the brilliantly leaping gazelle

So, with a tedious inevitability, here we are again and no, I don’t mean you reading my blog. That’s on you, it was your choice, there wasn’t a gun to your head. I mean newspapers presenting opinion survey results as anything other than complete and utter nonsense. Reading tea-leaves is by comparison a much better way to predict the future, having required the drinking of a cup of tea to do so.

It’s not as if the woeful incompetence of opinion polls to accurately predict the result of the 2015 General Election wasn’t remarked upon. One could easily be forgiven for mistaking the same newspapers that were so thunderous in their condemnation as somehow bastions of moral probity. One could then question how it is that a beneficial amnesia has afflicted these same newspapers as they callously presume their readers will have forgotten as well.

The latest – but by no means the last –example of this utter nonsense was to be found in yesterdays ‘Guardian’, which ran a story based on the results of an opinion poll, and claimed on that basis,

Brexit party tops Westminster election poll for first time

Survey by Opinium suggests Nigel Farage would be 20 seats short of a majority, with Conservatives reduced to 26 MPs

How much support the Brexit party would receive at a General Election can only be accurately calculated from the results of that election, because opinion polls only tells one what a statistically insignificant proportion of the population are thinking. There are estimated to be somewhere in the region of 51 million eligible voters in the UK – whether they’ve bothered to register to vote or not is the subject of another post – and yet typically a survey will ask no more than 4,000 what they think. Therefore an opinion poll has all the evidential qualities of a beauty product advert on television. Actually, thinking about it, they’re remarkably similar. Both get widespread attention, but make huge claims based on not very much at all. I mean, have you ever read the small print on a beauty ad? You’d have to be quick though, as they don’t leave it on the screen for very long, for a very good reason. Typically around 80% of about 150 people agree with some claim, sometimes more, sometimes less.

But back to Nigel Farage and his Brexit party. Is he clever or dangerous in the way that so far, there have been no policy idea’s given about what the Brexit party stands for, so therefore everyone can believe it stands for what they believe in. This doesn’t alienate anyone but allows them to project whatever they want.

By doing this, good old Nige, beer drinking, cigar smoking, man of the people Nige, isn’t dangerous. Neither is he clever.

He’s dangerously clever.

And now back to the opinion poll. Have the pollsters massively enlarged the amount of people questioned? And was it conducted using a combination of different interview techniques – face to face, telephone, or online – to give a truly representative sample?

No.

Buried down at the bottom of the article – which has worrying similarities to those beauty adverts – which most people wouldn’t bother to read down to, proved how utterly worthless the survey was the detail that.

Opinium polled 2,005 people online from 28-30 May.

It gets worse. You mean that asking a self-selecting, statistically insignificant sample, unrepresentative of anyone other than themselves isn’t bad enough? No, you don’t mean that….oh you do….they’re paid!