Lies, damn lies and polling

I used to be a pollster (before they lost their way) and here’s a quick brief on why you really should ignore them this time around: first, front page figures are still using national samples – not detailed results from the seats that are actually in contention; second, there is more reliance on the ‘art’ of polling now than before (it has always been a balance of art vs science) with different companies just guessing likely turnout among key voter groups; and third, there’s evidence that voters are increasingly aware of the impact of polling on the political narrative, so they use it to send a message rather than describe how they’ll vote – think petitions rather than commitments. There are one or two honourable exceptions, but most polling companies will prove they’re lucky rather than good if they get close to the final result.