More Tosh From The Pollsters

I’ve blogged in the past what I think about opinion polls. I still completely refuse to believe that public opinion is anything like as volatile as the polls make out. If you back the polls, than a 13-point Labour lead in late September got turned into a 12-point Conservative one just before Christmas, which has now been cut to four points. This just cannot be.

I believe that many of these UK polls are hogwash. But, I have to admit, very few of them are talking as much gibberish as their brothers and sisters across the Atlantic were spouting last night. USA Today was reporting a 13 per cent lead for Obama over Clinton yesterday, and yet what actually happened was a two-point win for Clinton. It reminds me of the Presidential Race of 2004, in which the UK election Night programmes, guided by the exit polls, started discussions with their panellists as to why Kerry had won. A few hours later, the same people were attempting to explain Bush’s victory. To me it demonstrates the level of scepticism needed when viewing polling data.

I don’t know a great deal about how pollsters arrive at their conclusions. I appreciate it’s not as simple as going up to 100 people in the street, asking them which way they intend to vote, and giving the raw results. As well as trying to make sure they get a representative sample of the electorate, they have to take into account that some parties are habitually under-represented / under-stated in polls, that not everyone’s going to vote, the precise timing of their polling and all sorts of other factors. But even so, I do find it pretty extraordinary that they can still get things so dramatically wrong.

I still think that the thing we have to be most wary of in interpreting polls is the “tide” factor. Since Iowa, Obama has been seen as having momentum – the “cool” guy to back. I suspect that influences respondents, who, not wanting to look “out of touch” or unfashionable, tell pollsters that he’s the one they’ll vote for. Once in the polling booth, however, they tell their ballot paper how they really feel. So Obama got a polling bounce because some people wanted to convince the pollsters that they had their fingers on the political pulse, whereas actually nearly 40% of them were planning to vote Clinton all along.

However much I criticise the pollsters, I have to say that they’re far better at the job than I ever managed in my one attempt at polling. On the day of the 1979 General Election, aged 14, I decided to do some polling of my own. I marched out into the street, clipboard in hand, and then walked around for ages, looking for someone who looked “nice” so that I could ask them whom they’d voted for. Finally I plucked up the courage, and asked someone who, quite understandably, didn’t want to discuss politics with a spotty teenager. At that point I decided it was all a waste of time, and retreated home, tail between legs, to watch the result unfold on telly (actually, when I found out the nothing would become clear until about 2 in the morning, I think I just went to bed). So according to the first ever (and only) Womble On Tour Opinion Poll, the 1979 General Election was a nil-nil draw.

For interest, and to underlined the volatility of the UK polling market at present, I’ve updated my graph showing the results of all polls since the start of September.