Tories, based largely around the issue of taxation, and a weak start for Labour, with much scorn heaped on the launch of the election in the south London school by the Prime Minister, generally regarded as overhyped. What did the public think of the Tories after their strong start? Not much. And Labour? Best of a bad bunch. And the Liberal Democrats? Who?

And the polls? Steady. Watch the share, not the gap.

WINNER: Stuart Miller's Campaign day in The Guardian, who has awarded me a much-deserved knighthood in his article Monday. Thanks Stuart. Sir Bob it is (don't tell President Bush; he wouldn't understand).

SINNER: Sunday Telegraph: The graphics of their MORI poll had a bar chart showing a clear Conservative lead on which party was thought to have the best policy on taxation (and delivery of public services) among those who thought tax an important issue (not mentioned), but said in the footnote (Labour lead = 6), when it was the Conservatives who led by six points.

WINNER: Michael Brunson, who used every warning device known to warn people watching the Election Night Programme on ITN to say not to pay much attention to the MORI Exit Poll they'd spent so much money on and we'd worked so hard to get right, as 'Exit polls have been wrong before…'. Interviewing me on Powerhouse (Channel 4, 12 noon each weekday during the election) yesterday, he was generous in his apology for his scepticism four years ago. Just keep your fingers crossed for us this time Mike.