Newspaper polls - a health warning
CAMPAIGNTV Sat Oct 06, 2007 at 09:33:27 AM GMT 
Facebook
Any Labour PM trying to decide when to call an election should treat the published polls with extreme caution.
Newspaper polls ALWAYS exaggerate Labour's lead. If you want a good fix on how many people would actually vote Labour on any given day, the safest policy is to knock five points off whatever they tell the pollsters.
Here are the results of the last four elections, compared with the averages of all newspaper polls published in the last week of campaigning.
Here are the results of the last four elections, compared with the averages of all newspaper polls published in the last week of campaigning.
Election What the polls predicted Actual vote
2005 6% Labour lead Labour win by 3%
2001 15% Labour lead Labour win by 9.3%
1997 18% Labour lead Labour win by 12.7%
1992 2% Labour lead Tories win by 7.6%
The polls have been slightly closer to the actual results since '92, when they got it wrong by nearly 10 per cent. The pollsters' inquest into how they screwed up so badly decided that - they asked the wrong samples, at the wrong time of day, and a lot of people were too embarrassed to say they would vote Tory.
Despite making changes to their sampling techniques, the last week polls were still wrong by more than 5% in '97 and 2001. The point - for anyone who might be spending tomorrow deciding whether we're having a November election - is that the polls are always wrong in Labour's favour.
Why don't the newspapers publish all polls with a little warning sign - "Please remember that this poll gives Labour a much bigger lead than they're actually going to get when you vote. Probably by about 5%. So don't panic." Seriously, why don't they? Either they're lulling us into a false sense of security, or they haven't spotted that the pollsters are providing a very poor service. They should ask for their money back.
In 1992 the pollsters said they hadn't realised people wouldn't own up to voting Tory. But this is exactly what you get on the doorstep, and why the pollsters were so surprised by it is a mystery. And a bit of tragedy because of the influence their rubbish polls had on Labour's campaign tactics. Let's hope we're not led up the garden path again.
Should have written this two weeks ago, but didn't get round to it, being busy setting up campaigntv.org - take a look at some of our films and tell me what you think - gez.sagar@campaigntv.org
2005 6% Labour lead Labour win by 3%
2001 15% Labour lead Labour win by 9.3%
1997 18% Labour lead Labour win by 12.7%
1992 2% Labour lead Tories win by 7.6%
The polls have been slightly closer to the actual results since '92, when they got it wrong by nearly 10 per cent. The pollsters' inquest into how they screwed up so badly decided that - they asked the wrong samples, at the wrong time of day, and a lot of people were too embarrassed to say they would vote Tory.
Despite making changes to their sampling techniques, the last week polls were still wrong by more than 5% in '97 and 2001. The point - for anyone who might be spending tomorrow deciding whether we're having a November election - is that the polls are always wrong in Labour's favour.
Why don't the newspapers publish all polls with a little warning sign - "Please remember that this poll gives Labour a much bigger lead than they're actually going to get when you vote. Probably by about 5%. So don't panic." Seriously, why don't they? Either they're lulling us into a false sense of security, or they haven't spotted that the pollsters are providing a very poor service. They should ask for their money back.
In 1992 the pollsters said they hadn't realised people wouldn't own up to voting Tory. But this is exactly what you get on the doorstep, and why the pollsters were so surprised by it is a mystery. And a bit of tragedy because of the influence their rubbish polls had on Labour's campaign tactics. Let's hope we're not led up the garden path again.
Should have written this two weeks ago, but didn't get round to it, being busy setting up campaigntv.org - take a look at some of our films and tell me what you think - gez.sagar@campaigntv.org
Newspaper polls - a health warning | 0 comments ( topical)
