by Stephen Tall on March 13, 2005
Though Tony Blair feigns bemusement that anyone could possibly think he’s yet set the date for the next general election, if it’s not the 5th May then I’m a banana. Unlike 2001, this election should be interesting – even if the result’s not in doubt, the scale of Labour’s victory certainly is. Frankly, it’s a mug’s game trying to predict anything… so here goes.
The obvious first thing to do is to look at the opinion polls
1992 General Election. Opinion Polls bias to Labour 9.5%
1997 General Election. Opinion Polls bias to Labour 3.5%
2001 General Election. Opinion Polls bias to Labour 6.5%
Average opinion poll bias to Labour in the last three General Elections – 6.5%
In only two of the General Elections over the past fifty years have the pollsters not over-stated Labour – February 1974 and the Tory landslide of 1983. Since then the record has been abysmal.
1992 General Election. The vast majority of polls were over the top with the Labour share with one or two notable exception Gallup and Harris. Not one single poll produced results that even hinted that John Major would win by a margin of 8%. Probably the biggest polling failure there has ever been.
1997 General Election. The pollsters fared better but forty-eight of the 50 polls in the final month over-stated Labour by upto 11%. Two ICM surveys stopped it being a clean sweep of failure
2001 General Election. Every single poll from every single pollster got Labour wrong and the error in every case was an over-statement. Even Rasmussen – the only pollster to get the Tories right – had Labour 2% higher than it was.