Are These Polls Serious ?

So THIS, if it is to be believed (and assuming you can read it - I've never uploaded anything of my own creation before) is a representation of voting intentions in the UK since the end of August. You will notice that it's all over the place.

Apparently Labour have moved from having a 13-point lead on 26th September to a three-point deficit today. If this is true voters don't float, they move around in a tsunami. And all this at a time when actually, not very much has happened. It's not as if there's been some huge cataclysmic event - just a lot of talk about a general election and not much else.

I'm sorry, folks, but I just don't buy it. There is no way that public opinion can be as volatile as this. Either the pollsters are just making it all up or something else is at work when they ask the general public who they're going to vote for. I think it's the latter.

I think that when pollsters ask about voting intentions, many people don't tell the truth. they tell the pollster what they think they want to hear, or rather, they tell the pollster what they think it the fashionable answer.

So when Labour are getting postive news stories, they are seen as the trendy party to be associated with, so people say they'll vote Labour. On the other hand, when the MSG gets a slating from the Press and Cameron comes over all confident on the news, they say they're going to vote Conservative. People tell the pollsters what they think everyone else is going to tell them.

But put a ballot paper in people's hands and leave them alone, and they react differently. What does on then is no one else's business. So when it comes to voting, people don't have to worry about what everyone else is thinking.

If we'd had a general election every week (Heaven help us) my guess is the results would have been far more uniform; less buffeted by what's going in in the media.

I know the parties (and the Press, bless 'em) read a huge amount into poll data. And if you look at a rolling "Poll of Polls" or the longer-term trends (say over two or three months) you might get a slightly more accurate view as to what's going on. But day-to-day polls ? Forget 'em.